Extreme
Collaboration

Cities worldwide invest billions in smart city infrastructure yet consistently fail to achieve interoperability, shared decision-making, or measurable impact. The data accumulates. The dashboards multiply. The decisions remain as fragmented and slow as before.

This paper presents an action-based research framework grounded in Extreme Collaboration — a methodology that inverts the conventional data-first approach by beginning with human outcomes and using shared reality as the coordination substrate. Shared reality, in this context, means digital twins: physics-based, visually navigable, temporally continuous models of the urban environment that all stakeholders can see, interrogate, and build decisions upon simultaneously.

Drawing from case engagements including Olympic city planning for Brisbane 2032 and LA 2028, smart city transformations across Europe and Asia, and Ukraine reconstruction feasibility studies, we propose a progression model from fragmented silos to autonomous cognitive city operations. The framework introduces the NUDED Digital Maturity Index, the SMILE implementation methodology, the PESTELED contextual lens, and a participatory decision-support architecture that repositions the city itself as the boundary-spanning object and tourism as the driver — not the consumer — of urban intelligence. Critically, this paper argues that cities must not become technology experts. They must become environments where experts can help them — fast, spatially, and on the city's own terms.

What follows is both a research agenda and a practitioner framework. It is grounded in field experience, designed to be reproduced, and built on a single conviction: that the hardest engineering problem in urban intelligence is not technical. It is human.

The modern city is not a system. It is an archipelago. Real estate does not speak to energy. Energy does not speak to water. Water does not speak to tourism. Tourism does not speak to GIS. GIS does not speak to traffic. Each island has developed its own data standards, its own organizational culture, its own political mandate, and its own definition of success. What they share is geography. What they do not share is context.

Technical interoperability efforts — open APIs, data standards, unified platforms — address the surface of this problem without touching its root. Even when data flows freely between departments, decisions are still made by disconnected people operating on disconnected timescales with disconnected objectives. A water department can receive real-time tourism data and still make infrastructure decisions in a four-year budget cycle that ignores every peak and trough in that data.

The OASC (Open and Agile Smart Cities) Minimal Interoperability Mechanisms framework addresses the first two layers with sophistication and real institutional traction. What it cannot address is the third and fourth — because no technical standard can mandate that a city department trust another department, share credit, or accept shared accountability for outcomes.

Sweden, as one example, has developed a cultural pathology around data. The word “data-driven” appears in virtually every public sector strategy document produced since 2015. Yet data-driven, as typically practiced, means: collect more data, build more dashboards, hold more meetings about what the dashboards show, and postpone decisions until the data is more complete. This is not data-driven decision-making. It is data-deferred decision-making.

The dashboard is a representation of reality. It is not reality. A city official who has spent thirty years walking Gothenburg's waterfront holds knowledge that no sensor array has yet encoded. The gap between that tacit knowledge and the data on the dashboard is where most smart city initiatives fail — not because the data is wrong, but because the people who understand context were never in the conversation.

Conventional smart city methodology follows a logical but dysfunctional sequence: collect data, turn it into information, extract insight, design action, and — eventually, if the project survives — achieve some approximation of the intended outcome. The problem is not the sequence. The problem is the rate of attrition. Organizations run out of budget at the insight stage. They run out of political will at the action stage. By the time the initiative reaches the outcome, it has been so attenuated by compromise and delay that the original impact goal is unrecognizable.

The inversion is not merely rhetorical. It changes what gets funded, who gets invited to the table, and what counts as progress. Starting with the outcome means beginning every engagement with a single question: what does success look like in this city, for these people, in five years? Not what data can we collect. Not what platform should we procure. What does impact look like — and can we describe it precisely enough that we will know it when we see it?

A practical example. A coastal European city is preparing for a 40 percent increase in tourist arrivals over the next decade. The conventional approach begins with sensors, a unified data platform, and a request for proposal. The SMILE approach begins with a different question: how do we absorb two million additional visitors per year without destroying the livability that makes the city worth visiting in the first place? That question — not the data — is the design brief.

The answer to that question requires tourism, water, transport, planning, and citizen engagement to be in the same conversation from day one. And that conversation cannot happen in a meeting room with PowerPoint slides. It requires a shared representation of reality that all parties can see, point to, and simulate against.

