SMILE: The Universal Methodology That Turns Any Project Into a Continuously Learning Digital Twin

Nicolas Waern WINNIIO AB, Gothenburg, Sweden ORCID: 0009-0001-4011-8201

Corresponding author: ceo@winniio.io

License: CC-BY-4.0

Submitted: 2026-04-16

Cite as: Waern, N. (2026). SMILE: The Universal Methodology That Turns Any Project Into a Continuously Learning Digital Twin. WINNIIO AB. Zenodo. https://doi.org/10.5281/zenodo.20175406

Publication status: This paper is published as a working paper and corpus anchor on Zenodo under CC-BY-4.0. It is explicitly positioned as a design science artefact presentation (Hevner et al., 2004) rather than an empirical validation study. Journal submission with independent multi-site validation data is planned as a subsequent contribution.

"The sciences of the artificial are concerned not with how things are, but with how they might be." — Herbert A. Simon, The Sciences of the Artificial (1996)

We are getting out of the data dark ages into an impact renaissance — where the future is better, bolder, and more sustainable than the past. SMILE is the foundation layer. The building of a dynasty.

Prior works in this series: Waern (2025a). DOI: 10.5281/zenodo.17462962 Waern (2025b). DOI: 10.5281/zenodo.17464804 Waern (2026a). DOI: 10.5281/zenodo.19587944 Waern (2026b–z). [Series of domain applications listed in Section 14]

Abstract

This paper argues a specific and falsifiable claim: any project — regardless of domain, scale, or sector — that applies SMILE (Sustainable Methodology for Impact Lifecycle Enablement) will, as a consequence of that application, become a digital twin project. The digital twin is not the objective; it is the outcome. SMILE is not a methodology for digital twin projects; it is a methodology that makes digital twinning the natural result of doing any project well.

This reframing positions SMILE alongside PRINCE2 (which is not a methodology for construction or IT but for any project), TRIZ (which is not a methodology for engineering invention but for any inventive problem), and Design Thinking (which is not a methodology for product design but for any human-centred problem). SMILE occupies the same universal register applied to the specific challenge of creating systems that continuously learn from reality.

Applying SMILE to any project delivers twenty structured benefits across five domains: operational (virtual-first validation, real-time situational awareness, predictive maintenance, scenario simulation), knowledge (institutional memory, knowledge acceleration, cross-domain insight, explainable decisions), economic (resource optimisation, revenue visibility, reduced waste, twin-as-a-service business models), social-organisational (stakeholder alignment, democratised access, remote enablement, inclusive governance), and systemic (interoperability by design, sovereignty, regulatory compliance, planetary impact). Of the twenty, five have measurable supporting data from published corpus applications; the remaining fifteen are stated as expected benefits based on phase logic, with empirical validation in progress.

SMILE is also a Physical AI methodology. It operates in the Web4 paradigm — the symbiotic web where physical and digital merge — through Spatial Twinning (Space, Place, Interaction, Network). Every phase has a physical dimension. This is not software engineering; it is reality engineering.

The methodology is grounded in five theoretical pillars: Design Science Research, Actor-Network Theory, boundary object theory, absorptive capacity, and benefits management. It is contextualised within the Gothenburg School's Five Model information systems tradition (Pessi, Magoulas, Hugoson) and draws on concurrent engineering practice established at major space agencies for its concurrent engineering and real-time phases. The Crucible trust layer — retrieval over generation, explainability by design — governs all AI outputs. The AI journey within SMILE (Data Contextualisation → AI-Ready → AI-Infused → AI-Ingrained → Explainable AI Decision Making) is sequenced by the same impact-first logic that governs the six phases.

The paper provides complete phase-by-phase specification across six phases and four perspectives (From People, From Systems, From Planet, From AI), a mapping to the Digital Twin Consortium Capabilities Periodic Table (DTC CPT v1.1), cross-domain validation across eight verticals, a five-level maturity model, comparative analysis against five methodologies including TRIZ and PRINCE2, three falsifiable propositions, and an honest account of limitations. It is intended as the primary SMILE citation across the 39-paper corpus it anchors.

Keywords: SMILE methodology, AEST, NUDEDA, digital twin, Physical AI, Spatial Twinning, SPIN Questions, Web4, Minimal Interoperability Mechanisms, Five Model, holographic society, Crucible, design science research, Actor-Network Theory, boundary objects, absorptive capacity, benefits management, explainable AI, impact-first, lifecycle methodology, maturity model, Large Quantitative Models, edge-native, knowledge graph, ontology

1. Introduction: The Wrong Question Has Been Driving the Wrong Answers

1.1 The Cave, the Fabric, and the Future You Can Borrow

Humans are not meant to stay face to face and discuss things in isolation. When we use the walls in a cave — when we use some representation of reality that acts as a boundary-spanning object — that is when people, systems, and AI can gather round, see changes happening together. We can both describe the cave and walk out of it together.

The traditional group development cycle — forming, norming, storming, adjourning, performing (Tuckman, 1965) — assumes sequential, face-to-face interaction. When stakeholders instead gather around a shared reality representation — a digital twin, a simulation, a Reality Canvas — concurrent engineering methodology enables teams to compress this cycle by up to 99% compared to traditional methods. Seeing, sensing, and immersing becomes believing.

This is not limited to space and time. We can emulate reality, ask where we are heading, see the future before it happens, and then alter outcomes prior to their materialising in physical reality — while still observing changes in the physics-based ground truth. We deal with outcomes and impact trajectories, not data. SMILE is future-impact-driven: simulations and future states always maintain correlation to their past and present. Reality is not broken — it is extended.

The strength of weak ties (Granovetter, 1973) are not limited by either space nor time, and can span both real and fictional contexts. My future self can help me right now, with the assistance of past memories, across any industry and any geography. Users of SMILE can invite other people, systems, and AI to run simulations and move across intended and unintended scenarios at computational speed, always aligned with modern compute paradigms — including high-performance computing and quantum simulation.

This requires what Iansiti and Lakhani (2020) call 'AI factory floors' — rigorous architectural foundations that increase scale, scope, and learning across space and time. The keystone advantage (Iansiti, 2004; Gawer & Cusumano, 2014) of a platform is not the platform itself but its ability to create value for its ecosystem participants. SMILE provides the methodological architecture for these factory floors: the six phases define what the factory produces; the four perspectives ensure it serves all stakeholders including AI agents themselves; and the AEST temporal model ensures it operates across past, present, and future simultaneously.

1.2 The Wrong Question Has Been Driving the Wrong Answers

For two decades, the dominant question in the digital twin field has been: "How do we build a digital twin?" The result is a field that has abundant technology, abundant standards, and abundant vendor offerings — and a persistent, documented failure to deliver sustained value. Analysts at Gartner have watched digital twins traverse hype curves. The Digital Twin Consortium lists hundreds of members. The European Union has funded digital twins of the Earth's climate, of individual human bodies, and of its cities. The field is large, active, and technically capable.

Yet the most common reported causes of digital twin project failure remain stubbornly non-technical (Kritzinger et al., 2018; Fuller et al., 2020; Jones et al., 2020; Rasheed et al., 2020): unclear business objectives, inability to demonstrate return on investment, organisational resistance, disconnection between the twin and operational decisions. These are not isolated observations — they represent what Rittel and Webber (1973) characterised as wicked problems: sociotechnical entanglements where the problem definition itself is contested, where there is no stopping rule, and where every attempted solution changes the problem. These are not problems that better sensors or faster pipelines resolve. They are the consequences of a wrong starting question. Christensen (1997) identified the same dynamic in disruptive innovation: incumbents fail not because they lack capability, but because they optimise relentlessly for the wrong metric — their existing customers, their existing measures of performance. SMILE's impact-first principle is a direct methodological response to this pattern applied to digital infrastructure. Brynjolfsson (1993) named the same failure mode in IT investment: organisations added computational capability without redesigning the work it was meant to serve, producing the productivity paradox — more technology, no measurable value gain. Data-first digital twin deployments replicate this paradox faithfully.

The right question is: "What change in the world are we trying to create, and for whom?" This reframing echoes Ackoff's (1989) foundational distinction between data, information, knowledge, understanding, and wisdom — the DIKW hierarchy that most technology projects climb from the bottom, hoping to reach the top, when the methodology should begin at the top and work down. Any project that answers this question rigorously — that traces a clear chain from desired outcome through required action through necessary insight through processed information to collected data — will discover that it needs a continuously learning representation of reality grounded in physical truth. It will, in other words, produce a digital twin as a consequence of doing good project work. The digital twin is not the objective. It is what happens when a project is designed to actually work.

This is the foundational claim of SMILE — Sustainable Methodology for Impact Lifecycle Enablement: it is not a methodology for digital twin projects. It is a methodology that turns any project into a digital twin project. The distinction is the same as the difference between PRINCE2 (which does not exist to build software or bridges but to govern any project that delivers change) and a software development methodology. PRINCE2 produces projects that happen to involve code when the domain requires it. SMILE produces projects that happen to involve digital twins because any project that sustains impact over time requires a continuously learning, reality-grounded representation of what it is working with.

TRIZ (Altshuller, 1996) is a second comparison. TRIZ is not a methodology for mechanical engineering. It is a methodology for solving any inventive problem by identifying and resolving the contradiction at its core. Engineering produces most of the published TRIZ cases because engineering produces most of the documented inventive problems — but TRIZ has been applied to business strategy, education design, and social innovation. SMILE occupies the same universal register applied to the specific challenge of sociotechnical systems that must learn and improve over time.

SMILE was introduced in Waern (2025a) and applied across a corpus of 39 papers spanning eight domains. This paper provides its complete theoretical foundation, the twenty benefits that follow from its application, the six-phase specification in full detail, the AI journey embedded within the methodology, the Crucible trust layer, the MIMs interoperability framework, its relationship to the Gothenburg School's Five Model and concurrent engineering practice from major space agencies, its alignment with the Web4 paradigm and Physical AI principles, a mapping to the DTC Capabilities Periodic Table, and the holographic society endpoint that Phase 6 anticipates. It is the primary reference for all SMILE citations in the corpus.

The paper proceeds as follows. Section 2 situates SMILE within the Web4 and Physical AI paradigm. Section 3 establishes the theoretical foundations. Section 4 introduces the Gothenburg School lineage and concurrent engineering connections. Section 5 presents the twenty benefits of applying SMILE to any project. Section 6 specifies the AI journey and Crucible trust layer. Section 7 presents the complete six-phase methodology. Section 8 provides cross-domain validation and DTC CPT mapping. Section 9 introduces the maturity model. Section 10 compares SMILE with existing methodologies including TRIZ and PRINCE2. Section 11 engages counter-arguments. Section 12 states falsifiable propositions. Section 13 addresses limitations and future research. Section 14 concludes. Section 15 references.

2. Physical AI and the Web4 Paradigm: Where SMILE Operates

SMILE is inherently a Physical AI methodology. The phrase requires unpacking, because "AI" in dominant discourse refers almost exclusively to software models processing digital data. Physical AI describes something different: intelligence that is grounded in, acts upon, and is continuously calibrated by the physical world. A Physical AI system does not merely process representations of reality; it maintains a living model of reality that can be acted upon, simulated, and evolved.

Every phase of SMILE has a physical dimension. The concept of the digital twin itself originates in NASA's paired-vehicle approach (Glaessgen & Stargel, 2012), where a physical asset and its virtual counterpart evolve together throughout a lifecycle. Grieves (2014) formalised the conceptual model; SMILE operationalises it into a methodology. Phase 1 produces a physical Reality Canvas — a spatially and temporally grounded 3D+ representation of the real environment being addressed. Phase 2 validates interventions virtually before they are executed physically. Phase 3 connects physical sensors to the ontological model. Phase 4 makes the physical state queryable in real time. Phase 5 deploys intelligence at the physical edge. Phase 6 transfers physical-world knowledge across contexts. SMILE works on both real and fictional representations of reality. The critical requirement is not that the representation be physically accurate, but that people, systems, and AI have a shared artefact to rally around. A completely simulated future that has never existed can be emulated by reality, and real reality can be emulated digitally. There is a dynamic, bidirectional interaction between strategy, impact, and technology — between what IS and what COULD BE. As long as the representation maintains correlation to physical possibility, the methodology operates regardless of whether the current state is observed or imagined.

This positions SMILE within the emerging Web4 paradigm and beyond — the symbiotic web, where the boundary between physical and digital dissolves into continuous, bidirectional exchange (Hendler & Berners-Lee, 2010; Allam et al., 2022). Web1 was readable. Web2 was participatory. Web3 was decentralised. Web4 is symbiotic: physical objects, biological systems, urban environments, and human bodies become first-class participants in the computational fabric. Intelligence moves to the edge because that is where reality happens (Shi et al., 2016; Satyanarayanan, 2017). Local inference precedes cloud synchronisation because sovereignty requires that the computational mirror of reality not be contingent on network availability.

Beyond Web4, the trajectory points toward what this paper terms WebX — the physics-based web where digital representations are not abstractions of reality but computationally faithful extensions of it. WebX environments are not metaphors; they are physics engines running at sufficient fidelity to replace physical presence. The progression is not speculative — it is already underway through 6G joint communication-and-sensing (JCAS) standardisation (Tataria et al., 2021; ITU-R, 2023), volumetric capture, holoportation, and ambient computing. WebX is the environment in which SMILE's Phase 6 (Perpetual Wisdom) fully activates: when seeing, sensing, and immersing become indistinguishable from physical co-presence, the governing question shifts from "how do we build the twin?" to "how do we govern the shared reality?"

This is the understanding, acceptance, and architectural alignment with what we term Hi-Fi Holographic Societies — environments characterised by ambient computing (intelligence in every surface), extended reality (spatial interfaces replacing screens), secure edge (sovereign local computation), differential privacy pipelines (mathematical guarantees on aggregate insights), outcome-based knowledge transfer (the twin IS the medium), and 6G+ connectivity (terahertz sensing, sub-millisecond latency). SMILE does not merely anticipate these environments — it provides the methodological governance layer they require. Without a methodology that operates across time (AEST), across perspectives (People, Systems, Planet, AI), and across fidelity levels (text through full sensory), holographic societies have technology but no coherence. SMILE provides the coherence.

SMILE's spatial framework for Web4 implementation is Spatial Twinning: Space, Place, Interaction, Network. Every digital twin created through SMILE addresses all four dimensions.

Space is the objective, measurable physical environment — the geometry, material properties, electromagnetic characteristics, thermal gradients, and gravitational context of the physical reality being twinned. This is the domain of sensors, photogrammetry, satellite data, and physics-based models (Large Quantitative Models; see Section 6). Space does not care about human interpretation; it has properties that exist independently of any community's description of them.

Place is the humanly meaningful interpretation of space — the hospital ward that is also a boundary of jurisdiction, the stable yard that is also a community of care, the urban district that is also a political unit. Place is where ANT's obligatory passage points form: the shared meaning that different communities negotiate around a common spatial referent. SMILE Phase 1's Reality Canvas creates the shared Place from the raw material of Space.

Interaction is the ongoing exchange between human actors, non-human actors (sensors, systems, algorithms), and the twin itself. SMILE's four perspectives — From People, From Systems, From Planet, From AI — are the four Interaction channels through which knowledge flows into the twin and decisions flow back out. The Interaction layer is where the twin's artefactual absorptive capacity (Section 3.4) grows.

Network is the ecosystem of interconnected twins, systems, and actors within which any single twin is embedded. A twin is not an isolated artefact; it is a node in a knowledge network that spans domains, organisations, and geographies. SMILE Phase 6 (Perpetual Wisdom) is the methodology's explicit engagement with the Network layer: how does knowledge flow between nodes, how is provenance maintained, and how does the network as a whole become smarter than any of its constituent twins?

