The mandate
to imagine.
Shifting from participation to collaboration — inside a city’s own planning and innovation teams, and with the people who live there. CoPlanAI is the link that lets those two worlds talk, so transformation arrives sooner, and stays affordable.
Cities are asked to deliver the future — but rarely allowed to imagine it.
The biggest constraint in urban transformation isn’t technical. It’s institutional: cities are handed ever-larger mandates to act on the future, and almost none to picture it first.
Cities have to change fast — the pressure has never been higher.
Climate adaptation, social cohesion, economic resilience — met with a decade of stacking frameworks, from the Paris Agreement to the EU Green Deal. Cities have answered at a scale that didn’t exist ten years ago, and participation itself is now a legal duty across Europe.
A decade of escalating commitments, 2013–2023. National adaptation strategies rose from 20 to 32; Covenant of Mayors signatories from zero to over 4,500. The mandate to deliver the future keeps growing; the mandate to imagine it does not. · Source: European Environment Agency.
Yet the room to imagine — and to act together — keeps shrinking.
Three institutional fractures block the very change cities are asked to deliver.
No mandate to imagine
Planning law is built to administer — zone, permit, archive — never to picture and choose between futures. Participation arrives late and narrow.
No capacity to match the urgency
Overloaded, rule-bound offices and thinning planner numbers leave no room to think forward.
A fraying common ground
Algorithmic silos and eroding trust outside; departmental silos and tightening budgets within.
A mandate that is missing, an urgency that is mounting, a unity that is fraying. This is the gap CoPlanAI is built for.
What cities ask for — and what we build instead.
Requests reach us in the language of automation. Underneath each is a better use case — the one we actually deliver.
What CoPlanAI is — and is not.
✓What it is
- ✓A participatory-planning platform — it structures how many people propose, compare, deliberate and converge on futures for a real place, in the room and online.
- ✓A shared visual language — a resident’s phrase becomes an image in seconds, so expertise is no longer the price of entry.
- ✓A way to produce a mandate — documented evidence of what a community can agree to want.
✗What it is not
- ✗Not an urban-design tool. No masterplans, no CAD, no technical drawings. The images are scenarios to think with — indicative, never construction-ready.
- ✗Not a render service. If the decision is already made, we’re the wrong tool.
- ✗Not automation of the public. AI never holds the pen alone; every generative step is wrapped in a human step.
CoPlanAI produces mandates, not masterplans. The masterplan stays your job — now with evidence of what people actually want behind it.
From a photo to a built street.
No survey, no zoning map, no jargon at the entrance. One loop — sense, decide, act — shown here in Helsinki Summer Streets, with the City of Helsinki.

A photo of a real street — the only input citizens need to start.

A citizen committee and entrepreneurs generate, debate and vote on scenarios together.

Helsinki Summer Streets, delivered — the loop closed on the real street.
The ballot box is the wrong template for design.
You choose once, blind, and live with it. Take a different Helsinki street — the Esplanadi cycle-lane experiment: we’ve all seen the same sequence play out — participation, then discontent, then demolition.



Democracy isn’t just about choice. It’s about rethinking your choice.
“The value of participation is not in deciding what most people want, but in discovering what we could want, together.”
In Helsinki, first instincts — bike lanes, everywhere — evolved once people could compare consequences side by side. That evolution is the democratic act.
The Hippocratic Oath of working with AI.
Bringing AI into a planning organisation is not a tooling decision — it transforms the institution. So the institution owes itself an explicit ethic. Ours, in four commitments.
Augment judgment — never automate it away
Agents cluster, theme and summarise. Officials review the synthesis instead of compiling it — the clerical work goes, the judgment stays.
Preserve the right to know and to change
Every participant can see consequences and revise their choice in public — the loop the whole method is built on.
Make the model’s biases legible
Before asking anyone to trust its images, we show where the model leans — and let the room correct it.
Represent those who cannot be in the room
Children, future residents, ecosystems — voiced by tuned agents, never confused with the living public.
A reparative technology.
The deepest repair is to participation itself — which has a default participant (adult, articulate, free on a weekday evening) while everyone else is structurally absent. Our agents can be tuned to centre the people, and the species, planning routinely overlooks.

A child-specific participatory AI — with an Architects’ Council of Europe grant, piloted in Finnish schools and at Milan Digital Week.

With UNDP Accelerator Labs Panama, students turned photos of their own waterfront into climate-adaptation scenarios they could own.

At BiodiverseCity? (Estonian Museum of Natural History, Tallinn) a tuned model let children reimagine communal areas as wilder habitats.
Imagination, evaluated against policy.
The newest layer answers the question every authority asks next: fine, but is it compliant — and is it good? Agents score co-designed scenarios against EU policy, local design guides and municipal strategy, while professionals keep the final word.

The scenario, assessed
A generated image is easy to love and easy to dismiss. Switch to Evaluated to see the assessment that makes it usable in planning.
Bologna’s own catalogue of permeable, vegetal and mineral materials becomes the evaluation criteria — checked while imagination is still cheap to change. Also in evaluation: Vienna Supergrätzl and Dubai Municipality. · Illustrative scoring.
Already in the field — and it travels.

3,000+ ideas, ~300 votes reimagining the space around the Siegestor.

An interactive play area and shaded cycling track — with planning law in the loop.

Collective imagination feeding directly into the municipal strategy document.

Participatory AI at the most consequential scale: a city that does not exist yet.

A UNDP-led regeneration with the mayor’s office at the table.

Seventh Hill’s Connect East Cleveland — community-led co-design, from the first photograph.
Behind the room, you stay in control.
Participants see no software to learn — just a photo and a question. Organisers configure the session before it starts, then watch and steer it as it runs.
Pick where images come from — upload, library or map — and how participants reshape them: theme, draw or prompt.
From first call to a mandate — in about two weeks.
Every engagement follows the same arc, built around your institution. You bring one real question, a place or a plan, and the people. We handle configuration, training, facilitation and the technical translation.
What the office keeps: a documented mandate, a configured, re-runnable platform, and a trained team — not a report that ages on a shelf.
Citizens propose.
The city responds.
Cities are asked to deliver the future; CoPlanAI gives them, and the people they serve, the instrument to imagine it in public — and the evidence to act on it. Start with one real question, in one office, and build from there.