Reality, rendered as a navigable digital twin, is not a deliverable at the end of the process. It is the medium through which the process happens. It is the table that all parties sit around. It is the shared language that allows a water engineer and a tourism director to have the same conversation at the same level of concreteness — pointing at the same map, seeing the same simulation, arguing about the same future.

A digital twin, properly understood, is not a visualization tool. It is a physics-based, continuously updated, time-navigable model of a system. For a city, this means a model that knows not just the current state of infrastructure, population distribution, energy load, and water stress, but can be moved backwards through historical states and forwards through simulated futures. The same model. The same coordinate system. The same ground truth — available to every stakeholder, simultaneously, without translation.

The spatial web — sometimes described as Web 4.0 — uses geography as its universal coordinate system. Every city, regardless of size, exists on the same planet. The coordinate system does not change. This means that Dubrovnik, Tampere, and Gothenburg can be overlaid and compared on the same substrate. The knowledge transfer between cities is not a document or a case study. It is a spatial comparison: here is how Tampere handled peak tourist flow at these coordinates, under these conditions. Here is what happened to water stress. Here is the decision that was made. Would it work here?

The time dimension is equally transformative. A city official who needs to understand the cumulative impact of thirty years of tourist development does not need to read thirty years of reports. She needs a time slider. Drag it backwards to 1995: see the city as it was. Drag it to today: see the city as it is. Drag it forward to 2035 under three different policy scenarios: see the city as it could become. The simulation shows water stress, traffic saturation, waste infrastructure capacity, air quality, and economic vitality — simultaneously, spatially, for each scenario.

This is not data visualization. Data visualization shows what has been measured. Shared reality augmentation shows what is, what was, and what could be. The difference is the difference between a photograph and a film that has not yet been made — one in which the viewer can change the ending.

When a tourism director says “show me Brisbane with fifty thousand additional visitors during the Olympics,” the simulation does not return a spreadsheet. It shows water stress propagating outward from the stadium district. It shows traffic collapse on the secondary arterials that were not modeled in the original infrastructure plan. It shows waste overflow at the three transfer stations that were built for a city of three million, not a city that briefly hosts five. These are not predictions. They are consequences, made visible before the decisions that generate them are finalized.

Consider Dubrovnik. A coastal city of forty-three thousand permanent residents that absorbs over a million visitors per year. Its tourism department has knowledge about peak flow management, seasonal staffing, and infrastructure stress that Gothenburg — a city ten times larger but only now experiencing rapid tourism growth — desperately needs. Conversely, Gothenburg has invested heavily in transport-integrated digital twins and energy-efficient district management that Dubrovnik has not yet attempted.

In the conventional model, these cities would never talk to each other. They exist in different countries, different administrative traditions, different languages. Whatever knowledge transfer occurs happens through conferences, published case studies, and consultancy firms that have worked in both contexts.

In the shared reality model, Dubrovnik and Gothenburg exist on the same planet — literally. Their digital twins occupy the same coordinate system. A city official in Gothenburg can zoom into Dubrovnik's harbor district, see how tourist flows were managed during a peak event, observe the infrastructure response, and ask: what would this look like if we applied it to our waterfront? The twin does not merely display. It simulates. It translates. It teaches.

Add Tampere, Helsinki, Brisbane, a cluster of Indian villages surveyed by drone. Each city contributes a different capability to the network. Tampere teaches winter infrastructure management. Brisbane teaches mega-event absorption. The Indian village cluster teaches peer-to-peer scaling with minimal resources. None of these cities needs to become expert in the others' domains. They share spatial knowledge — anchored in reality, visible on the map, translatable across contexts.

This is the vision that changes everything: cities sharing AI agents, sharing simulation results, sharing resource allocation strategies — not through bureaucratic agreements that take years to negotiate, but through their digital twins, which can interoperate at the speed of computation. Gothenburg's water management agent notices a pattern that matches what Dubrovnik's tourism agent learned last summer. The insight propagates. The cities get smarter together. The network becomes the intelligence.

