Democratizing the power to shape the future

Democracy is a process by which we decide our future. Before anyone decides, someone asks the questions, develops the alternatives and explores what each would do. Whoever can do that work shapes the options everyone else chooses from.

Today, that work mostly happens inside organizations that can pay for studies. Everyone else is invited to respond to the options they prepared. Widening access means more people can originate alternatives, examine consequences and bring work of their own into a conversation. That changes who can begin, as well as who can respond.

Cities also take shape through millions of everyday decisions: where to open a shop, how to organize deliveries, which home makes daily life possible, where to hold a market, what to build on a street. Most people making those decisions have never had access to simulation. Whathappensify exists for them. People who use it for their own decisions gain practice for public ones: a shopkeeper who has compared delivery routes can use the same skills when the city proposes closing their street to traffic.

The ambition is to redistribute foresight. Ask a question in your own words, have an AI assistant do the specialist work, and watch the possibilities unfold in your browser. Examine the reasoning. Bring what you know. Develop an alternative of your own.

The case rests on one change. A simulation that carries its basis with it stops being an authority to accept or reject. It becomes a statement someone must answer for, and others can answer it with work of their own. Repeated across many places, that could change who proposes a city's future and who benefits.

Who can ask the question and develop an alternative

A person who knows a place may see a possibility others have missed. A shopkeeper sees customers struggling to collect orders. A parent wonders whether a shared school ride could work. Neighbors who sweltered through the last heat wave wonder whether the library could open as a cooling room, and whether the people who need it most could reach it. Testing an idea like that can demand expertise and software far beyond their reach. Whathappensify puts that capacity in their hands, so more ideas get a chance before a lease is signed, money is committed or a plan is settled.

Small cities and towns are often in the same position. A town of twenty thousand rarely has a modeling team or the budget for a consultant's study, yet it decides on bus routes, parking rules, opening hours and new services just as a large city does. Its staff could test a policy or a new service before committing to it, asking what happens if the bus runs every twenty minutes instead of every forty, if market-day parking moves to the edge of the center, or if the town starts a shuttle for residents who no longer drive.

Personal enterprise, care for neighbors and collective action all belong here. The same person may be a business owner, renter, parent, pedestrian and volunteer. What they need to know changes with the question they are asking.

Explore what people could do

The capabilities converge

AI makes the work accessible through ordinary questions. Simulation lets you explore what might happen. Imagine arriving at a meeting with a proposal you have already explored. You can show what you changed, who it might help, and which assumptions deserve another look. Someone else can question your work and develop a different version.

That requires a way to check work an AI assistant helped produce. An assistant can explain a result fluently and still claim more than the calculation shows. The Epistemic Protocol Machine, or EPM, keeps a record of what each conclusion rests on and how strongly the evidence supports it. An official who doubts your proposal can check it against that record, and you can examine how a conclusion was reached and pursue the next question.

Perspectives make what matters to people explicit. A street change may look different to a wheelchair user, a delivery driver and a shopkeeper. Each can bring their concerns into the analysis and examine how those concerns become measures. A concern that no number captures stays in the record with its reasons.

Together, these capabilities let people explore a question on their own terms. You can challenge an assumption, distinguish a disagreement about evidence from a disagreement about priorities, and develop alternatives that respond to both.

Follow one question · Explore EPM · Understand Perspectives

Dialogue in place of decree

A simulation shown at a public meeting is a statement about the future. Someone chose the question, the assumptions, what to count and what to leave out. A traffic study that grades an intersection by how long cars wait has already decided whose time counts. The home care aide whose bus sits in that queue, and who reaches her client half an hour late, may not appear in it at all.

For most people, such a statement arrives as a decree. Residents who object are arguing about whose time counts, but their argument reaches the hearing as one about numbers, and only one side has numbers.

Whathappensify treats a simulation as a statement its author must answer for. A result carries a record of what it rests on, and the EPM labels it by the kind of work behind it, never by who paid for it. Anyone who disagrees with a published result can rebuild it to see whether it holds, read it through their own Perspective, or build their own version with the disputed assumption changed and publish it beside the original. They can also add evidence the original left out. The reply takes the same form as the original and is held to the same rules.