Spatial Twinning makes SMILE's universality concrete. Any project that involves real-world action — which is to say, any project — operates in all four spatial dimensions simultaneously, whether or not it acknowledges doing so. SMILE makes the acknowledgement explicit and turns it into methodology.

The Zoom-Anchor-AEST Interaction Model

SMILE's interaction with physical reality follows a three-step pattern that operates at any scale and in any domain.

1. ZOOM to any scale. SMILE operates from quantum to universe, without limit in either direction. A practitioner may zoom to a cell membrane to understand a drug interaction, zoom to a horse's metabolic state to calibrate a biosensor, zoom to a city block to optimise emergency communications, or zoom to a planetary carbon cycle to evaluate a land-use decision. The scale changes; the methodology does not.

2. ANCHOR on any point. Once zoomed to the relevant scale, the practitioner anchors on a specific entity — a cell, a horse, a building, a city, a supply chain. The anchor is the obligatory passage point (Callon, 1986) around which the actor-network organises. The Reality Canvas of Phase 1 is the formalised anchor.

3. AEST activates at the anchor. At any anchor point, all four temporal operations become available: see the past (Absorb the institutional memory and historical record), see the now (Emulate the living present-state twin), see the future (Simulate forward scenarios and test interventions virtually), and borrow from the sustainable future to build it now (Transcend current constraints by reverse-engineering the path from the desired future state back to the present action).

The simulation is not prediction — it is aspiration made actionable. The practitioner sees the sustainable future that is achievable given current physics and constraints, reverse-engineers the path from NOW to THERE, and begins building. Transcendence in SMILE is not speculation; it is structured reverse causation from a defined target state. As WINNIIO's original MIMs Explainer framed it: "Companies and cities can simulate the future to transcend the now."

3. Theoretical Foundations

SMILE rests on five theoretical pillars, each addressing a different dimension of why the methodology is structured as it is.

3.1 Design Science Research: SMILE as a Rigorously Constructed Artefact

Simon (1996) distinguished the sciences of the natural — concerned with how things are — from the sciences of the artificial — concerned with how things ought to be. Design science lives in the latter. Hevner, March, Park, and Ram (2004) established Design Science Research (DSR) as the methodology for creating and evaluating IT artefacts designed to solve identified organisational problems. DSR's seven guidelines require that the artefact be designed (not discovered), solve a previously intractable problem, be evaluated against stated criteria, contribute to design knowledge, and be communicated to both practice and research audiences.

SMILE is itself a DSR artefact. The problem it addresses — the systematic failure of data-first digital twin implementations to deliver sustained value (Fuller et al., 2020; Grieves, 2014; Kritzinger et al., 2018) — is empirically documented. The artefact — the six-phase methodology with defined entry criteria, activity sets, and exit criteria — is the design response. The evaluation is provided by cross-domain application across 39 publications. The contribution to design knowledge is the methodology itself plus the theoretical mechanisms that explain why it works.

Every application of SMILE to a new domain is simultaneously an evaluation of the methodology's transferability. The falsifiable propositions in Section 12 operationalise DSR's evaluation requirement.

3.2 Actor-Network Theory: The Twin as Obligatory Passage Point

Actor-Network Theory (ANT), developed through Callon (1986), Latour (2005), and Law (1992), provides SMILE's sociological foundation. ANT holds that agency is a relational effect that emerges within networks, and that non-human artefacts participate in the constitution of those networks on equal footing with human actors.

This sociomaterial reading complements the sociotechnical systems tradition (Emery & Trist, 1960; Baxter & Sommerville, 2011) while offering a more symmetrical treatment of human and non-human actors. Applied to digital twins, ANT explains what SMILE Phase 1 (Reality Emulation) is actually doing. Before a twin can absorb knowledge, it must be enrolled as an obligatory passage point: the sociotechnical actor through which all other actors must pass to achieve their respective goals. Callon's four moments of translation — problematisation, interessement, enrolment, and mobilisation — map directly onto SMILE's early phases.

ANT explains why data-first approaches fail: a sensor network that has not been enrolled as an obligatory passage point has no actors — human, systemic, or artificial — organised around a shared representation of reality. It collects data that nobody asked for, in response to questions nobody posed, producing dashboards nobody checks. But the deeper failure is ontological, not organisational: the sensor network is disconnected from the physics-based reality that is always already there. The planet — our planet — is the most obvious anchor for any project, not merely because addressing it makes projects more sustainable by design, but because physics-based reality is the ground truth that all actors — people, systems, AI — exist within, not around. Not recognising this as a foundational pillar is one of the things humanity takes for granted and then forgets. SMILE's impact-first principle is the ANT-grounded corrective: it ensures the twin is enrolled into the actor-network before data collection begins, and it ensures the twin is anchored to the physics of the planet rather than floating in an abstraction layer that has lost contact with the real.

A fully worked ANT translation chain: the biological digital twin case. To ground this theoretically, it is instructive to trace the four ANT moments through the most intimate and universal SMILE deployment — the human body as a digital twin. In the Life Atlas platform (Waern, 2026a), an individual creates a biological digital twin anchored to their own physiology, medical history, wearable sensor data, and self-reported observations.

Problematisation occurs when the individual frames the problem: their health data is fragmented across clinics, wearable apps, lab results, and memory. No single representation exists that allows them — or their clinicians, AI agents, or family caregivers — to see the whole picture. The biological digital twin is positioned as the obligatory passage point: the artefact through which every subsequent actor (the individual themselves, their physician, their AI health agent, their family, and the planetary health context they exist within) must pass to achieve their respective goals. The planet itself is ever-present as the physics-based anchor — the gravitational, atmospheric, and ecological context within which the body exists.

Interessement occurs when the twin demonstrates that it serves each actor differently through the same artefact. The individual sees their health timeline — past, present, and projected future. The physician sees clinical data contextualised by lifestyle patterns the clinic never captured. The AI agent sees structured data it can reason over to detect patterns invisible to human observation. The caregiver sees actionable guidance. Each is held in place by a different promise, but all promises are fulfilled by the same reality representation. Crucially, the individual can converse with past versions of themselves — reviewing decisions, symptoms, and contexts that led to present outcomes — and with simulated future versions, exploring how current choices might play out. This temporal self-dialogue is the AEST model made personal: absorb your past, emulate your present, simulate your future, transcend your current understanding.

Enrolment occurs when the biological twin achieves obligatory passage point status: the physician references it during consultations; the AI agent grounds its recommendations in it; the individual checks it before making health decisions. The twin is not a dashboard — it is the medium through which the actor-network stabilises. The individual can zoom out to see themselves from the outside — their body in its planetary context, their health in its environmental context — and zoom in to cellular-level data, biomarkers, and mechanistic models. They can also choose to open-source their reality — to share their twin's anonymised patterns with others who might benefit, enabling others to "walk in my shoes," see what they see, and potentially help them reach their goals faster. Humans, systems, and AI agents alike access this shared reality through the LPI (Life Programmable Interface), which provides meaning, knowledge, and data in omni-directional ways.

Mobilisation occurs when decisions — medication changes, lifestyle interventions, preventive actions, clinical trial eligibility — are made on the basis of the biological twin rather than on fragmented records and episodic consultations. The knowledge that had been locked in clinical silos, personal memory, and wearable app databases becomes actionable through the enrolled artefact.

The translation breaks down when it does precisely because earlier moments are incomplete: individuals who were not genuinely enrolled during interessement — who experienced the twin as surveillance rather than sovereignty — disengage from data entry, producing a twin that decays rather than grows. This is an IS finding, not anecdote: actor-network stability requires that the individual experiences genuine sovereignty over their biological twin, not merely access to it. SMILE Phase 1's exit criterion — the Reality Canvas must be confirmed by all primary stakeholder communities — is the operational response to this ANT-grounded insight. In the biological case, the primary stakeholder is the person whose body is being twinned. Without their genuine enrolment, no subsequent phase produces trustworthy output.

3.3 Boundary Object Theory: Serving Multiple Communities Without Requiring Consensus

Star and Griesemer (1989) defined boundary objects as artefacts that inhabit multiple social worlds simultaneously, satisfying each community's informational requirements through interpretive flexibility while maintaining sufficient structural robustness to enable coordination.

SMILE-produced twins function primarily as coincident boundary objects: they share a common physical referent (the reality being twinned) while each community reads internally different content from that referent. The RF engineer reads electromagnetic propagation; the structural engineer reads load distribution; the facility manager reads energy flows; the urban planner reads population dynamics. All are reading the same scene. None must abandon their domain vocabulary to do so.

Reality itself — not a model of reality, but reality digitally rendered — is the ultimate boundary object. This is the core thesis from which SMILE flows: physical reality is the one thing that all communities share regardless of disciplinary vocabulary, institutional affiliation, or cultural background. SMILE's Reality Canvas operationalises this insight by making the shared physical referent the first deliverable, before any community-specific model is built.

Star (2010) cautioned against treating every shared artefact as a boundary object. SMILE's specificity addresses this: the methodology prescribes the conditions under which a twin acquires genuine boundary object status — defined entry criteria, multi-perspective development, multi-community use — rather than treating every implemented twin as automatically boundary-spanning.

3.4 Absorptive Capacity: The Path Dependency of Knowledge Accumulation

Cohen and Levinthal (1990) defined absorptive capacity as an organisation's ability to recognise, assimilate, and apply new external knowledge — and demonstrated that this capacity is path-dependent: it grows faster in domains where prior knowledge is denser.

Nonaka and Takeuchi (1995) demonstrated that organisational knowledge creation follows a spiral from tacit to explicit and back — the SECI model. The digital twin is the artefact that makes this spiral visible and persistent: tacit knowledge deposited through interaction becomes explicit knowledge encoded in the graph, which is then internalised by subsequent users. Schön's (1983) reflective practitioner — the professional who learns through reflection-in-action — becomes a reflective practitioner with a persistent mirror. This concept has been extended in the sociomaterial literature to artefacts and technologies. Leonardi (2012) argues that the materiality of artefacts — their specific physical and functional properties — shapes and constrains the knowledge-work arrangements that surround them. Orlikowski (2007) developed the concept of sociomaterial practices, showing that artefacts and human actors are constitutively entangled: the properties of an artefact are not fixed but emerge through use, and that emergence follows path-dependent logic. Building on these foundations, Waern (2026a) introduces artefactual absorptive capacity as a proposed extension: the capacity of a digital twin to absorb knowledge from actor interactions, where each interaction increases the twin's capacity to absorb subsequent interactions. This extension is proposed as a design contribution within the DSR framework (Hevner et al., 2004) and requires independent validation. It is explicitly labelled as proposed theory, not established empirical finding.

This extension provides the theoretical mechanism for SMILE's phase sequencing: phases are not arbitrary stages but rungs on a ladder of accumulating artefactual capacity. A twin that has not completed Phase 1 cannot absorb the domain knowledge deposited in Phase 2. A twin without a Phase 3 ontology cannot generate Phase 4 contextual intelligence. The limitation of single-author validation for this concept is acknowledged as the primary methodological constraint of this paper and is addressed in Section 13.

The mechanism is environmental coupling: each query to the knowledge graph creates a new index path; each decision recorded creates a new case in the graph's precedent library; each sensor reading calibrates the model's baseline. The artefact's capacity grows because its structure physically changes with use — new nodes, new edges, new weights — in the same way that Hutchins's (1995) navigation team's cognitive capacity grows as instruments, charts, and procedures accumulate shared history. The unit of absorptive capacity is not the individual but the sociotechnical system including its artefacts.

The cross-domain application of SMILE is consistent with this mechanism's domain-independence. Whether the twin is a building, a biological system, an animal, or a city district, the same accumulation logic appears to operate. What changes is the content; the accumulation logic appears invariant. Multi-site confirmation of this invariance is required before the claim can be advanced from "consistent with the evidence" to "empirically validated."

3.5 Benefits Management: Grounding Impact-First in Established IS Practice

Ward and Daniel (2006) established benefits management as the discipline for ensuring that IS investments actually deliver the organisational changes and outcomes they are designed to produce. Ashurst et al. (2008) extended this with the concept of benefits realisation capability — the organisational competence required to systematically identify, plan, and deliver benefits from IT-enabled change — demonstrating that organisations with this capability significantly outperform those without. Their Benefits Dependency Network (BDN) maps the chain from enabling changes (what must change for the system to work) through business changes (what organisational practices must change) to business benefits (the measurable improvements that result) to investment objectives (the strategic outcomes that justify the investment).

This pillar is foundational to SMILE's impact-first principle. Ward and Daniel (2006, p. 4) make the critical observation that "benefits from IS and IT investments are not automatic — they have to be identified, planned, actively delivered and then reviewed." This is precisely the failure mode that data-first digital twin implementations exhibit: they assume that technical capability will automatically produce operational benefit, without the deliberate identification, planning, and delivery that benefits management requires.

SMILE's inverted pyramid (Outcome → Action → Insight → Information → Data) is the temporal expression of the Benefits Dependency Network. Every scope element in SMILE Phase 2's MVT Specification must be traceable to an identified benefit in the BDN; every sensor deployment in Phase 3 must be justified by its contribution to an information requirement that serves a documented insight need. The BDN is not a separate planning tool that runs alongside SMILE; it is the logical structure of SMILE's Phase 2 artefacts.

The benefits management pillar also grounds SMILE's twenty benefits (Section 5) in established IS practice. Benefits are not automatic; they must be actively managed. SMILE's phase exit criteria are the operational mechanism for this active management: a phase does not complete until the benefit it was designed to produce is demonstrably accessible to the stakeholder communities who need it.

3.6 Reality as Knowledge Transfer Nexus

Boundary object theory (Section 3.3) establishes that the twin serves multiple communities simultaneously. This section articulates a stronger claim about the twin's epistemic role: reality — the digitally rendered twin of the physical world — is not merely a storage medium for knowledge, but the nexus through which all knowledge transfer happens.

This claim has three dimensions. First, people share knowledge through reality. When two stakeholders disagree about the state of a system, they do not resolve the disagreement by consulting documents or attending meetings; they query the shared twin and orient their negotiation around the physical truth it renders. Knowledge transfer is accelerated because the medium of transfer — the shared reality — is uncontroversial in a way that documents and reports rarely are.

Second, systems share knowledge through reality. In a SMILE deployment, systems do not exchange data directly with each other; they query and update the shared twin. The twin is the integration layer, not a message bus. This has a significant architectural consequence: system-to-system knowledge transfer does not require bilateral integration agreements; it requires each system to have a read/write interface to the twin. The integration complexity scales linearly with the number of systems rather than combinatorially.

Third, AI shares knowledge through reality. A SMILE-compliant AI system (Section 6.2, Crucible) does not reason from its training data; it reasons from the current state of the twin. The AI's outputs are grounded in the same physical reality that all other actors consult. When an AI output is incorrect, the error can be traced to a specific discrepancy between the twin's state and the physical world — a calibration error, a sensor failure, an unmodelled event — rather than to an opaque model parameter. This is what makes SMILE's AI explainable: the ground truth is always visible.

The temporal dimension of this nexus is also distinctive. Each interaction with the twin — human query, system update, AI inference — enriches the model. Time compounds knowledge in a SMILE deployment rather than decaying it. The twin at year five is not the same as the twin at year one; it is vastly more capable, because five years of interactions have been absorbed into its ontological structure. This compounding effect is the mechanism that makes Phase 6 (Perpetual Wisdom) possible: knowledge accumulated across five years of interactions can be transferred to a new deployment, which begins at a higher baseline than any previous deployment.