Every city exists within a web of external forces that shape what is possible, what is permissible, and what is urgent. The PESTELED framework — Political, Economic, Social, Technological, Environmental, Legal, Ethical, and Demographic — provides the structured external scan that any city transformation must account for. No digital twin operates in a vacuum. A city's AI agents must understand not just internal infrastructure but the regulatory environment (the EU AI Act constrains what autonomous agents can decide), the economic context (tourism revenue dependency creates political pressure that distorts priority-setting), the demographic trajectory (an aging population changes what “livability” means), and the ethical boundaries (surveillance capabilities that are technically possible but socially corrosive).

PESTELED is not a one-time assessment. It is a continuous environmental scan that feeds into the digital twin's decision context. When a city's agent recommends a tourism optimization, it does so with awareness that the local political climate has shifted toward resident protection, that new EU data regulations constrain cross-border data sharing, that the demographic profile of visitors has changed, and that the ethical implications of predictive crowd management have been publicly debated. Context-aware autonomy — not context-blind automation — is the design goal.

Most digital maturity frameworks measure technology adoption. How many sensors does the city have? What percentage of services are digitized? Is the city using AI? These are questions about inputs. They do not measure outputs. They do not measure whether the technology is changing how decisions are made, how quickly, or with what quality.

The NUDED index takes a different approach. It scores how organizations — at the role level, the department level, and the city level — relate to technology in terms of decision velocity and organizational trust. The acronym is deliberate: None, Utility, Dependent, Enabler, DNA — and then the critical extension — Distributed. Six levels, not five.

The index is applied per role, per department, and per city. A tourism manager scoring Enabler while the water department scores None reveals a collaboration gap that no amount of technical integration will resolve — because the issue is not the absence of data but the absence of shared decision-making culture. The water department does not know how to use tourist flow data even if it arrives in perfect format.

The most counterintuitive finding from field applications of the NUDED index is this: cities scoring DNA in technology adoption but None in inter-departmental collaboration consistently perform worse on impact outcomes than cities scoring Utility in technology but Enabler in collaboration. The quality of cross-silo relationships is a stronger predictor of smart city success than the sophistication of the technology deployed.

The Distributed level — the final D — is the horizon. It describes a city whose intelligence is not contained within its own boundaries. A city that has reached Distributed does not merely use technology well internally. It exports its learning to peer cities. It imports best practices from other contexts. Its AI agents negotiate with other cities' agents. Its digital twin is not an island — it is a node in a network of city twins that learn from each other, teach each other, and share resources across geographic and political boundaries.

A city that makes high-quality cross-departmental decisions in days, using basic tools and strong relationships, will outperform a city that makes low-quality siloed decisions in months, using expensive platforms. The technology matters. It matters less than the collaboration architecture it is deployed within.

Extreme Collaboration is not a metaphor. It is a specific methodology for accelerating the organizational conditions under which intelligent urban systems become possible. Its core premise is simple: before any data platform, AI system, or digital twin can deliver value, the humans who will use it must be in conversation with each other — and that conversation must happen in shared reality, not in the abstract.

The intervention begins with a room. All relevant stakeholders — tourism, water, transport, planning, energy, emergency services, and citizen representatives — are brought into the same physical or virtual space. Not for a workshop about technology. For a workshop about reality. Their reality. The city they actually inhabit, serve, and are responsible for.

The standard questionnaire approach to organizational research produces abstract answers to abstract questions. “What are your main challenges?” yields generic responses. The map-anchored interview replaces this with a concrete, spatial alternative: every interview is conducted in front of the city's digital twin. Every challenge, every bottleneck, every opportunity is located. “Here is where our water infrastructure cannot scale. Here is where the construction project conflicts with the tourist corridor. Here is the intersection that fails every August.”

Spatial grounding transforms the quality of information collected. It also transforms the interview itself into a collaborative act — the interviewer and interviewee are both looking at the same reality, exploring it together, discovering connections that neither would have surfaced in a conventional conversation.

Every external engagement — whether from a cloud partner, an AI consultancy, or a smart city integrator — begins by needing to understand the same four things. The SPIN diagnostic (Situation, Problem, Implication, Need) provides the structure. But the critical insight is where the SPIN knowledge lives. It does not live in the external partner. It lives in the city's own people.