That would change the conversation: which number is right becomes a question anyone can check, and better data can settle a dispute about data. A dispute about priorities would show as one and could be argued in the open, which is the work democracy exists to do. Once a reply sits beside the original on the record, setting it aside is a choice made in public, one that reporters, appeals and the next election can reach.

The decision belongs to the people and institutions accountable for it, who must explain it. Software cannot make a hearing listen, but it could raise the cost of ignoring a reply. In one of the imagined stories, a tenants' association took its results to a council hearing, and the council added a requirement for affordable homes.

Simulation adds a means of participation, never a condition of it. People keep their standing in a decision whether or not they produce a model, and an objection based on experience or rights must be weighed on its own terms.

Anyone can make these replies, including a developer. The gain would be largest for whoever arrives with nothing: a neighborhood group would go from no analysis to analysis anyone can check.

How useful work accumulates

A community should be able to return to its work when a proposal changes. A researcher should be able to examine a method, reproduce a result and extend the question. A new business should be able to borrow an approach that helped someone else, using its own information and circumstances.

1. One group shares its Perspective

   Tenants near a planned rail stop found that rising rents could push out many of the people the stop was meant to benefit. They share the Perspective they used, which asks whether people living near the stop can stay and use it. Their lease letters stay private.

2. Another neighborhood applies it

   Tenants near a different planned stop apply the same Perspective to their own blocks, with their own rents and leases. They start from an approach another group has already used and spend their time on what is different about their neighborhood.

3. Later evidence improves the method

   After the first stop opens, a researcher compares what the simulation showed with what happened to rents and to who stayed. Where the two differ, the correction goes into the shared method, and the next group starts from a better version.

Each new version keeps a connection to the work it came from. People can see what changed, try an alternative and carry the lessons into another place.

Capacity can accumulate in people too. A group that answers one proposal keeps its Perspective, its data and the practice of building a reply, so the next answer costs less.

Shared simulations, prepared data, methods and perspectives can become a public resource. Local knowledge can travel with its source and context.

You decide what to share, within the permissions attached to the work. Private questions and private collaboration have a place in public infrastructure too.

The long-term difference

Most use would be private and ordinary: a caterer checking whether a street fair will block her van, a family deciding when to leave before a storm. Together, such questions could make asking what would happen as ordinary as checking the weather.

Climate change makes this more pressing. Heat waves, floods and storms arrive street by street, and the people who live there often see first what could help: a shaded walk to the bus, trees along the route to school, storm drains cleared before a downpour, a second route the ambulance can take when the low road floods. Ideas like these deserve a fair test before the next summer or storm, in a form the people who decide can check.

Institutions would feel the change next. Where a reply can arrive in a form anyone can check, an agency or a developer would have reason to publish its analysis from the start and to adjust a plan while changes are cheap.

That is how this work could make cities fairer, more equitable, more resilient and more inclusive:

• Fairer, when a burden is counted before a decision instead of discovered after it, such as the route a new depot's trucks will take past a school.
• More equitable, when people whom official data barely describe can add what they know, and a gap in the data counts as a gap, never as proof that nobody is affected.
• More resilient, when households and blocks can anticipate a flood or a heat wave on their own street and test their own ideas for meeting it, not only the agencies planning for the whole city.
• More inclusive, when a grandmother who cannot stand long at a bus stop and a street vendor can each bring their concern into the comparison of a plan, even a concern no number captures.

Software alone does none of this. A forecast can reveal a risk a family cannot afford to escape. Better conditions depend on organizing, law, funding and institutions willing to act.

Shared infrastructure that people can run themselves

Whathappensify's public purpose is to make this capacity widely available, including where a person's question offers little commercial return. United Cities Foundation provides it as free, open-source digital public infrastructure, with software that people can inspect, extend and operate independently.

A city could run it for residents, businesses, community groups, researchers and public servants, so they can explore ideas of their own and build on useful shared work. Because the software is open, a community group or a university can also run its own copy. Then no public decision would have to rest on a tool that one owner, the Foundation included, could withdraw, reprice or quietly change.

Communities need the capacity to question powerful institutions, carry their work from one decision to the next, and choose the purposes their tools serve. The long-term possibility is a public ability to explore and shape change, shared far beyond the organizations that can purchase it today.