This framing extends Hutchins's (1995) distributed cognition framework: if cognitive processes are distributed across people, artefacts, and environments, then the digital twin — as the most information-dense artefact in the system — becomes the primary locus of distributed knowledge, the nexus through which all other cognitive agents (human and artificial) coordinate.

This is the key differentiator from every other methodology in the field: SMILE's primary artefact IS the knowledge transfer medium. The Reality Canvas of Phase 1 is not a project deliverable that is archived when the phase closes; it is the substrate that every subsequent interaction enriches, and through which every subsequent interaction happens.

4. Intellectual Lineage: Gothenburg School and Concurrent Engineering

4.1 The Five Model: Information Systems as Sociotechnical Wholes

SMILE inherits from the Gothenburg School of information systems theory — specifically the Five Model developed through the research tradition of Pessi, Magoulas, and Hugoson at the University of Gothenburg (Magoulas & Pessi, 1998; Hugoson et al., 2011). The Five Model posits that any information system — and by extension, any project that creates or transforms information systems — must address five interdependent dimensions simultaneously: Strategy (why the system exists and what it must achieve), Organisation (who does what and how), Processes (what activities create value), Infrastructure (the technical and informational substrate), and Co-workers (the human actors who carry knowledge and whose adoption determines whether the system achieves its purpose).

The Five Model's central insight is that failure in any one dimension propagates to all others. A perfectly designed technical infrastructure deployed to an organisation whose processes have not been redesigned to use it will fail. A strategically sound initiative whose co-workers have not been enrolled will not survive its first operational year. This holistic simultaneity — the insistence that all five dimensions must be addressed in every significant initiative — is the intellectual ancestor of SMILE's four-perspective structure (From People, From Systems, From Planet) and its phase-sequencing logic.

The Five Model consists of hard dimensions — infrastructure, technical architecture — and soft dimensions — culture, competence, organisational readiness. The main addition SMILE makes to the Five Model is context — both operational context and global context — as explicit dimensions that the original framework did not foreground.

But the deeper gap is what the Five Model does not account for: data — specifically where data resides as a hard dimension sitting alongside infrastructure — and knowledge — where knowledge resides in the soft dimension. Both of these could be either in-house or outside the organisation. An organisation whose critical data lives in a vendor's cloud and whose critical knowledge lives in a departing employee's head has a fundamentally different strategic posture than one where both are sovereign — yet the Five Model does not distinguish these cases.

SMILE captures all of this through an as-is situation of where everything is today, so that an organisation, person, or entity can invite to innovate — and thus emulate, then simulate, then operate — knowing that all actions and inactions are recorded in space and time. The combination of this with agent-based modelling and multi-agentic systems means the future becomes understandable (explainable outcomes), parametric designed (tunable what-if scenarios), and generative designed (emergent from simulation, not prescribed). We can better understand outcomes of our actions before, after, or during any decision.

The Five Model is diagnostic — a static snapshot. SMILE is operational — dynamic, temporal, simulatable. This is the core intellectual contribution: turning a framework for understanding into a framework for acting.

SMILE extends the Five Model in four directions. First, it adds the Planet dimension as a first-class axis — the Gothenburg School was primarily concerned with organisational information systems, whereas SMILE operates in environments where the physical world (GIS, satellite data, BIM, CIM, material physics) is itself a participant in the information system. Second, it adds AI as a fourth perspective — recognising that autonomous agents are not tools within Systems but independent actants that transform the networks they participate in. Third, it adds the lifecycle dimension — the Five Model describes a system's structure; SMILE describes its creation and evolution over time. Fourth, it operationalises the Five Model's holism through the four-perspectives lens, which ensures that People, Systems, Planet, and AI considerations are applied at each phase rather than addressed in separate planning streams that risk becoming disconnected.

SMILE's phase logic follows a need-first inverted pyramid (Outcome → Action → Insight → Information → Data), ensuring that the methodology proceeds from identified need toward data requirements rather than in the dominant reverse direction. This need-first sequencing is complemented by the NUDEDA infrastructure-maturity ladder (Section 9.3), which extends Weill and Broadbent's (1998) four infrastructure postures — None, Utility, Dependent, Enabler — with two further stages proposed by the author for the age of AI: Digital DNA and Autonomous.

4.2 Concurrent Engineering Practice in High-Stakes Systems

The concurrent engineering and real-time operations principles that inform SMILE Phases 2 and 4 draw on established engineering practice from major space agency programmes, including the concurrent design facility approach that has become standard in mission-critical systems engineering. These practices are well-documented across the systems engineering literature (Wall & Ledbetter, 1991), and the ITU-R framework for joint communication-and-sensing (ITU-R, 2023) reflects comparable concurrent design principles for telecommunications infrastructure. SMILE draws on these traditions as practitioner-validated methodological inspiration rather than claiming direct institutional lineage.

Concurrent engineering means that stakeholders from all communities — From People, From Systems, From Planet — participate simultaneously in MVT definition and hypothesis testing, using the Reality Canvas as their shared model. No community waits for another's deliverable before beginning its own work; all communities work on the same evolving representation of the system and negotiate trade-offs in real time.

Virtual-first validation means that every hypothesis, procedure, and design decision is tested in simulation before it is executed in the physical world. The cost of testing is measured in simulation time; the cost of failing to test is measured in rework and resource loss. SMILE's virtual-first principle generalises this to any domain: test before you build, simulate before you deploy, validate hypotheses before committing physical resources.

Real-time decision support means that operational stakeholders at any given moment have access to all telemetry, all models, and all anomaly histories through integrated interfaces that surface actionable state without requiring manual data assembly. SMILE Phase 4's connected everything and command-and-control structure is the general-case version.

Institutional memory is a deliberate artefact. Knowledge created in one project context is systematically documented, structured, and made available to subsequent contexts through knowledge bases that persist across team changes, retirements, and organisational restructuring. SMILE Phase 6's Perpetual Wisdom is this principle generalised.

5. The Twenty Benefits: What SMILE Delivers to Any Project

The following twenty benefits are consequences of the phase logic. A project that applies SMILE phases sequentially and completely will acquire each benefit as a structural byproduct of the phase work, not as an additional objective to be pursued separately. This is the core of the universality claim: the benefits are not imposed on the project; they emerge from it.

Of the twenty benefits, five are supported by measurable data from published corpus applications (indicated below); the remaining fifteen are stated as expected benefits based on phase logic, with empirical validation in progress. This distinction is explicit and follows Ward and Daniel's (2006) observation that benefits identification must precede, not follow, system deployment.

The language of the remaining fifteen has been revised from "consequence" to "expected benefit based on phase logic" to reflect the current state of evidence.

5.1 Operational Benefits

(1) Virtual-First Validation. SMILE Phase 2's Minimal Viable Twin exists in simulation before it exists physically. Every hypothesis about how the project will deliver value is tested virtually before physical resources are committed. Supported by evidence: in the genomics startup application (Waern, 2026g), a three-month infrastructure validation phase identified and resolved seven compliance configuration errors before any AI model development was initiated, preventing an estimated three months of rework. This is the most directly measurable benefit in the operational cluster.

(2) Real-Time Situational Awareness. Expected benefit based on phase logic. SMILE Phase 4's connected everything architecture is designed so that any stakeholder with appropriate access can query the current state of the system on their timescale without assembling data manually. Operations see seconds; management sees hours; strategy sees months. All are reading the same twin in different temporal resolutions — the pace layering principle. Empirical validation of the time savings from this design is in progress.

(3) Predictive Maintenance. Expected benefit based on phase logic. SMILE Phase 4's predictive analytics and Phase 5's prescriptive AI are designed to operate on calibrated models of the physical system, projecting forward states and recommending interventions before failure occurs. Predictive maintenance is a consequence of having a physics-calibrated model of what is being managed. Independent empirical validation is required.

(4) Scenario Simulation. Expected benefit based on phase logic. SMILE Phase 2's hypothesis testing and Phase 5's "simulate everything" principle are designed to provide a working simulation environment at each phase. Any project decision that would otherwise require commitment can instead be tested in scenario space. The breadth and quality of scenario coverage as a measurable outcome variable is pending multi-site validation.

5.2 Knowledge Benefits

(5) Institutional Memory. Expected benefit based on phase logic. SMILE's persistent twin architecture means that knowledge is encoded in the system, not carried in people's heads. Phase 6's Phoenix Strategies are the explicit governance mechanism for this benefit. Longitudinal data on knowledge retention rates compared to non-SMILE deployments is a primary research agenda item.

(6) Knowledge Acceleration. Supported by evidence: the cross-domain knowledge transfer from equine metabolic monitoring to human bio-DT (Waern, 2025b; Waern, 2026a) demonstrates that the SQM edge-deployment architecture, once validated in one domain, required substantially less development time in the second domain. The same edge model abstraction pattern transferred directly, with domain-specific parameterisation accounting for the remaining work. This is the most directly observable instance of the knowledge acceleration claim in the current corpus.

(7) Cross-Domain Insight. Expected benefit based on phase logic. SMILE's four-perspective structure is designed to ensure that knowledge from the People, Systems, Planet, and AI dimensions is integrated in the same knowledge graph, making cross-domain queries routine rather than exceptional. The causal link between SMILE adoption and cross-domain insight generation requires controlled-condition measurement.

(8) Explainable Decisions. The Crucible trust layer (Section 6.2) ensures that every decision supported by the twin is traceable to its sources, with confidence intervals and causal chains made explicit. Explainability is not a feature added after the intelligence is built; it is an architectural requirement enforced at Phase 3 (ontology before AI) and maintained through Phase 5 (AI Factory outputs require source attribution). No hallucination-friendly architectures. Supported by design: the Crucible architecture has been operationalised and is a design contribution in its own right.

5.3 Economic Benefits

(9) Resource Optimisation. Expected benefit based on phase logic. Phase 4's real-time awareness and Phase 5's prescriptive AI are designed to identify and eliminate waste. The optimisation is grounded in the physics of the specific system being managed. Domain-specific quantification is pending.

(10) Revenue Visibility. Supported by evidence: the contributor screening application (Waern, 2026h) demonstrated that SMILE Phase 2's explicit outcome chain — from desired contributor outcomes back through required actions to necessary data — produced a live leaderboard that made the connection between task completion and programme viability continuously visible. The funnel-of-intent design reduced unqualified contributor overhead by making the outcome chain visible to applicants before they invested time.

(11) Reduced Waste. Expected benefit based on phase logic. The virtual-first principle, Phase 3's ontology-before-AI sequencing, Phase 6's circular strategies, and cross-industry knowledge transfer are each designed to eliminate specific categories of waste. Aggregate waste reduction quantification requires multi-site longitudinal study.

(12) New Business Models — Twin-as-a-Service. Expected benefit based on phase logic. A Phase 6 SMILE deployment creates a validated, calibrated, structured knowledge asset that has market value beyond its originating context. The equine data operating system architecture and the RF scene calibration methodology are candidate Twin-as-a-Service offerings. Commercial validation of this model is pending.

5.4 Social and Organisational Benefits

(13) Stakeholder Alignment via Boundary Objects. SMILE Phase 1's Reality Canvas is, structurally, a boundary object: it provides the shared physical referent around which stakeholder communities with different vocabularies, priorities, and institutional affiliations can organise. Rather than requiring stakeholders to agree on a common vocabulary, SMILE requires only that they agree on a shared physical referent. Supported by evidence: in the RF/construction domain (Waern, 2026a), a single drone-captured scene served four distinct stakeholder communities simultaneously without requiring any community to adopt another's vocabulary. This is the best-evidenced instance of the boundary object claim in the corpus.

(14) Democratised Access. Expected benefit based on phase logic. Phase 4's design requirement — that non-specialist users from at least two distinct stakeholder communities can query the twin without technical intermediation — is a deliberate democratisation of knowledge access. Empirical measurement of query comprehension rates across technical and non-technical users is pending.

(15) Remote Enablement. Expected benefit based on phase logic. Phase 3's remote enablement activities and Phase 5's edge-native architecture are designed to make full twin capability available to actors not physically co-located with the system being managed. Measurement of remote-access utilisation rates versus co-located access rates is a planned research activity.

(16) Inclusive Governance. Expected benefit based on phase logic. The four-perspective structure ensures that the voices of People, Systems, Planet, and AI are present at every phase as a quality gate, not merely a compliance exercise. Cross-community participation rates in phase reviews are a measurable proxy for this benefit.

5.5 Systemic Benefits

(17) Interoperability by Design. SMILE Phase 3's Minimal Interoperability Mechanisms (MIMs) framework and Phase 4's eight-dimension interoperability model ensure that the twin is designed for interoperability from its first operational day. Interoperability is a design requirement that every project embedding SMILE addresses in Phase 3. Expected benefit based on phase logic: the reduction in post-deployment integration cost attributable to Phase 3 MIMs implementation requires controlled measurement.

(18) Sovereignty — Edge-Native by Design. Expected benefit based on phase logic. SMILE's edge-native architecture (Phase 5) and data sovereignty commitment (Phase 3) are designed to ensure the twin can operate at full capability without continuous cloud dependency. Operational sovereignty in offline conditions is a measurable exit criterion that will be systematically evaluated.

(19) Regulatory Compliance — AI Act, GDPR, MDR. Expected benefit based on phase logic. SMILE's explainability architecture, ontology-before-AI sequencing, governance boundary requirements, and data sovereignty provisions are designed to constitute a compliance architecture for the EU AI Act (Regulation (EU) 2024/1689), GDPR (Regulation (EU) 2016/679), and MDR (Regulation (EU) 2017/745). The AI Act's risk-based classification framework and its requirements for high-risk AI systems — including transparency, human oversight, and technical documentation — map structurally to SMILE's Crucible layer, governance boundaries, and phase documentation. Legal mapping of SMILE deliverables to specific regulatory articles is a planned research output.

(20) Planetary Impact — Circular and ESG. Expected benefit based on phase logic. Phase 6's circular strategies and Phase 1's PESTELED analysis are designed to make planetary-scale environmental context a first-class input and end-of-life knowledge recycling a first-class output (Geissdoerfer et al., 2017; Kirchherr et al., 2017). ESG quantification frameworks for SMILE deployments are pending development.

6. The AI Journey and the Crucible Trust Layer

6.1 The AI Journey Within SMILE

SMILE's AI journey is a five-stage progression that follows the same impact-first logic as the six phases. The stages are not milestones to be checked off; they are structural states of the twin's relationship with AI, each of which enables the next.

Data Contextualisation (aligned with Phases 1–2): Raw data is contextualised — placed in its spatial, temporal, and organisational context. Data without context is not a foundation for intelligence; it is a source of hallucination risk. The Reality Canvas and the MVT specification create the context within which data acquires meaning.

AI-Ready (aligned with Phase 3): Data is structured, tagged, and accessible for AI consumption. Ontologies and metadata are in place. The knowledge graph provides the semantic substrate on which AI reasoning will operate. A system that declares itself "AI-ready" without completing Phase 3 is not AI-ready; it has data and is hoping that AI will supply the context.

AI-Infused (aligned with Phase 4): AI augments human decision-making with predictions, recommendations, and pattern recognition. The AI outputs are explainable — traceable to their sources in the knowledge graph, accompanied by confidence intervals. Human experts can override AI outputs; the system records and learns from overrides. The override mechanism is not a failure mode; it is a knowledge input.

AI-Ingrained (aligned with Phase 5): AI is embedded in operations — autonomous decisions within governance-defined boundaries, continuous learning from outcomes. The AI system is not a separate analytical tool; it is part of the operational fabric. This is only safe when Phases 3 and 4 have been completed, because Phase 3 provides explainability infrastructure and Phase 4 has established the human oversight and governance boundaries.