When the city's own stakeholders have collectively articulated their SPIN — on the map, together, in shared reality — external partners can be brought in with surgical precision. The cloud provider, the AI partner, the smart city consultant: they do not need to spend six months assessing the situation. The city has already done it. The city hands them a spatial brief. The experts arrive knowing exactly what the problems are, where they are, who is affected, and what success looks like. This compresses external engagements from years to weeks.

The most important discovery from applying this methodology is that the city itself — rendered as a shared digital twin — becomes the boundary-spanning object. Not a department. Not a steering committee. Not a consultant's framework. The city. The map. The reality that everyone shares but no one has been looking at together.

In organizational theory, a boundary-spanning object is any artifact that allows different communities of practice to coordinate without requiring each community to fully understand the other's domain. The digital twin of the city serves this function exactly. A water engineer does not need to understand tourism economics. A tourism director does not need to understand pipe capacity. But both can point at the same district on the same map and say: here is where your world and my world collide. Here is where we need to decide together.

This is why the framing matters. The city is not a client that hires experts. The city is the collaboration partner — the primary one. The knowledge required to design the next stage of the city's intelligence resides, overwhelmingly, within the city's own staff — in departments that have never been asked to sit in the same room at the same time. External experts are accelerators, not architects. The architecture is the city itself.

Small cities hold a structural advantage in this methodology. A city of fifty thousand people may have five to ten critical decision-makers across all relevant departments. Those people can be in one room. A decision that would take eighteen months in a large metropolitan bureaucracy can be made in a day — if the right people are looking at the same reality simultaneously, with the authority to commit.

This is not a soft skill observation. It is a structural design principle. Smart city investments in small cities, anchored in Extreme Collaboration methodology, can produce decision-cycle compression of an order of magnitude over large-city implementations. The question is whether large cities can architect internal structures that replicate the conditions small cities enjoy by default.

The path from siloed city operations to autonomous cognitive city management is not a leap. It is a measurable progression, each stage building the organizational and technical conditions required for the next. The stages are not sequential in the sense that one must be fully complete before the next begins — but each stage creates the substrate that makes the subsequent one possible.

A provocation worth defending: the tourism department is the most likely leader of this transformation in most mid-sized cities. Tourism departments understand revenue. They understand visitor experience. They understand the relationship between external demand and internal capacity. They are accustomed to working across organizational boundaries — with transport, with hospitality, with cultural institutions. And they are, in most cities, chronically underestimated as a source of urban intelligence.

This is less a prediction than a diagnostic. In cities where this transformation has advanced furthest, the initiative has consistently been anchored in departments that combine external accountability — they face real visitors, real revenues, real consequences — with cross-organizational reach. Tourism fits both criteria better than any other urban department.

The following cases are drawn from direct field engagement, published research, and documented consultations. They are presented not as proofs but as evidence — patterns that recur across geographies, scales, and political contexts, which any theoretical framework for urban intelligence must account for.

The digital twin, as coordination substrate, enables a form of civic participation that no previous urban technology has made possible: citizens can see the simulated consequences of decisions before those decisions are made. Not as projections in a planning document. As navigable spatial simulations, available to any resident with a smartphone.

The practical consequence of this is the prevention of what we might call the Barcelona problem: success metrics that ignore citizen experience. When residents can query the shared reality — show me what happens to my neighborhood if we add ten thousand visitors per month — and receive a simulation grounded in real infrastructure capacity, real transport loads, and real historical patterns, the conversation about what counts as success changes. It becomes impossible to claim that a tourism strategy is working well while the residents most affected by it are experiencing displacement, crowding, and degraded services.

The city's AI agent, in the mature stage of this progression, serves as a public interface for this participatory layer. Not a chatbot with scripted responses, but a genuinely knowledgeable system that can answer: “You are planning to visit the harbor district next Saturday. The weather will be overcast. Based on historical patterns, I expect 40,000 visitors in that zone between 11:00 and 16:00. Here are three alternative itineraries that would give you the same experience with significantly lower crowding. Shall I route you?”