Many services, one shared foundation

People reach Whathappensify through their AI assistant. They could also meet it inside services built for something they already do: a lender's guide to a renovation loan, a planning tool a small-business adviser uses, a tenants' association's toolkit, a city's program for new shops. Each is built by people who know whom it serves and what they ask. Many such services become possible when the hardest parts are shared.

Often the hardest part is the model: a careful account of how something works, written by someone who understands it. Much of what matters in a city builds up and runs down over time: money in a business account, orders waiting to be filled, homes waiting for repairs, people trained in a trade. Each affects the others, often after a delay. Someone who understands one of these can describe it once, with its assumptions, its evidence and the situations it does not fit, and many services can use it.

A housing cooperative plans to insulate and replace the heating in two hundred apartments, to cut energy bills and emissions. The plan depends on installers, and trained installers are scarce: training a new one takes most of a year. One specialist has described how trained workers build up in a trade and leave it. Another has described how a retrofit changes a household's bills. The cooperative's assistant combines the two to compare plans. The comparison could show that spending the first year training apprentices finishes the work sooner than starting every building at once with the installers available today, and leaves fewer families living with open walls through the winter.

The same workforce model could help a credit union design its retrofit loans, a college plan its courses and a city set a target it can meet. Each builds something of its own on top, and none has to rebuild the model.

Many kinds of contribution could meet here, and one organization may make several:

• Models from people who know how something works, with their assumptions, evidence and limits.
• Data, kept up to date, with evidence of how well a model has done in particular places.
• Services that turn shared parts into something useful for a particular group of people.
• Operation: running the platform for others and connecting it to their data.
• Testing: comparing models with what happened, and improving them.

Reuse is worth having only if limits travel with the work. A model keeps its name, its author, its assumptions and the situations it was built for wherever it is used. Writing an equation for a relationship does not show that the relationship holds. Joining two sound models adds an assumption about how they affect each other, and the evidence for each one does not vouch for the combination. The EPM governs what any conclusion may claim, whoever built the model, and Perspectives keep it clear whose concerns decide which consequences matter.

Organizations working for change already write down how they expect it to happen. A theory of change names what should shift, through which steps and over what time. Parts of it become an inspectable model whose consequences can be traced and compared, bringing out assumptions the written account leaves unstated. Some links cannot yet be measured responsibly. They stay in the record as open questions and are never filled with a convenient number.

Shared models make an assistant more useful too. Asked a question, it can look for evidence, find a model suited to it, set up the alternatives, compare them and explain what follows, starting from knowledge someone maintains instead of building every model from nothing.

This is also how the shared foundation can be paid for and improved. Services that earn money on it can help fund the infrastructure they rely on. Organizations that build on it can give back models, data connections, ways of testing and corrections. Public and philanthropic funding can then carry the same capacity to people and places that commercial demand would leave out.

Whether this works depends less on the engine than on the ordinary work around it: models documented well enough to use, easy to find and supported by evidence for the places they are used, results that fit together, and fair terms for the people who contribute.

How others build on the platform

How we would know it is working

Access is not use, and use is not change: a reply in front of a council is not a changed decision, and a changed decision is not a safer street. The signs to look for, roughly in the order they would appear:

• people asking about their own streets again, on their own initiative;
• a neighborhood's alternative appearing in the conditions or findings of a decision, not only in a list of comments received;
• a group reusing what it built for one proposal to answer the next;
• a model or Perspective that one organization built, in use in another's service;
• an agency publishing its analysis, with its basis, before anyone asks;
• predictions compared with what happened, in the open.

Warning signs matter as much: decisions moving to rooms where no analysis is shared, replies acknowledged and then ignored, and services that use a shared model while dropping its limits.

One test governs the rest. The change counts only if neighborhoods with the least access to paid analysis and the least organization gain the practical capacity to propose and to answer. If only well-organized neighborhoods got better at contesting plans, the effort would have failed, whatever the usage numbers say. These are signs to look for, not results to report.

Work with us

What would you explore if this capability were yours? We welcome conversations about the questions, places and possibilities that matter to you.

Contact Robert Kirkpatrick · See the technical work