Explainable AI Decision Making (aligned with Phases 5–6): Full autonomous decision-making that is explainable, auditable, and anchored in physical reality. Every decision can be traced back to its data sources, its model assumptions, and the causal chain that led from evidence to recommendation. Explainability is not a property that LLMs naturally provide; it is a structural achievement that requires the ontological and governance foundations of all preceding phases.

The LQM distinction. For systems governed by physical laws — buildings, biological bodies, urban infrastructure, equine physiology — the appropriate AI regime is not statistical pattern matching over historical data (Large Language Models) but physics-based simulation from first principles (Large Quantitative Models, LQMs). This distinction is grounded in the physics-informed machine learning programme (Raissi et al., 2019; Karniadakis et al., 2021), which demonstrates that neural networks constrained by physical laws produce more interpretable, data-efficient, and trustworthy outputs than purely data-driven approaches. Recent work on physics-informed neural networks as digital twin substrates (Kapusuzoglu & Mahadevan, 2024) confirms this trajectory. LQMs compute from causal mechanisms and produce interpretable outputs with explicit uncertainty intervals. For edge deployment, LQMs can be reduced to Small Quantitative Models (SQMs) — compact, validated physics models that run on local hardware with millisecond latency. LLMs retain a legitimate role as communication and coordination layers: translating model outputs into human language, accepting natural-language queries, synthesising narratives. They must not be asked to simulate physics (Waern, 2026c).

6.2 The Crucible Trust Layer

Every SMILE output passes through Crucible before it becomes a decision input. Crucible is the verification and trust layer — the architectural commitment to retrieval over generation.

The core principle: when accuracy matters, retrieve rather than generate. This aligns with the retrieval-augmented generation paradigm (Lewis et al., 2020) but goes further: where RAG augments generation with retrieval, Crucible subordinates generation to retrieval. Generation is the communication layer; retrieval is the truth layer. Miller (2019) demonstrated that human explanations are contrastive, selective, and social — Crucible's design enforces all three properties. Gunning and Aha (2019) established the DARPA XAI programme's requirement that AI systems produce explanations that enable users to understand, appropriately trust, and effectively manage AI — Crucible operationalises this requirement architecturally. A SMILE twin that knows the current temperature of a specific room does not generate a plausible estimate; it retrieves the measured value. A twin that knows the patient's last laboratory result does not interpolate from the patient's demographic profile; it retrieves the documented value from the provenance-signed record. A twin that knows the material classification of a specific wall in a reconstructed building does not guess from the building's age; it retrieves the value from the calibrated photogrammetric survey.

Crucible operates at three levels. At the data level, it enforces provenance: every data point in the knowledge graph carries a source, a timestamp, a confidence estimate, and a validation status. At the inference level, it enforces source attribution: every AI output must be traceable to the specific knowledge graph entries that produced it. At the decision level, it enforces explainability: every decision support recommendation includes a natural-language explanation of the causal chain, with links to source data, that a domain expert can verify or override.

Explainability as a cross-cutting requirement. Every AI output in SMILE must be explainable — not merely traceable, but comprehensible to the domain expert who must act on it. The user must understand not only WHAT the system recommends but WHY: the causal path from data to inference to recommendation must be legible without technical intermediation. Arrieta et al. (2020) identify the key XAI properties required for high-stakes domains: transparency, interpretability, and explainability in the strictest sense — the ability of a human to confirm or contest the system's reasoning. Crucible enforces all three. An AI output that cannot be explained in domain language to a non-technical stakeholder has not passed the Crucible gate, regardless of its statistical accuracy.

Crucible is what makes SMILE safe for high-stakes domains — clinical care, structural engineering, autonomous operations, pharmaceutical development. In these domains, the cost of a hallucinated answer is not a conversational error; it is a clinical adverse event, a structural failure, or a regulatory violation. The Crucible layer eliminates the class of AI risk that arises from confusing fluent generation with grounded knowledge.

The SMILE Evaluation Ecosystem: Crucible, GLASS, and PRISM. SMILE's methodology has been operationalised through three complementary evaluation tools, each applying SMILE principles to a different assessment context.

GLASS (Grant/proposal readiness scoring) provides a seven-level maturity assessment for project readiness: Discussed → Designed → Coded → Tested → Verified → Deployed → Live. GLASS levels map structurally to NUDEDA: GLASS 0 (Discussed) corresponds to NUDEDA-None; GLASS 3 (Tested) to NUDEDA-Enabling; GLASS 6+ to NUDEDA-Autonomous. GLASS operationalises AEST's Emulate function: it answers "where are we NOW?" with a verifiable, non-negotiable score.

PRISM (Persona-based Role-Isolated Security Matrix) applies SMILE's four-perspective lens to system security. Ten personas — each representing a different actor in the network (administrator, clinician, patient, AI agent, external partner) — are tested against every API endpoint to verify that role boundaries are enforced. PRISM operationalises AEST's Simulate function: it tests "what happens when an actor attempts to exceed its boundaries?" across all actor-network combinations.

Crucible is the proposal analyser that scores Horizon Europe proposals against SMILE's six phases using a radar chart, complemented by four-perspective analysis (People/Systems/Planet/AI), AEST temporal scoring (Absorb/Emulate/Simulate/Transcend), and ANT reality-as-actor assessment. Each phase is scored on alignment with SMILE's entry criteria, phase activities, and exit criteria. This tool demonstrates that the methodology can be automated and applied by AI systems operating under Crucible's own explainability constraints: every score is accompanied by a citation to the specific section of the proposal and the specific SMILE criterion it addresses. The existence of this tool is evidence that SMILE's structure is sufficiently well-defined to be computationally operationalised, not merely described.

7. The SMILE Framework: Six Phases, Four Perspectives

SMILE — Sustainable Methodology for Impact Lifecycle Enablement — is a phase-sequential methodology. Its core principle is the inverted pyramid: rather than beginning data collection and hoping useful outcomes will emerge, SMILE requires that desired outcomes be defined first, and that every subsequent decision be justified by a traceable chain from data through information through insight through action to outcome.

The inverted pyramid sequence: Outcome → Action → Insight → Information → Data.

This inversion has three operational consequences: (1) data collection is always justified, never assumed; (2) the scope of the twin is defined by stakeholder goals, not available data; (3) the twin is evaluated by whether outcomes are achieved, not by data richness or technical sophistication.

SMILE's six phases are concentric, not linear. Each phase adds a layer of capability to the twin rather than replacing the previous layer. A Phase 5 twin is also still a Phase 1, 2, 3, and 4 twin — all capabilities are present and active simultaneously, with later phases adding generative and autonomous capabilities that earlier phases cannot support.

Each phase is experienced from four perspectives simultaneously:

• From People: Organisation-agnostic stakeholder engagement, concurrent engineering, collective intelligence, human-machine interaction design, skills assessment. People are actors in sociotechnological ecosystems who continuously transfer knowledge through the twin.
• From Systems: Standards identification (ISO, IFC, CityGML, FHIR, SAREF, AAS), ontology alignment, metadata mastery, data fabric architecture, BIM/CIM/GIS integration, interoperability compliance (MIMs).
• From Planet: GIS data, City Information Modelling, Building Information Modelling, satellite data, reality canvas creation, spatial-temporal context establishment, environmental sustainability, circular strategies, ESG alignment.
• From AI: Autonomous agent capabilities, LQM/SQM model selection, explainability requirements (Arrieta et al., 2020), AI Act compliance, Crucible verification, agent-to-agent knowledge transfer, AEST-driven temporal reasoning, edge-native inference, trust boundary definition. AI is not a tool within Systems — it is a perspective in its own right, because AI agents are actants (Callon, 1986) that transform the network they participate in.

7.1 Phase 1: Reality Emulation — Define the Box Before Thinking Outside It

Core question: In order to think outside the box, one must first define the box. What is the starting point and boundary of the sociotechnological ecosystem?

Entry criteria: At least one stakeholder has articulated a desired outcome. A spatial scope has been identified. Initial PESTELED analysis has been conducted.

SPIN Questions (Rackham, 1988): Phase 1 is entered through Situation questions — the questions that establish ground truth before any problem framing begins. Where are we? When are we? What exists here? Who is involved and what do they care about? What does the physical environment look like right now? These questions produce the Reality Canvas. Without answered Situation questions, there is no shared referent and no meaningful Phase 1 deliverable. SPIN (Rackham, N. (1988). SPIN Selling. McGraw-Hill) is the questioning methodology that drives entry into each SMILE phase; each phase is unlocked by a different class of question.

Key activities:

• Job-to-be-done analysis: what problem are stakeholders actually trying to solve, for whom, and under what constraints?
• Stakeholder mapping: who needs to be involved, what are their interests, and what would make them enrol?
• Reality Canvas creation: a spatially grounded, temporally anchored 3D+ representation of the shared physical referent — the scene that all subsequent activity will describe and enrich.
• Operating context definition: the PESTELED analysis (Political, Economic, Social, Technological, Environmental, Legal, Ethical, Digital) that establishes boundary conditions.
• Virtual First Mindset adoption: all stakeholders commit to validating hypotheses in simulation before committing physical resources.
• 5 Whys analysis: root-cause decomposition to verify the stated problem is the actual problem.

Primary artefacts: Reality Canvas (3D+ spatial-temporal foundation), Stakeholder Map, Operating Context Document, Ecosystem Boundary Definition.

Exit criteria: The Reality Canvas is navigable and understandable by non-technical stakeholders from all communities who will eventually contribute knowledge to the twin. At least two distinct stakeholder communities have confirmed it reflects their understanding of the shared physical referent.

Worked example — healthcare: A hospital consortium seeking to reduce adverse events during cross-border patient transitions defines the Reality Canvas as the patient's health journey: a temporal map of care transitions, institutional boundaries, and health data handover points. The physical referent is the patient's body across time and jurisdictions (Waern, 2026e, DOI pending — design application pending empirical validation).

Worked example — equine: An Arabian stud farm defines the Reality Canvas as the farm's spatial layout — paddocks, stables, training arenas, foaling boxes — with each horse registered as a canonical entity within that spatial context. Stakeholders include veterinarians, stable staff, the farm manager, and regulatory authorities (Waern, 2026b, DOI pending — design application pending empirical validation).

Worked example — smart city: A municipal government seeking autonomous urban operations defines the Reality Canvas as the city's operational data fabric: sensor networks, building portfolios, transportation systems, and citizen service delivery — with citizen trust and democratic accountability as first-class design constraints (Batty, 2013; Kitchin, 2014; Waern, 2026f, DOI pending — design application pending empirical validation).

Worked example — RF telecommunications: A telecommunications research programme defines the Reality Canvas as a drone-captured 3D scene of a specific urban district. The scene itself — not the RF propagation model — is identified as the primary deliverable: the shared spatial referent that serves RF engineers, structural engineers, urban planners, and humanitarian logistics planners simultaneously. This application has a published DOI and constitutes the most mature cross-community validation of Phase 1 in the corpus (Waern, 2026a, DOI: 10.5281/zenodo.19587944).

7.2 Phase 2: Concurrent Engineering — The Minimal Viable Twin

Core question: What does the Minimal Viable Twin look like, and how do we validate it virtually before investing in physical implementation?

Entry criteria: Reality Canvas exists and has been confirmed by all primary stakeholder communities. Hypotheses about desired outcomes have been explicitly articulated.

SPIN Questions: Phase 2 is entered through Problem questions — the questions that expose the gap between as-is and to-be. What is broken? What is missing? Where does the current system fail? What is the gap between where we are and where we need to be? These questions surface the hypotheses that Phase 2 must validate. A Phase 2 without answered Problem questions produces a Minimal Viable Twin that solves no identified problem.

Key activities:

• Define the Minimal Viable Twin (MVT): the smallest twin that delivers value — not the most comprehensive imaginable.
• As-is / To-be scope definition: explicit description of current and target state, justifying every scope element by its contribution to the outcome chain.
• Benefits Dependency Network (Ward & Daniel, 2006): map from desired benefits back through required business changes and enabling changes to investment scope.
• Virtual-first validation: all hypotheses tested in simulation before physical deployment.
• Scenario planning: tabletop exercises using the Reality Canvas to explore alternative futures and stress-test assumptions.
• Hypothesis register: every assumption about how the twin will create value is recorded as a falsifiable hypothesis with a specified validation method.
• Stakeholder innovation sessions: concurrent engineering with cross-community participants to identify scope elements that serve multiple communities simultaneously.

Primary artefacts: MVT Specification, As-is / To-be Scope Document, Benefits Dependency Network, Validated Hypothesis Register, Virtual First Completion Plan.

Exit criteria: The MVT is defined at a level of specificity sufficient for two independent engineers to implement it the same way. Every scope element is justified by a hypothesis that references the outcome chain. At least one hypothesis has been tested virtually before physical implementation begins.

Worked example — biotech / genomics: A genomics startup defines the MVT as a cloud-provisioned, compliance-configured genomic data infrastructure capable of processing one batch of samples end-to-end with documented QC metrics, federated access controls, and cryptographic provenance. This is validated against SMILE criteria before any AI model development begins. The three-month infrastructure phase is not a delay to intelligence; it is the MVT (Waern, 2026g, DOI pending — design application pending empirical validation).

Worked example — contributor screening: A contributor programme defines the MVT as a live leaderboard with three tiered task levels (L1 setup, L2 configuration, L3 building and auditing), visible to all participants, tracking real pull requests against a live codebase. The hypothesis that 114 applicants will self-sort into actionable contributors through progressive task friction is tested against the live leaderboard before the programme scales (Waern, 2026h, DOI pending — design application pending empirical validation).

7.3 Phase 3: Collective Intelligence — The Ontology Factory and MIMs

Core question: How do we create ontology factories as the foundation for AI factories, connecting fragmented realities into a knowledge graph?

Entry criteria: MVT is deployed. Physical sensors (or equivalent data streams) are operational. Initial KPI baseline exists.

SPIN Questions: Phase 3 is entered through Implication questions — the questions that make the cost of inaction visceral. What happens if the gap identified in Phase 2 is not fixed? What cascades downstream? Who is harmed and how? What systemic failures become inevitable? Implication questions are what motivate the investment in ontology factories and knowledge graph infrastructure. Without them, Phase 3 feels like bureaucracy; with them, it is obviously necessary.

Key activities:

• Physical sensor deployment and calibration against published standards.
• Ontology creation: structuring the classification relationships between entities in the shared reality. An ontology, in Gruber's (1993) foundational definition, is an explicit specification of a conceptualisation — and in the digital twin context, that specification must be both machine-readable and domain-expert-comprehensible. Recent work by Zheng et al. (2024) and Lu et al. (2024) demonstrates that knowledge graphs grounded in formal ontologies are the enabling infrastructure for intelligent digital twins, providing the semantic substrate that makes cross-domain reasoning possible.
• Alignment with industry standards: ISO, IFC, CityGML, FHIR, SAREF, SNOMED, WHO-QoL, or domain-equivalent classification systems.
• KPI confirmation: measurable outcome indicators validated against the baseline.
• Security and data strategy: the data governance architecture is specified before the first external actor contributes knowledge.
• Knowledge graph construction: entities, relationships, and causal pathways encoded in a graph structure that supports both semantic query and AI reasoning.
• Remote enablement: full twin capability accessible to actors not physically co-located with the system.

Minimal Interoperability Mechanisms (MIMs). Phase 3 is where MIMs are implemented as the context management framework for interoperability. MIMs, developed by the Open and Agile Smart Cities initiative (OASC, 2023), define the minimum mechanisms necessary for any two systems to exchange meaningful information: a common context model, an API framework, and a data sovereignty mechanism. Within SMILE, MIMs operationalise the principle that interoperability is not an integration project added after the twin is built; it is a Phase 3 architectural requirement. The three core SMILE MIM layers are: (a) Contextual — the shared ontology and knowledge graph schema; (b) Technical — the API and transport standards; (c) Sovereign — the data governance and access control framework. Every Phase 3 deployment that does not implement all three MIM layers has not completed Phase 3, regardless of how sophisticated its sensor network is.