Knowledge transfer is perhaps the most underappreciated benefit of the shared reality approach. Institutional memory, in conventional city management, resides in people. When those people leave — through retirement, career change, or political transition — decades of contextual knowledge leave with them. The digital twin, properly maintained, holds that knowledge in a form that is not dependent on any individual. The causal relationships between decisions and their outcomes are spatially and temporally encoded. The new arrival can navigate them. The institutional knowledge does not disappear when the institutional memory does.

Participatory approaches must also include dissenting voices — not as a democratic formality but as a methodological requirement. Anti-tourism activists, technology skeptics, and residents who have been burned by previous smart city promises carry knowledge about failure modes that enthusiastic adopters systematically underweight. A research and implementation methodology that does not include mechanisms for dissenting voices to be heard spatially and specifically will reproduce the errors it is trying to prevent.

This paper presents a framework derived from field engagement and practitioner observation. What it requires, to advance from framework to theory, is systematic empirical grounding. The following research agenda is proposed for the period 2026 through 2030.

Research teams embedded in active city transformation processes, participating in the Extreme Collaboration methodology rather than observing it from outside. The researcher is both instrument and participant — present in the room when stakeholders first encounter the shared map, present when the first cross-silo decisions are made, present when the AI copilot begins surfacing patterns. The knowledge produced by this proximity cannot be replicated by case study analysis or retrospective interviews.

The same cities measured at six-month intervals on the collaboration-to-autonomy progression, over a minimum three-year research window. Comparative framework spanning at least three European contexts — drawing on the existing network of cities engaged with OASC MIMs, augmented by specific selection criteria for small city representation (population thresholds consistent with established small city literature).

Participatory research methodology requires that the voices of those who will be most affected by the systems being designed are present in the design process — not as consultation formalities but as genuine co-producers of the research. This includes residents at risk of displacement, workers in sectors that will be disrupted by automation, and communities with historical reasons to distrust urban technology deployments. The SMILE methodology provides a structured approach to incorporating these perspectives without allowing the research process to be captured by the loudest or most organized voice in any given engagement.

Reproducibility is a non-negotiable research commitment. The SMILE methodology is the standardized implementation framework — documented with sufficient specificity that another research team, in another city, on another continent, could replicate the engagement protocol and produce comparable data. This is not merely academic. It is the only mechanism through which the knowledge produced by this research can scale to the cities that need it most.

The road to autonomous cognitive city operations does not begin with AI. It does not begin with sensors, platforms, or procurement. It does not begin with data. It begins with people in the same room, looking at the same reality, and being willing — for the first time, in many cases — to see what the person next to them sees.

Extreme Collaboration is the hardest engineering problem in urban intelligence. Not because the technology is difficult. The technology is, at this point, well ahead of the organizational capacity to use it. The hard problem is human: the trust required for a water department to act on tourist flow data, the political will required to share credit for outcomes, the institutional courage required to put the city's collective knowledge on a map and let everyone see the gaps.

The NUDED index gives cities a language for where they are — and, at its final level, a vision of where they can go: Distributed, sharing intelligence across city boundaries, learning from peers, teaching what they know. The SMILE methodology gives them the path. The PESTELED lens ensures they walk it with eyes open to the political, economic, ethical, and demographic forces shaping their context. And the shared reality substrate — the digital twin as boundary-spanning coordination medium — gives them the table to sit around while they make the journey.

Small cities will get there first. They have the structural advantage of small numbers, short distances, and decision-makers who already know each other's names. The question is not whether small cities can achieve cognitive urban operations. The question is whether large cities can learn from them before their own complexity becomes irrecoverable.

A city that knows itself — its history, its constraints, its latent capacities, its citizens' actual needs — is a city that can be governed with wisdom rather than approximation. The digital twin is not the city's mirror. It is the city's mind. And a mind, to function well, requires not just information but the integration of perspectives that have too long been kept apart.

The progression from collaboration to cognition is measurable, teachable, and replicable. Every city is already somewhere on that progression, whether or not it knows it. The work of this research agenda is to make the path visible — so that every city can find where it stands, see where it needs to go, and understand precisely what the next step requires.

Not a blueprint. A compass. The city already contains the direction. Our task is to help it find north.