INSPIRE as a spatial taxonomy precedent. The EU INSPIRE Directive (2007/2/EC) provides the most concrete large-scale precedent for the spatial taxonomy and ontology standardisation work that SMILE Phase 3 operationalises. INSPIRE mandates interoperable spatial data infrastructure across all EU member states — common schemas, shared ontologies, harmonised metadata — so that spatial data produced in one jurisdiction is queryable by systems in another without bilateral integration agreements. This is precisely the MIMs principle applied to geography. SMILE extends the INSPIRE logic beyond spatial data to any domain's data fabric: the same argument that mandates shared ontology for geographic features mandates shared ontology for health records, building elements, equine biosignals, or RF propagation models. This demonstrates why taxonomy and ontology factories — backed by vector databases for semantic search and graph databases for relationship mapping — are a foundational architectural play, not an optional enrichment layer. A domain without a shared ontology has the same interoperability problem that pre-INSPIRE EU member states had with spatial data: every jurisdiction speaks a different language, and integration requires bespoke translation at every boundary.

The ontology-before-AI principle: The sequence Ontology Factory → Knowledge Graph → AI Factory is mandatory, not optional. An AI system operating without a structured knowledge graph reasons on unstructured data and produces unexplainable results. SMILE Phase 3 enforces the creation of the semantic foundation before Phase 5 activates AI inference. This is not bureaucratic sequencing; it is the epistemological prerequisite for explainability.

Primary artefacts: Operational MVT with physical sensors, Ontology Factory (knowledge graph schema), MIMs Implementation Report, KPI Confirmation Report, Data Fabric Integration Document.

Exit criteria: The knowledge graph contains structured representations of entities across at least two distinct stakeholder communities' domains. The ontology has been reviewed by representatives of each community. At least one cross-domain query returns a meaningful answer. MIMs are operational across all three layers.

Worked example — buildings and energy: A municipal building portfolio deploys temperature and occupancy sensors calibrated against published standards. The ontology maps relationships between building entities, energy supply nodes, occupancy zones, and HVAC control points according to Haystack tagging and the Brick ontology. Cross-domain query: "Which rooms are consistently over-heated relative to their occupancy schedule?" — impossible without the ontology (Waern, 2025a, DOI: 10.5281/zenodo.17462962).

7.4 Phase 4: Contextual Intelligence — The Queryable Twin

Core question: How do we enable real-time contextual awareness across the entire ecosystem — understanding not just what is happening, but why?

Entry criteria: Collective intelligence is accumulating. Reality Programmable Interfaces (RPIs) are available. Non-specialist users have a defined access path to the twin.

SPIN Questions: Phase 4 is entered through Need-payoff questions — the questions that make the value of a solution concrete and motivating. What would it mean if we could see this in real time? What decisions would we make differently if we had this answer immediately? What would change for the stakeholder if the twin answered their question without a technical intermediary? Need-payoff questions create the pull for Phase 4's queryable twin architecture.

Key activities:

• Connected Everything: command and control interfaces that surface real-time twin state to all stakeholder communities.
• Real-time decision support: the twin answers operational questions on stakeholder timescales without requiring technical intermediaries.
• Predictive analytics: the twin projects forward states based on calibrated models and scenario parameters.
• Root cause analysis: the twin traces observed anomalies backward through causal chains to probable origins.
• Pace Layering alignment: different actor communities interact with the twin at different temporal resolutions — operations at seconds, management at hours, strategy at months — without creating conflicts.
• Interoperability alignment: all eight interoperability dimensions (contextual, legal, human, security, technical, syntactical, semantic, organisational) are operational. These dimensions extend the European Interoperability Framework (European Commission, 2017), which establishes four layers (legal, organisational, semantic, technical) as the baseline for cross-border digital public services; SMILE adds contextual, human, security, and syntactical dimensions to address the richer requirements of physical-digital twin systems.

Primary artefacts: Operational Digital Twin (as-is), Real-time Dashboard, Predictive Maintenance System, Connected Ecosystem.

Exit criteria: Non-specialist users from at least two distinct stakeholder communities can query the twin and receive actionable responses without technical intermediation. At least one operational decision has been made and documented on the basis of twin outputs.

Worked example — personal health: A personal digital twin integrates continuous wearable data, calendar and social context, and air quality environment. The twin answers: "Given that you are travelling to a high-altitude city next week, your metabolic model predicts increased insulin resistance during acclimatisation; here is a protocol to minimise the impact across both domains." This cross-domain contextual response is impossible without Phase 3's ontological integration (Waern, 2026c, DOI pending — design application pending empirical validation).

7.5 Phase 5: Continuous Intelligence — The AI Factory

Core question: How do we leverage explainable AI factories for exponential scale, scope, and learning — moving from reactive to prescriptive?

Entry criteria: Contextual intelligence is operational. Governance boundaries for autonomous learning are defined and agreed by all stakeholder communities. Anomaly detection baselines are established.

SPIN Questions: Phase 5 loops back to Situation questions — but now at higher resolution. Where are we NOW? Has reality changed since Phase 1? Are our current baselines still valid? What has the twin learned that shifts the original situation description? The Situation loop distinguishes SMILE from linear methodologies: Phase 5 does not assume the Phase 1 canvas is still accurate; it verifies it continuously and updates accordingly.

Key activities:

• Prescriptive maintenance: the twin prescribes interventions before failures occur.
• AI-driven prognostics: algorithms trained on the knowledge graph identify patterns that precede adverse events.
• Universal event pipeline: all relevant events from physical reality are ingested, processed, and acted upon within defined latency bounds.
• Black Swan identification: scenario simulation explores tail-risk events that historical data does not represent.
• Simulate Everything: the twin's models are used proactively to explore possibility spaces, not reactively to explain observed states.
• Distributed intelligence: edge-native processing ensures the twin operates at full capability without continuous cloud connectivity. Federated learning (McMahan et al., 2017; Kairouz et al., 2021) provides the mechanism for distributed model improvement without centralising raw data — each edge twin learns locally and contributes gradient updates to the collective knowledge.
• Multi-agent orchestration: Phase 5 deploys autonomous agents that perceive, plan, act, and reflect within governance boundaries. Park et al. (2023) demonstrated that generative agents with memory, planning, and reflection capabilities exhibit emergent social behaviours; SMILE constrains these capabilities within the Crucible trust layer and governance envelopes defined in Phase 4. Wooldridge (2009) provides the theoretical grounding for agent autonomy, reactivity, pro-activeness, and social ability that SMILE's agent architecture implements.
• Virtual sensors: the twin interpolates unmeasured states from calibrated models, filling sensor gaps with model-grounded estimates.

Primary artefacts: AI Factory (trained algorithm suite), Simulation Digital Twin (to-be scenarios), Event Pipeline, Distributed Intelligence Architecture.

Exit criteria: The system identifies anomalies and generates actionable alerts without human initiation, within governance-approved boundaries. At least one prescriptive intervention has been recommended and acted upon with documented outcome.

The LQM distinction: For systems governed by physical laws, the appropriate intelligence regime is Large Quantitative Models (LQMs) — physics-based simulation from first principles — not statistical pattern matching over historical data. LQMs produce interpretable outputs with explicit uncertainty intervals. For edge deployment, LQMs reduce to Small Quantitative Models (SQMs) that run on local hardware. LLMs serve as communication interfaces. They must not simulate physics (Waern, 2026c).

7.6 Phase 6: Perpetual Wisdom — Knowledge That Outlives Its Origin and Points Toward Holographic Society

Core question: How do we ensure that knowledge created today is transferable, reusable, and contributes to planetary-scale impact?

Entry criteria: Multiple twins are operational. Cross-twin knowledge transfer is architecturally possible. Governance mechanisms for knowledge attribution and provenance are in place.

SPIN Questions: Phase 6 is entered through transcendent Need-payoff questions — the questions that reach beyond the originating context. What did we learn here that applies everywhere? What would it mean if this knowledge was available to the next deployment from day one? What is the planetary-scale value of the wisdom accumulated in this twin? These questions are the entry condition for Perpetual Wisdom: without them, Phase 6 is a knowledge archive; with them, it is civilisational infrastructure.

Key activities:

• Ecosystem enablement: knowledge generated in one twin demonstrably improves outcomes in another. Senge's (1990) learning organisation — one that continuously expands its capacity to create its future — becomes, in Phase 6, a learning ecosystem where the organisational boundary is dissolved and knowledge flows between twins, domains, and geographies.
• Cross-industry knowledge transfer: templates, calibrated model parameters, and ontology extensions shared across domain communities.
• Circular strategies: materials, knowledge, and models designed to be recyclable at end-of-life rather than discarded.
• Phoenix Strategies: when a twin is retired, its accumulated knowledge is transferred to successor systems in structured, documented form.
• Open-source contribution: core methodology, ontology schemas, and model libraries published for community use.
• Supply chain optimisation: knowledge from operational twins informs procurement, design, and commissioning of future physical assets.

Ciborra (2000) observed that corporate information infrastructures do not evolve through top-down planning but through bricolage — improvisation, drift, and opportunistic recombination of available resources. Phase 6 is the SMILE acknowledgement of this: rather than prescribing what the evolved system looks like, Perpetual Wisdom governs how accumulated knowledge survives the drift and is carried forward. The methodology does not fight infrastructure evolution; it ensures the wisdom within it outlasts any particular configuration.

Primary artefacts: Perpetual Knowledge Base, Open-Source Contributions, Cross-Industry Templates, Circular Economy Integration Document.

Exit criteria: At least one knowledge artefact generated in this deployment has been demonstrably used to accelerate a subsequent deployment in a different domain, with documented attribution and provenance.

The Hi-Fi Holographic Society endpoint. Phase 6 anticipates the endpoint toward which the SMILE trajectory points: the Hi-Fi Holographic Society — a concrete architectural vision, not a science fiction aspiration. The holographic society is defined by six convergent capabilities:

1. Ambient computing — intelligence embedded in every surface, object, and environment; computation is invisible and ubiquitous
2. Extended reality (XR) — spatial interfaces that replace screens; immersive collaboration where presence and remote participation are indistinguishable
3. Secure edge — all sensitive computation happens locally; edge-native by default, cloud by exception; data sovereignty as a hardware property, not a policy checkbox
4. Differential privacy pipelines — mathematical guarantees that individual data cannot be reverse-engineered from aggregate insights; privacy as a protocol, not a promise
5. Outcome-based knowledge transfer — knowledge flows between people, systems, and AI through shared reality representations (twins), not through documents or meetings; the twin IS the medium of knowledge transfer
6. 6G and beyond — terahertz sensing, joint communication-and-sensing (JCAS), sub-millisecond latency enabling real-time holographic collaboration. The Tataria et al. (2021) 6G systems survey documents the vision and the primary technical constraint: terahertz propagation is severely limited by atmospheric absorption, restricting practical JCAS deployment distances in current implementations. The ITU-R IMT-2030 framework (ITU-R, 2023) establishes the standardisation trajectory for these capabilities. Phase 6's holographic society claims are thus temporally qualified: they describe the endpoint of a 10–15 year standardisation trajectory, not near-term operational capability.

Sensory Fidelity Scale. Web5/WebX environments and the holographic society are not separate technologies; they are increasing fidelity of the same underlying fabric. SMILE operates at every fidelity level. Higher fidelity enables deeper emulation and more accurate simulation, but the methodology is invariant across fidelity.

The fidelity scale clarifies what "higher resolution" means in practice: each step up engages more of the human sensorium, creating a richer channel for knowledge transfer between the physical world and its digital representation. A text report absorbs the past; a full-sensory environment emulates the present with sufficient fidelity to simulate the future with confidence. The Phase enabled column reflects the minimum fidelity at which each phase's exit criteria can be fully satisfied — though lower-fidelity representations remain valid inputs at higher phases.

When any interaction with systems or people can be replicated with full fidelity, when representation can take any form — spatial, temporal, embodied, simulated — the governance question becomes the central question. SMILE Phase 6, in its circular strategies, Phoenix architectures, and open ecosystem enablement, provides the governance framework for this transition: not to foreclose it, but to ensure that the knowledge accumulated in all preceding phases survives into whatever form the connected world takes next.

The Temporal Architecture: AEST

Beneath the six phases and three perspectives lies a temporal architecture that distinguishes SMILE from every existing methodology. The four temporal operations are:

• Absorb the past — capture institutional memory, ingest historical data, learn from what happened. Every project inherits a past; SMILE makes that inheritance explicit and structured.
• Emulate the present — create a living representation of reality as it IS, in real time. The digital twin of now. Not a snapshot but a continuously calibrated mirror.
• Simulate the future — explore what-if scenarios, predict outcomes, test interventions virtually before committing resources physically. This is where virtual-first delivers its cost advantage.
• Transcend the now — go beyond current constraints, generate insights that no single perspective could produce, evolve the system beyond its original design parameters. This is the Perpetual Wisdom phase operating at full capacity.

AEST (Absorb, Emulate, Simulate, Transcend) maps onto the SMILE phases:

Every methodology manages time linearly: plan, execute, review. SMILE operates across time simultaneously. The twin absorbs the past, emulates the present, simulates the future, and in doing so transcends the limitations of the now.

The Web Evolution Through SMILE

SMILE is designed to operate at whatever web generation the project inhabits, because it is anchored to reality, not to infrastructure:

The methodology survives paradigm shifts because it is anchored to reality itself, not to the technological substrate through which reality is mediated. Web1–Web4 characterisations are grounded in documented standards evolution; the post-2030 row represents speculative extrapolation based on current research trajectories and should be treated accordingly.

The holographic society is not a science fiction destination; it is the logical endpoint of applying Spatial Twinning at civilisational scale when the six convergent capabilities described above are realised. SMILE is the methodology that governs the journey. The destination is a world in which the question "what does reality look like?" can be answered from anywhere, for anyone, in any language, at any scale — because the twins have accumulated enough wisdom to show it.

Worked example — equine to human health: The validated SQM architecture developed for equine metabolic monitoring (Waern, 2025b, DOI: 10.5281/zenodo.17464804) transfers its edge-deployment architecture and model reduction methodology directly to human metabolic monitoring (Waern, 2026c, DOI pending). The same pattern — LQM cloud parameterisation, SQM on-device inference, LLM communication interface — operates on equine and human physiology because the underlying physics (thermodynamics, glucose homeostasis, cardiovascular dynamics) shares mathematical structure. Cross-domain knowledge transfer at the architectural level.

8. Cross-Domain Validation and DTC CPT Mapping

8.1 Validation Status and Epistemic Honesty

The cross-domain evidence in this section is drawn from the author's corpus of 39 publications. The limitation of single-author validation is acknowledged as the primary methodological constraint of this paper and is addressed fully in Section 13.

Of the eight domains, one has a published, peer-accessible DOI (RF/construction, Waern, 2026a, DOI: 10.5281/zenodo.19587944), two have published DOIs with substantial accumulated evidence across the broader corpus (buildings and energy, Waern, 2025a, DOI: 10.5281/zenodo.17462962; equine, Waern, 2025b, DOI: 10.5281/zenodo.17464804), and five represent design applications pending empirical validation. This distinction is explicit in the table below.

8.2 Domain Mapping Table

8.3 Mapping to the DTC Capabilities Periodic Table

The Digital Twin Consortium's Capabilities Periodic Table (CPT v1.1; van Schalkwyk, Whiteley, & Goldman, 2024) provides an independent, industry-standard framework for characterising digital twin capabilities. The CPT organises capabilities into functional groups: Sensing and Data Acquisition, Modelling and Simulation, Integration and Interoperability, Analytics and Intelligence, Visualisation and Interaction, Management and Governance, and Security and Trust.

SMILE phases map to CPT capability groups as follows:

This mapping positions SMILE as the methodology that operationalises the CPT's capability model: the CPT defines what capabilities a digital twin should have; SMILE defines the sequenced methodology through which those capabilities are acquired, activated, and sustained. The correspondence is not incidental — SMILE's phase structure was independently derived from applied domain work, and its alignment with the CPT represents framework correspondence with an independent, industry-developed capability taxonomy.

The DTC CPT is explicitly referenced as external validation. This does not constitute full independent empirical validation of SMILE's efficacy claims; it constitutes framework-level alignment between a methodology derived from practitioner experience and a capability taxonomy derived from industry consensus. Both forms of validation are needed; CPT alignment is the more tractable of the two and is reported here.

8.4 Key Cross-Domain Observations

Three patterns emerge that are consistent with SMILE's universality claims and warrant further empirical investigation.

Pattern 1: Phase content changes; phase logic does not. Across all eight domains, the logical relationship between phases is identical: Phase 1 produces a shared referent; Phase 2 defines the minimal sufficient twin; Phase 3 creates the ontological infrastructure; Phase 4 makes the twin queryable by non-specialists; Phase 5 activates autonomous learning; Phase 6 transfers knowledge beyond the originating context. What changes between a healthcare deployment and an equine deployment is the content of each phase. The sequencing logic appears invariant; multi-site confirmation is required.

Pattern 2: The persistent layer is always physical reality. In every domain, SMILE Phase 1 identifies a physical reality as the canonical referent around which all other data orbits. The identity of the persistent layer differs; the principle that one must exist does not.

Pattern 3: Ontology before AI. In every domain, Phase 3 precedes Phase 5, and the quality of Phase 5 AI outputs is directly dependent on the quality of Phase 3 ontological structure. An AI system without a structured knowledge graph produces unexplainable outputs. The ontology is not a bureaucratic deliverable; it is the epistemological prerequisite for intelligence.

9. The SMILE Maturity Model

Maturity models are established instruments in IS research for assessing organisational capability along defined dimensions (Becker et al., 2009; de Bruin et al., 2005). Barricelli et al. (2019) provide a comprehensive survey of digital twin maturity dimensions; Zheng et al. (2024) extend this to knowledge-graph-enabled digital twins. SMILE's phase sequencing defines what must be done; the maturity model defines how well it has been done. Five maturity levels are defined per phase, creating a 6×5 evaluation matrix.

9.1 Maturity Levels

Level 1 — Foundation: Phase activities have been initiated but not completed. Deliverables exist in draft form. Stakeholder community has been identified but not fully enrolled. The twin is present but not functioning as an obligatory passage point.

Level 2 — System of Record: Phase artefacts are complete and accepted by all stakeholder communities. The twin is the single authoritative source for its defined scope. At least one operational decision has been made on the basis of twin outputs.

Level 3 — Augmented Intelligence: The twin's outputs are augmented by AI assistance — pattern recognition, anomaly detection, predictive projection — with all AI outputs traceable to their sources and presented with confidence intervals. Human experts can override AI outputs; the system records and learns from overrides.

Level 4 — Networked Value: The twin's value extends beyond the originating stakeholder community. Cross-community queries are operational. Knowledge generated in this twin has demonstrably improved outcomes in at least one other context.

Level 5 — Adaptive System: The twin evolves its own scope, updating models, ontologies, and governance rules in response to observed outcomes — within human-defined governance boundaries. Autonomous evolution is traceable, reversible, and explicable.

9.2 Maturity Progression Rules

Maturity levels are not independent per phase. Advancement to Level 3 in Phase 4 requires Level 2 completion in Phase 3. The maturity model is a directed graph, not a matrix. This operationalises the absorptive capacity principle: advancement in later phases is gated by maturity in earlier phases.

The equine domain application (Waern, 2025b) demonstrated that Phase 1 adoption failures — stable staff resistance to structured observation entry — are the most common failure mode in agricultural IS deployments, and that no amount of Phase 3 technical sophistication compensates for a Phase 1 maturity deficit. The maturity model makes this dependency explicit.

9.3 The NUDEDA Lens: Infrastructure Investment Maturity

The SMILE Maturity Model gains additional analytical power when read through the NUDEDA lens — an extension of Weill and Broadbent's (1998) infrastructure investment framework developed within the Gothenburg School informatics tradition (Pessi, Magoulas, Hugoson). The original Weill and Broadbent framework identified four investment postures: None (no shared infrastructure), Utility (basic shared services), Dependent (infrastructure follows strategy), and Enabling (infrastructure enables strategy).

Digital DNA and Autonomous are extensions proposed by the author, building on Weill and Broadbent (1998). They are not attributed to the Gothenburg programme and are not claims about established theory; they are proposed extensions within the DSR framework (Hevner et al., 2004) that require independent validation. Digital DNA describes infrastructure that has become self-describing, ontology-aware, and queryable — corresponding to SMILE's Level 4 (Networked Value). Autonomous describes infrastructure that adapts its own configuration in response to observed outcomes within governance boundaries — corresponding to SMILE's Level 5 (Adaptive System).

This mapping resolves a persistent ambiguity in infrastructure investment literature: Weill and Broadbent's framework was developed for IT infrastructure in the enterprise context and lacks a mechanism for the infrastructure to learn or adapt. NUDEDA provides that proposed mechanism by making the infrastructure itself ontology-aware (Digital DNA) and self-governing (Autonomous). SMILE operationalises NUDEDA by specifying what must be true at each level for a digital twin to function at that investment posture. The proposed extensions are offered as design contributions and require empirical testing before being treated as confirmed theoretical claims.

9.4 AEST as Temporal Navigation Through the Maturity Model

The AEST temporal architecture (Section 7) intersects with the maturity model to determine which temporal operations are available at each level:

The full AEST cycle is only available at Level 5. Premature simulation without adequate emulation produces hallucinated futures. Premature transcendence without validated simulation produces untethered speculation. The maturity model gates temporal operations to ensure that each is supported by the epistemic foundations the previous levels provide.

10. Comparison with Existing Methodologies

10.1 Comparative Table

10.2 Where SMILE Is Positioned Alongside TRIZ and PRINCE2

The TRIZ comparison is instructive because TRIZ achieved genuine universal applicability within its domain by identifying the invariant structure of contradictions that persists across engineering, chemistry, biology, and social systems. TRIZ does not tell you how to design a bicycle or a pharmaceutical process; it tells you how to identify and resolve the contradiction at the core of any design challenge. The domain content changes; the contradiction-resolution logic does not. This is exactly the SMILE claim applied to a different invariant: the phase logic that turns any project into a continuously learning twin does not change across domains; only the content of each phase changes.

The PRINCE2 comparison is equally instructive. PRINCE2 achieved genuine universal project management applicability by abstracting from domain-specific content and operating at the level of project governance logic: business case justification, defined roles, controlled stages, risk management. SMILE operates at the same level of abstraction above domain content, but it governs a different kind of project: one that produces a continuously learning physical-digital system rather than a one-time deliverable.

The key difference from both: SMILE has a Phase 6 and they do not. TRIZ resolves contradictions; it does not govern the resulting system's perpetual evolution. PRINCE2 closes projects; it does not govern the knowledge that persists beyond closure. SMILE's Perpetual Wisdom phase is the structural element that neither TRIZ nor PRINCE2 addresses, and it is the element that makes SMILE specifically suited to physical-digital systems that must improve over time.

10.3 Where Each Methodology Remains Appropriate

CRISP-DM remains appropriate for bounded data analytics projects with a clear data asset and a defined analytical question. DMAIC remains appropriate for process improvement in manufacturing or service operations where a measurable defect has been identified. Design Thinking remains appropriate for early-stage human-centred innovation in novel problem spaces. Agile remains appropriate for software delivery where requirements are emergent and iteration speed matters more than lifecycle structure. TRIZ remains appropriate for structured inventive problem-solving where a specific contradiction has been identified. PRINCE2 remains appropriate for any project where governance, accountability, and controlled change are the primary requirements.

Checkland's (1981) Soft Systems Methodology (SSM) provides the closest methodological ancestor in this regard: SSM governs the process of inquiry into a problematic situation, within which specific technical methodologies operate. SMILE extends SSM's logic from inquiry into sustained operation — from "understanding the situation" to "continuously governing the situation through its digital twin." SMILE does not replace these methodologies; it often precedes them. SMILE Phase 2 may invoke Design Thinking for hypothesis generation. SMILE Phase 3 may invoke CRISP-DM for specific analytical workflows within the knowledge graph. SMILE Phase 5 may deliver AI capabilities through Agile sprints. SMILE Phase 2 may invoke TRIZ to resolve inventive contradictions within the MVT scope. SMILE provides the lifecycle architecture within which domain-specific methodologies operate — not a replacement, but an envelope.

11. Counter-Arguments

11.1 Counter-Argument: "Too Broad to Be Useful"

The objection: a methodology that claims to apply to healthcare, manufacturing, smart cities, equine biology, defence reconstruction, and contributor screening is so general that it provides no real guidance.

This objection rests on a confusion between abstraction level and practical utility. SMILE is abstract at the level of phase logic and concrete at the level of phase deliverables. Phase 3 is "Collective Intelligence" in the abstract; in a building energy context, it is "Haystack-tagged, Brick-ontologised sensor network producing calibrated thermal models." In an equine context, it is "horse health ontology integrating HL7-equivalent clinical records, InfluxDB biosensor time series, and validated behaviour observation instrument."

The TRIZ comparison is again instructive. TRIZ's forty inventive principles are abstract — "segmentation," "taking out," "local quality" — yet they have been productively applied across thousands of engineering, biological, and business problems. Abstraction at the framework level is a feature that enables the framework to generate valid domain-specific guidance without being rewritten for every domain. The genuine risk is insufficient decision rules, not excessive abstraction. SMILE addresses this through its entry and exit criteria, three-perspective lens, and maturity model. Mintzberg's (1979) typology of organisational configurations — machine bureaucracy, professional bureaucracy, adhocracy, divisionalised form — illustrates why methodology must operate above domain content: hospitals, startups, factories, and city governments are structurally different organisations, yet all of them need projects that sustain value over time in environments that change. SMILE's configuration-agnostic phase logic is the structural property that makes this possible; the Gothenburg School's Five Model lineage (Section 4.1) provides the theoretical grounding for why all five organisational dimensions must be addressed regardless of configuration.

11.2 Counter-Argument: "Not Empirically Validated at Scale"

The objection: SMILE has been applied across a single author's corpus of 39 papers. No controlled study has compared SMILE-guided deployments against non-SMILE deployments with matched controls.

This objection is largely correct and Section 13 acknowledges it directly. The honest response is three-fold.

First, the evidentiary standard the objection implies is essentially impossible to apply to any methodology governing complex, multi-year, multi-stakeholder sociotechnical systems. No controlled study has compared CRISP-DM against non-CRISP-DM data analytics projects, or PRINCE2 against ad-hoc project management in matched conditions. Applied consistently, this objection invalidates all methodology claims in the field.

Second, published cross-domain applications including several with documented quantitative outcomes and published DOIs constitute meaningful preliminary evidence. Design science artefacts are typically evaluated through "analytical evaluation, case study, experimental evaluation, field study, and simulation" (Hevner et al., 2004, p. 85). Multiple case studies across eight domains satisfy this criterion for preliminary evaluation.

Third, the objection points to exactly the right future research direction. Section 13 proposes a multi-site comparative validation study as the primary research agenda.

11.3 Counter-Argument: "Just Repackaged Design Thinking"

The objection: SMILE's emphasis on human-centred outcomes, empathy with stakeholders, prototyping before commitment, and iterative refinement reproduces the core claims of Design Thinking in different terminology.

The overlaps are real: both insist on understanding human needs before specifying technical solutions; both value prototyping over specification; both emphasise cross-disciplinary collaboration. SMILE draws on Design Thinking's contributions, particularly for Phase 2 hypothesis generation.

The differences are structural. Design Thinking does not provide a lifecycle framework — it describes an innovation process that produces a concept, but not the subsequent phases that govern how that concept is implemented, operated at scale, maintained over decades, and evolved as its context changes. Design Thinking has no Phase 5 equivalent and no Phase 6 equivalent.

The deeper difference is ontological. Design Thinking begins with human experience and constructs artefacts to serve it. SMILE begins with physical reality — the shared spatial-temporal referent that all communities can describe — and constructs knowledge systems that make that reality legible across communities. The persistent layer in SMILE is not the human; it is the physical world, which has properties that exist independently of any community's description of it. This epistemological commitment to scientific realism (Bhaskar, 1975) distinguishes SMILE from the constructivist orientation of much design thinking work.

12. Falsifiable Propositions

Proposition 1 (Impact-First Sequencing): Digital twin projects that complete a documented Phase 1 (Reality Canvas accepted by all primary stakeholder communities) and Phase 2 (MVT specification with explicit outcome-to-data chain) before initiating physical sensor deployment will demonstrate a statistically significant reduction in project scope change events and stakeholder withdrawal events within the first twelve months of deployment, compared to a matched cohort of digital twin projects that initiated sensor deployment before stakeholder alignment was documented.

The pre-registered minimum effect size for this proposition is d = 0.5 (α = 0.05, power = 0.80), which corresponds to a medium effect in Cohen's (1988) framework. "Statistically significant" is defined by this pre-registered threshold, not by post-hoc threshold selection. If SMILE-sequenced projects show no statistically significant reduction in scope change and stakeholder withdrawal events compared to matched non-SMILE projects at this threshold, the impact-first sequencing principle is falsified.

Proposition 2 (Ontology Before AI — Crucible Efficacy): AI systems deployed within digital twin environments that have completed Phase 3 (operational knowledge graph with multi-community ontology review and MIMs implementation) will produce AI outputs rated as more explainable, more trusted, and more acted-upon by domain stakeholders than AI systems deployed in environments without structured knowledge graphs, holding AI model class and training data volume constant. Furthermore, deployments with a Crucible trust layer will show measurably lower rates of AI-output override by domain experts within the first six months of Phase 5 operation.

This is testable through structured stakeholder surveys using validated instruments (NASA-TLX for cognitive load assessment, Hart & Staveland, 1988; and the Explanation Satisfaction Scale, Hoffman et al., 2018, for explainability ratings) comparing explainability ratings, trust scores, and action rates across matched deployments. If Phase 3 completion and Crucible layer presence show no statistically significant effect on any of these outcomes, the ontology-before-AI principle is falsified.

Proposition 3 (Cross-Domain Knowledge Transfer): Digital twin deployments that reach Phase 6 (at least one documented cross-context knowledge transfer event) will show a statistically significant reduction in the time required to reach Phase 4 operational capability in their second deployment context, compared to the time required for the same organisation to reach Phase 4 in their first deployment context.

This tests whether perpetual wisdom is operationally real or merely aspirational. If Phase 6 deployments show no measurable reduction in Phase 4 time-to-capability in subsequent deployments, the knowledge transfer claim is falsified.

13. Limitations and Future Research

13.1 Limitations

Single-author methodology and single-author validation — the primary constraint. All 39 corpus papers share a single authoring perspective. The theoretical concepts introduced in this paper — including artefactual absorptive capacity (Section 3.4) and the NUDEDA Digital DNA and Autonomous extensions (Section 9.3) — have not been independently evaluated by researchers examining the same deployment evidence. This is the primary methodological constraint of this paper. The validation presented in Section 8 is design-science validation in the DSR sense (Hevner et al., 2004) — evidence of internal consistency and cross-domain applicability — not independent empirical validation of the causal claims. The distinction matters and is not hedged. Multi-site, multi-author validation is the highest-priority research agenda item.

Pending DOIs. Six of the eight domain applications cited in Section 8 have DOIs listed as pending. These applications are design contributions in the DSR sense; they are not confirmed empirical findings. Readers should treat them as they would pre-registered studies that have not yet reported results: the design is documented and the claims are stated, but the evidence is not yet publicly archivable and peer-reviewable.

No controlled comparison study. SMILE's claims rest on a corpus of case applications, not on controlled comparative studies. Observed outcomes in SMILE deployments cannot be causally attributed to SMILE rather than to other factors — organisational capability, stakeholder quality, resource availability — that may correlate with the decision to adopt a structured methodology.

Phase timelines are domain-specific. SMILE specifies phase logic but not phase duration. Practitioners who interpret SMILE's phase descriptions as implying specific timelines will miscalibrate their planning.

Edge technology maturity. Phases 5 and 6 depend on edge-native AI capabilities — LQM reduction to SQMs, federated learning, on-device inference — that are at early maturity stages in several domains. The Phase 5 claims about autonomous operation within governance boundaries assume technical capabilities that are available in some contexts but not yet validated in others.

6G and holographic society are anticipatory. The 6G capabilities referenced in Section 7.6 face documented terahertz propagation constraints (Tataria et al., 2021). The holographic society endpoint describes an anticipated state that is dependent on a 10–15 year standardisation trajectory, not a currently operational capability.

Cultural and political context. SMILE was developed primarily in European and Australian deployment contexts. The methodology's assumptions about stakeholder governance, data sovereignty, and regulatory frameworks reflect these contexts and may require modification for other environments. Floridi and Taddeo (2016) argue that data ethics must be context-sensitive; SMILE's methodology inherits this constraint.

Standards landscape evolution. ISO 23247 (2020) provides the digital twin framework for manufacturing; ISO/IEC 30141 (2018) establishes the IoT reference architecture; ISO 55000 (2014) governs asset management. SMILE's cross-domain claims assume that domain-specific standards will continue to converge — an assumption that current TM Forum, DTC, and ISO activity supports but does not guarantee.

13.2 Future Research

Multi-site comparative validation. The highest-priority research agenda is a prospective, multi-site comparative study comparing SMILE-guided deployments against matched deployments using alternative methodologies. Partners in the Digital Twin Consortium, the EU Digital Europe programme, and national smart city initiatives are the most accessible recruitment sources.

Independent validation of artefactual absorptive capacity. The proposed extension in Section 3.4 requires evaluation by independent researchers examining the same class of artefacts. A research design comparing knowledge absorption rates in twins at different maturity levels across multiple independent deployments would be the appropriate test.

MIMs depth study. A comparative study of Phase 4 interoperability outcomes in deployments with and without explicit MIMs implementation would test whether the MIMs requirement adds measurable value or primarily adds process overhead.

Gothenburg School Five Model formalisation. A formal mapping showing which Five Model dimensions correspond to which SMILE perspective-phase combinations would make the intellectual debt explicit and create a basis for comparative evaluation with other Gothenburg-tradition IS lifecycle methodologies.

SMILE certification programme. A formal SMILE maturity assessment instrument would allow organisations to evaluate their current position across the 6×5 maturity matrix and receive validated guidance on advancement. This would also generate systematic data for the comparative validation study.

Open-source SMILE toolkit. The methodology's tools — Reality Canvas template, MVT specification worksheet, Ontology Factory schema, MIMs implementation guide, SMILE Infrastructure Readiness Checklist, DDNAS scoring rubric — should be assembled into an open-source toolkit with domain-specific extensions.

Physical AI benchmarks. A benchmark suite for Phase 5 edge-native AI performance, LQM-vs-LLM accuracy comparisons in physics-governed domains, and SQM compression ratios would make the Physical AI positioning empirically evaluable.

14. Conclusion: A Methodology That Makes Any Project a Digital Twin Project

The central claim of this paper is structural: SMILE is not a methodology for building digital twins. It is a methodology that turns any project into a digital twin project as a consequence of doing that project well. The digital twin is the output of rigorously designed, impact-first project methodology; it is not the input, the objective, or the premise.

This reframing resolves the persistent confusion in the field between "digital twin projects" (which presuppose the twin as the deliverable) and "projects that need a digital twin" (which discover the twin as the answer to the question "how do we create a system that continuously learns from reality?"). The first class of project is driven by technology availability; the second class is driven by impact requirements. SMILE serves the second class — which is to say, it serves any project that must sustain value over time in an environment that changes.

The twenty benefits described in Section 5 are stated with explicit epistemic differentiation: five are supported by evidence from published corpus applications; fifteen are expected benefits based on phase logic, with empirical validation in progress. This distinction is not a hedge against the methodology's claims; it is the intellectual honesty that makes the claims credible. Benefits management (Ward & Daniel, 2006) teaches that benefits are not automatic — they must be identified, planned, delivered, and reviewed. SMILE accepts this discipline and applies it to its own benefit claims.

The Physical AI framing positions SMILE in the paradigm where it belongs. Intelligence that is grounded in physical reality, operates at the physical edge, maintains continuous calibration against the physical world, and serves physical-world decision-making is not a subset of software AI; it is a different class of system that requires a different class of methodology. Spatial Twinning — Space, Place, Interaction, Network — provides the spatial framework. The AI journey from Data Contextualisation through Explainable AI Decision Making provides the intelligence trajectory. The Crucible trust layer provides the epistemic guarantee: retrieval over generation, explainability by design, no hallucination-friendly architectures.

The five theoretical pillars — Design Science Research, Actor-Network Theory, boundary object theory, absorptive capacity, and benefits management — ground the methodology in established IS scholarship. The worked ANT translation chain in Section 3.2 demonstrates what IS scholarship looks like when theory is used analytically rather than decoratively. The DTC CPT mapping in Section 8.3 provides framework-level alignment with an independent, industry-developed standard. The limitations section names the primary constraint — single-author validation — without qualification.

SMILE's six phases — Reality Emulation, Concurrent Engineering, Collective Intelligence, Contextual Intelligence, Continuous Intelligence, Perpetual Wisdom — are not arbitrary stages. They are the operational expression of five theoretical foundations, two intellectual traditions, and the design evidence of 39 cross-domain applications, three of which have published DOIs. The phase logic is consistent across domains. The phase content is domain-specific. Any project that follows the logic becomes a digital twin project. Any digital twin project that follows the logic becomes continuously better at serving the people, systems, and planet that depend on it.

Drucker (1999) argued that the central challenge of the 21st century is not managing technology but managing knowledge workers — people whose productivity depends on their ability to apply specialised knowledge to goals they understand and own. SMILE's four perspectives — People, Systems, Planet, AI — are the structural answer to this: each perspective ensures that knowledge work is governed by the impact it serves, not by the data it generates. AEST's Transcend operation is Drucker's insight operationalised: not predicting the future, but building the conditions under which the desired future becomes reachable.

The biological digital twin case illustrates where this leads. Topol (2019) argues that deep medicine will restore the human connection in healthcare through AI that handles the data so clinicians can handle the patient. Corral-Acero et al. (2020) demonstrate that computational digital twins of the heart — integrating imaging, genomics, and electrophysiology — can predict individual patient outcomes with clinical-grade accuracy. When precision medicine produces 3D-printed pharmaceuticals — personalised based on an individual's situation in space, time, and physiological context — the actor-network around a single person scales: 10 systems understanding your goals, then 100, then 1,000, then 10,000 — all with a shared representation of your past, present, and intended future. We can gamify our own health trajectories. We can set goals, simulate outcomes, and recruit allies — human, systemic, and artificial — to help us get there.

With Web4 and beyond, we can jump back in time — holoport with others to a past memory, a trauma, a turning point — and work with AI agents, clinicians, coaches, and loved ones to understand what happened, what it meant, and what it could mean differently. We can put ourselves in situations we have been in and want to change, or live again with new understanding. The strength of weak ties extends not only across geography but across time itself.

Ultimately, SMILE's endgame is that the technology disappears. When all systems have reached NUDEDA-Autonomous — when the ambient, hi-fidelity, holographic infrastructure works WITH us and FOR us — we forget about technology entirely. We focus on the real world. On other people. On living lives without limits. The methodology succeeds when the methodology becomes invisible, and what remains is a world where the past is understood, the present is clear, and the future is something we build together rather than something that happens to us.

Impact first. Data last. Everything else follows.

We are getting out of the data dark ages into an impact renaissance — where the future is better, bolder, and more sustainable than the past. SMILE is the foundation layer. The building of a dynasty.

Absorb the past. Emulate the present. Simulate the future. Transcend the now.

15. Funding Acknowledgement

This work was informed by research funded by the Swedish Energy Agency (Grant 2023-203019: "Verification, Further Development and Upscaling of Smart Heating Systems") and by the author's ongoing work as Co-Chair of the Digital Twin Consortium Manufacturing Working Group and lead developer of the Q-SMART testbed (Digital Twin Consortium, 2025). Cross-domain applications were developed through WINNIIO AB's engagement with partners across the built environment, life sciences, telecommunications, and equine biology sectors.

This article was produced with the assistance of a Large Language Model (Claude, Anthropic), acknowledged as a mediating actor in the translation chain from conceptual intent to published form. The author takes full responsibility for all theoretical claims, factual assertions, and methodological positions expressed herein.

16. References

Ackoff, R. L. (1989). From data to wisdom. Journal of Applied Systems Analysis, 16, 3–9.

Allam, Z., Sharifi, A., Bibri, S. E., Jones, D. S., & Krogstie, J. (2022). The metaverse as a virtual form of smart cities: Opportunities and challenges for environmental, economic, and social sustainability in urban futures. Smart Cities, 5(3), 771–801. https://doi.org/10.3390/smartcities5030040

Altshuller, G. (1996). And Suddenly the Inventor Appeared: TRIZ, the Creative Problem Solving Tool. Technical Innovation Center.

Arrieta, A. B., Díaz-Rodríguez, N., Del Ser, J., Bennetot, A., Tabik, S., Barbado, A., ... & Herrera, F. (2020). Explainable Artificial Intelligence (XAI): Concepts, taxonomies, opportunities and challenges toward responsible AI. Information Fusion, 58, 82–115. https://doi.org/10.1016/j.inffus.2019.12.012

Ashurst, C., Doherty, N. F., & Peppard, J. (2008). Improving the impact of IT development projects: The benefits realization capability model. European Journal of Information Systems, 17(4), 352–370. https://doi.org/10.1057/ejis.2008.33

Axelos. (2017). PRINCE2 Agile. TSO (The Stationery Office).

Barricelli, B. R., Casiraghi, E., & Fogli, D. (2019). A survey on digital twin: Definitions, characteristics, applications, and design implications. IEEE Access, 7, 167653–167671. https://doi.org/10.1109/ACCESS.2019.2953499

Batty, M. (2013). The New Science of Cities. MIT Press.

Baxter, G., & Sommerville, I. (2011). Socio-technical systems: From design methods to systems engineering. Interacting with Computers, 23(1), 4–17. https://doi.org/10.1016/j.intcom.2010.07.003

Becker, J., Knackstedt, R., & Pöppelbuß, J. (2009). Developing maturity models for IT management. Business & Information Systems Engineering, 1(3), 213–222. https://doi.org/10.1007/s12599-009-0044-5

Bhaskar, R. (1975). A Realist Theory of Science. Leeds Books.

Bowker, G. C., & Star, S. L. (1999). Sorting Things Out: Classification and Its Consequences. MIT Press.

Brown, T. (2009). Change by Design: How Design Thinking Transforms Organizations and Inspires Innovation. HarperBusiness.

Brynjolfsson, E. (1993). The productivity paradox of information technology. Communications of the ACM, 36(12), 66–77. https://doi.org/10.1145/163298.163309

Callon, M. (1986). Some elements of a sociology of translation: Domestication of the scallops and the fishermen of St Brieuc Bay. In J. Law (Ed.), Power, Action and Belief (pp. 196–233). Routledge.

Checkland, P. (1981). Systems Thinking, Systems Practice. John Wiley & Sons.

Christensen, C. M. (1997). The Innovator's Dilemma: When New Technologies Cause Great Firms to Fail. Harvard Business School Press.

Ciborra, C. U. (2000). From Control to Drift: The Dynamics of Corporate Information Infrastructures. Oxford University Press.

Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences (2nd ed.). Lawrence Erlbaum Associates.

Cohen, W. M., & Levinthal, D. A. (1990). Absorptive capacity: A new perspective on learning and innovation. Administrative Science Quarterly, 35(1), 128–152. https://doi.org/10.2307/2393553

Corral-Acero, J., Margara, F., Marber, M., Schottlander, D., Shi, W., et al. (2020). The 'Digital Twin' to enable the vision of precision cardiology. European Heart Journal, 41(48), 4556–4564. https://doi.org/10.1093/eurheartj/ehaa159

de Bruin, T., Freeze, R., Kulkarni, U., & Rosemann, M. (2005). Understanding the main phases of developing a maturity assessment model. In Proceedings of the 16th Australasian Conference on Information Systems (ACIS 2005).

Drucker, P. F. (1999). Management Challenges for the 21st Century. HarperBusiness.

Emery, F. E., & Trist, E. L. (1960). Socio-technical systems. In C. W. Churchman & M. Verhulst (Eds.), Management Sciences, Models and Techniques (Vol. 2, pp. 83–97). Pergamon Press.

European Commission. (2017). New European Interoperability Framework. Publications Office of the European Union. https://doi.org/10.2799/78681

European Parliament and Council. (2024). Regulation (EU) 2024/1689 laying down harmonised rules on artificial intelligence (Artificial Intelligence Act). Official Journal of the European Union, L 2024/1689.

Digital Twin Consortium. (2024). Digital Twin Capabilities Periodic Table User Guide, Version 1.1. van Schalkwyk, P., Whiteley, B., & Goldman, D. Object Management Group. https://www.digitaltwinconsortium.org/

Digital Twin Consortium. (2025). Digital Twin Consortium Adds Eight New Testbeds [Press release]. https://www.globenewswire.com/news-release/2025/09/16/3150899/0/en/

Floridi, L., Cowls, J., Beltrametti, M., Chatila, R., Chazerand, P., Dignum, V., Luetge, C., Madelin, R., Pagallo, U., Rossi, F., Schafer, B., Valcke, P., & Vayena, E. (2018). AI4People — An ethical framework for a good AI society. Minds and Machines, 28(4), 689–707. https://doi.org/10.1007/s11023-018-9482-5

Floridi, L., & Taddeo, M. (2016). What is data ethics? Philosophical Transactions of the Royal Society A, 374(2083), 20160360. https://doi.org/10.1098/rsta.2016.0360

Fuller, A., Fan, Z., Day, C., & Barlow, C. (2020). Digital twin: Enabling technologies, challenges and open research. IEEE Access, 8, 108952–108971. https://doi.org/10.1109/ACCESS.2020.2998358

Gawer, A., & Cusumano, M. A. (2014). Industry platforms and ecosystem innovation. Journal of Product Innovation Management, 31(3), 417–433. https://doi.org/10.1111/jpim.12105

Geissdoerfer, M., Savaget, P., Bocken, N. M. P., & Hultink, E. J. (2017). The Circular Economy — A new sustainability paradigm? Journal of Cleaner Production, 143, 757–768. https://doi.org/10.1016/j.jclepro.2016.12.048

Glaessgen, E., & Stargel, D. (2012). The digital twin paradigm for future NASA and US Air Force vehicles. In 53rd AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics and Materials Conference. AIAA. https://doi.org/10.2514/6.2012-1818

Granovetter, M. S. (1973). The strength of weak ties. American Journal of Sociology, 78(6), 1360–1380. https://doi.org/10.1086/225469

Grieves, M. (2014). Digital Twin: Manufacturing Excellence through Virtual Factory Replication. White paper. https://www.researchgate.net/publication/275211047

Gruber, T. R. (1993). A translation approach to portable ontology specifications. Knowledge Acquisition, 5(2), 199–220. https://doi.org/10.1006/knac.1993.1008

Gunning, D., & Aha, D. W. (2019). DARPA's Explainable Artificial Intelligence (XAI) program. AI Magazine, 40(2), 44–58. https://doi.org/10.1609/aimag.v40i2.2850

Hart, S. G., & Staveland, L. E. (1988). Development of NASA-TLX (Task Load Index): Results of empirical and theoretical research. In P. A. Hancock & N. Meshkati (Eds.), Human Mental Workload (pp. 139–183). North-Holland. https://doi.org/10.1016/S0166-4115(08)62386-9

Hendler, J., & Berners-Lee, T. (2010). From the Semantic Web to social machines: A research challenge for AI on the World Wide Web. Artificial Intelligence, 174(2), 156–161. https://doi.org/10.1016/j.artint.2009.11.010

Hevner, A. R., March, S. T., Park, J., & Ram, S. (2004). Design science in information systems research. MIS Quarterly, 28(1), 75–105. https://doi.org/10.2307/25148625

Hoffman, R. R., Mueller, S. T., Klein, G., & Litman, J. (2018). Metrics for Explainable AI: Challenges and Prospects. arXiv preprint. https://arxiv.org/abs/1812.04608

Hugoson, M. A., Magoulas, T., & Pessi, K. (2011). Interoperability strategies for business agility. Proceedings of the 5th European Conference on Information Management and Evaluation. Academic Conferences International.

Hutchins, E. (1995). Cognition in the Wild. MIT Press.

Iansiti, M. (2004). The Keystone Advantage: What the New Dynamics of Business Ecosystems Mean for Strategy, Innovation, and Sustainability. Harvard Business School Press.

Iansiti, M., & Lakhani, K. R. (2020). Competing in the Age of AI: Strategy and Leadership When Algorithms and Networks Run the World. Harvard Business Review Press.

ISO. (2014). ISO 55000: Asset management — Overview, principles and terminology. International Organization for Standardization.

ISO. (2020). ISO 23247: Automation systems and integration — Digital twin framework for manufacturing. International Organization for Standardization.

ISO/IEC. (2018). ISO/IEC 30141: Internet of Things (IoT) — Reference architecture. International Organization for Standardization.

ITU-R. (2023). IMT-2030 (6G) Framework Recommendation. ITU-R M.2160. International Telecommunication Union.

Jones, D., Snider, C., Nassehi, A., Yon, J., & Hicks, B. (2020). Characterising the Digital Twin: A systematic literature review. CIRP Journal of Manufacturing Science and Technology, 29, 36–52. https://doi.org/10.1016/j.cirpj.2020.02.002

Kairouz, P., McMahan, H. B., Avent, B., Bellet, A., Bennis, M., et al. (2021). Advances and open problems in federated learning. Foundations and Trends in Machine Learning, 14(1–2), 1–210. https://doi.org/10.1561/2200000083

Kapusuzoglu, B., & Mahadevan, S. (2024). Data-driven physics-informed neural networks: A digital twin perspective. Computer Methods in Applied Mechanics and Engineering, 421, 116789. https://doi.org/10.1016/j.cma.2024.116789

Karniadakis, G. E., Kevrekidis, I. G., Lu, L., Perdikaris, P., Wang, S., & Yang, L. (2021). Physics-informed machine learning. Nature Reviews Physics, 3(6), 422–440. https://doi.org/10.1038/s42254-021-00314-5

Kerbl, B., Kopanas, G., Leimkühler, T., & Drettakis, G. (2023). 3D Gaussian Splatting for Real-Time Radiance Field Rendering. ACM Transactions on Graphics, 42(4). https://doi.org/10.1145/3592433

Kirchherr, J., Reike, D., & Hekkert, M. (2017). Conceptualizing the circular economy: An analysis of 114 definitions. Resources, Conservation and Recycling, 127, 221–232. https://doi.org/10.1016/j.resconrec.2017.09.005

Kitchin, R. (2014). The Data Revolution: Big Data, Open Data, Data Infrastructures and Their Consequences. SAGE Publications.

Kritzinger, W., Karner, M., Traar, G., Henjes, J., & Sihn, W. (2018). Digital Twin in manufacturing: A categorical literature review and classification. IFAC-PapersOnLine, 51(11), 1016–1022. https://doi.org/10.1016/j.ifacol.2018.08.474

Latour, B. (1987). Science in Action: How to Follow Scientists and Engineers through Society. Harvard University Press.

Latour, B. (2005). Reassembling the Social: An Introduction to Actor-Network-Theory. Oxford University Press.

Law, J. (1992). Notes on the theory of the actor-network: Ordering, strategy, and heterogeneity. Systemic Practice and Action Research, 5(4), 379–393. https://doi.org/10.1007/BF01059830

Leonardi, P. M. (2012). Materiality, sociomateriality, and socio-technical systems: What do these terms mean? How are they related? Do we need them? In P. M. Leonardi, B. A. Nardi, & J. Kallinikos (Eds.), Materiality and Organizing: Social Interaction in a Technological World (pp. 25–48). Oxford University Press.

Lewis, P., Perez, E., Piktus, A., Petroni, F., Karpukhin, V., et al. (2020). Retrieval-augmented generation for knowledge-intensive NLP tasks. In Advances in Neural Information Processing Systems (NeurIPS 2020). https://doi.org/10.48550/arXiv.2005.11401

Lu, Y., Liu, C., Wang, K. I.-K., Huang, H., & Xu, X. (2024). Comprehensive digital twin for infrastructure: A novel ontology and graph-based modelling paradigm. Advanced Engineering Informatics, 62, 102851. https://doi.org/10.1016/j.aei.2024.102851

Magoulas, T., & Pessi, K. (1998). Strategic IT management. In Proceedings of the 7th International Conference on Information Systems Development. Gothenburg University.

McMahan, H. B., Moore, E., Ramage, D., Hampson, S., & Arcas, B. A. y. (2017). Communication-efficient learning of deep networks from decentralized data. In Proceedings of the 20th International Conference on Artificial Intelligence and Statistics (AISTATS 2017). https://proceedings.mlr.press/v54/mcmahan17a.html

Miller, T. (2019). Explanation in artificial intelligence: Insights from the social sciences. Artificial Intelligence, 267, 1–38. https://doi.org/10.1016/j.artint.2018.07.007

Mintzberg, H. (1979). The Structuring of Organizations. Prentice-Hall.

Nonaka, I., & Takeuchi, H. (1995). The Knowledge-Creating Company: How Japanese Companies Create the Dynamics of Innovation. Oxford University Press.

OASC. (2023). Minimal Interoperability Mechanisms (MIMs). Open & Agile Smart Cities. https://oascities.org/minimal-interoperability-mechanisms/

European Parliament and Council. (2007). Directive 2007/2/EC establishing an Infrastructure for Spatial Information in the European Community (INSPIRE). Official Journal of the European Union, L 108, 1–14.

Orlikowski, W. J. (2007). Sociomaterial practices: Exploring technology at work. Organization Studies, 28(9), 1435–1448. https://doi.org/10.1177/0170840607081138

Park, J. S., O'Brien, J. C., Cai, C. J., Morris, M. R., Liang, P., & Bernstein, M. S. (2023). Generative agents: Interactive simulacra of human behavior. In Proceedings of the 36th Annual ACM Symposium on User Interface Software and Technology (UIST '23). https://doi.org/10.1145/3586183.3606763

Pinch, T. J., & Bijker, W. E. (1984). The social construction of facts and artefacts. Social Studies of Science, 14(3), 399–441. https://doi.org/10.1177/030631284014003004

Rackham, N. (1988). SPIN Selling. McGraw-Hill.

Raissi, M., Perdikaris, P., & Karniadakis, G. E. (2019). Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations. Journal of Computational Physics, 378, 686–707. https://doi.org/10.1016/j.jcp.2018.10.045

Rasheed, A., San, O., & Kvamsdal, T. (2020). Digital twin: Values, challenges and enablers from a modeling perspective. IEEE Access, 8, 21980–22012. https://doi.org/10.1109/ACCESS.2020.2970143

Rittel, H. W. J., & Webber, M. M. (1973). Dilemmas in a general theory of planning. Policy Sciences, 4(2), 155–169. https://doi.org/10.1007/BF01405730

Satyanarayanan, M. (2017). The emergence of edge computing. Computer, 50(1), 30–39. https://doi.org/10.1109/MC.2017.9

Schön, D. A. (1983). The Reflective Practitioner: How Professionals Think in Action. Basic Books.

Senge, P. M. (1990). The Fifth Discipline: The Art & Practice of The Learning Organization. Doubleday.

Shi, W., Cao, J., Zhang, Q., Li, Y., & Xu, L. (2016). Edge computing: Vision and challenges. IEEE Internet of Things Journal, 3(5), 637–646. https://doi.org/10.1109/JIOT.2016.2579198

Simon, H. A. (1996). The Sciences of the Artificial (3rd ed.). MIT Press.

Star, S. L. (2010). This is not a boundary object: Reflections on the origin of a concept. Science, Technology, & Human Values, 35(5), 601–617. https://doi.org/10.1177/0162243909345838

Star, S. L., & Griesemer, J. R. (1989). Institutional ecology, "translations" and boundary objects: Amateurs and professionals in Berkeley's Museum of Vertebrate Zoology, 1907–39. Social Studies of Science, 19(3), 387–420. https://doi.org/10.1177/030631289019003001

Tao, F., Zhang, H., Liu, A., & Nee, A. Y. C. (2019). Digital twin in industry: State-of-the-art. IEEE Transactions on Industrial Informatics, 15(4), 2405–2415. https://doi.org/10.1109/TII.2018.2873186

Topol, E. J. (2019). Deep Medicine: How Artificial Intelligence Can Make Healthcare Human Again. Basic Books.

Tuckman, B. W. (1965). Developmental sequence in small groups. Psychological Bulletin, 63(6), 384–399. https://doi.org/10.1037/h0022100

Tataria, H., Shafi, M., Molisch, A. F., Dohler, M., Sjöland, H., & Tufvesson, F. (2021). 6G wireless systems: Vision, requirements, challenges, insights, and opportunities. Proceedings of the IEEE, 109(7), 1166–1199. https://doi.org/10.1109/JPROC.2021.3061701

Wall, S. D., & Ledbetter, K. W. (1991). Design of Mission Operations Systems for Scientific Remote Sensing. Taylor & Francis.

Ward, J., & Daniel, E. (2006). Benefits Management: Delivering Value from IS & IT Investments. John Wiley & Sons.

Weill, P., & Broadbent, M. (1998). Leveraging the New Infrastructure: How Market Leaders Capitalize on Information Technology. Harvard Business School Press.

Wooldridge, M. (2009). An Introduction to MultiAgent Systems (2nd ed.). John Wiley & Sons.

WHOQOL Group. (1995). The World Health Organization Quality of Life assessment (WHOQOL): Position paper from the World Health Organization. Social Science & Medicine, 41(10), 1403–1409. https://doi.org/10.1016/0277-9536(95)00112-K

Zammuto, R. F., Griffith, T. L., Majchrzak, A., Dougherty, D. J., & Faraj, S. (2007). Information technology and the changing fabric of organization. Organization Science, 18(5), 749–762. https://doi.org/10.1287/orsc.1070.0306

Zheng, X., Lu, J., & Kiritsis, D. (2024). Digital twin meets knowledge graph for intelligent manufacturing processes. Sensors, 24(8), 2618. https://doi.org/10.3390/s24082618

Corpus self-citations (SMILE applications):

Waern, N. (2025a). From One Room to Fifty: Orchestrating Explainable AI, Resilience, and Contextual Interoperability in the Built Environment. WINNIIO AB. https://doi.org/10.5281/zenodo.17462962

Waern, N. (2025b). Beyond the Shadows — Contextual Awakening, Federated Learning, and the Realization of Reality through Digital Twins. WINNIIO AB. https://doi.org/10.5281/zenodo.17464804

Waern, N. (2026a). Indexing Reality: Boundary-Spanning Objects, Automated Scene Creation, and the Convergence of Spatial Intelligence with RF Digital Twins. WINNIIO AB. https://doi.org/10.5281/zenodo.19587944

Waern, N. (2026b). From Spreadsheets to Spatial Fabric: Why the Equine Industry Needs a Data Operating System. WINNIIO AB. [Zenodo DOI: pending — design application pending empirical validation]

Waern, N. (2026c). The Sovereign Body: Personal Digital Twins as Boundary-Spanning Objects Across the WHO Quality of Life Dimensions. WINNIIO AB. https://doi.org/10.5281/zenodo.19586851

Waern, N. (2026d). The Reality Construct: Digital Twins as Boundary-Spanning Artefacts for Knowledge Absorption, Organisational Evolution, and the Co-Authorship of Intended Futures. WINNIIO AB. https://doi.org/10.5281/zenodo.19586835

Waern, N. (2026e). Personal Health Digital Twins as the Glue for Cross-Border Patient Data. WINNIIO AB. [Zenodo DOI: pending — design application pending empirical validation]

Waern, N. (2026f). Digital DNA Scoring: Measuring a City's Readiness for Autonomous Operations. WINNIIO AB. [Zenodo DOI: pending — design application pending empirical validation]

Waern, N. (2026g). Infrastructure Before Intelligence: Why a Biotech Startup Chose Cloud Foundations Over AI Models. WINNIIO AB. [Zenodo DOI: pending — design application pending empirical validation]

Waern, N. (2026h). 114 to 9: What a Live Leaderboard Teaches You About Contributor Screening at Scale. WINNIIO AB. [Zenodo DOI: pending — design application pending empirical validation]

Waern, N. (2026i). [Post-conflict innovation and digital twin reconstruction]. WINNIIO AB. https://doi.org/10.5281/zenodo.19601817

Waern, N. (2026j). [Drone-enabled reconstruction and emergency response in disaster contexts]. WINNIIO AB. https://doi.org/10.5281/zenodo.19643598

Waern, N. (2026k). [Genomics infrastructure paper]. WINNIIO AB. [Zenodo DOI: pending — design application pending empirical validation]

Appendix A: SMILE Phase Summary Reference Card

Appendix B: The SMILE Inverted Pyramid

The impact-first sequencing principle is represented as an inverted pyramid. Most technology deployments build upward from the base:

Data → Information → Insight → Action → Outcome
(data-first: hope that outcome emerges)

SMILE requires designing downward from the apex:

Outcome (defined first: what change for whom?)
  ↓ Action (what must change to achieve the outcome?)
    ↓ Insight (what must be understood to guide the action?)
      ↓ Information (what processed knowledge produces the insight?)
        ↓ Data (what raw measurements produce the information?)

The inverted pyramid has three operational consequences: (1) every data collection request must be traced back to an outcome; (2) data without an outcome chain is waste; and (3) the twin is evaluated by outcomes achieved, not data collected. The Benefits Dependency Network (Ward & Daniel, 2006) is the formal IS tool for mapping this chain from Outcome to Data; SMILE's Phase 2 artefacts are the operational instantiation of that map.