Governments in the AI era

The Government AI Playbook, Part 3: Make Participation Something Citizens Want

Cari · 2026-09-24 · 7 min read

TL;DR: Most civic engagement fails because it asks citizens to attend, submit, and wait. Participation works when it is structured like the things people already choose to do: clear missions, visible progress, and real recognition. Cari's civic missions turn participation into measurable civic work that citizens can complete from their phones, and every completed mission builds trust in the next digital service.

In Part 1 we argued that government AI should start with the queue, the visible backlog that citizens actually feel. In Part 2 we made the case for credentials before chatbots, because verified identity is the foundation every other digital service stands on. Part 3 is about what happens once those foundations exist: getting citizens to show up.

Why traditional engagement fails

Look at how most governments ask citizens to participate. A town hall meeting on a weekday evening. A public comment period buried on a ministry website. A survey that arrives by post and disappears into a filing cabinet. A consultation that closes, produces a report, and is never heard from again.

Each of these formats asks the citizen to do three things: attend, submit, and wait. Attend at a time and place the government chooses. Submit into a process they cannot see. Wait for an outcome that may never be communicated.

Nobody designs consumer products this way, because nobody would use them. Yet we act surprised when the same citizens who spend hours a day on their phones do not show up to a 6 p.m. consultation about zoning.

The problem is not apathy. The problem is design. Citizens participate constantly in systems that respect their time: they review restaurants, contribute to open maps, answer questions in community forums, and report problems in apps that show them the fix. People will do civic work. They will not do civic waiting.

What people already choose to do

The activities people volunteer for share a common structure. The task is concrete and bounded: you know what done looks like before you start. Progress is visible: you can see what you have contributed and what it added up to. And the contribution is acknowledged: by the system, and often by other people.

None of this is exotic. It is simply how humans decide whether an activity is worth their time. Civic participation has to compete for attention with everything else on a citizen's phone. It wins when it offers the same clarity, and it loses when it offers a form and silence.

Civic missions: participation as real work

This is the model behind Cari's civic missions. Instead of inviting citizens to comment on abstractions, governments publish concrete missions: specific pieces of civic work that a citizen can complete from their phone.

Missions fall into a few natural types:

• Report. Document something the government needs eyes on: a broken streetlight, an invasive plant, a flooded road, a damaged sign. The citizen becomes a sensor network the state could never afford to build.
• Verify. Confirm information the government holds: is this business still operating at this address, is this facility accessible, is this record accurate. Verification missions turn stale registries into living ones.
• Contribute. Add something new: a photo of a landmark, a translation, local knowledge that only residents have. Contribution missions build public assets.
• Learn. Complete a short course on disaster preparedness, digital safety, or a new public service, and be recognized for it. An informed citizen is a public good in itself.

Each mission has a clear scope, a clear completion state, and a clear owner inside government who receives the result. When a citizen completes one, the completion is acknowledged and recorded. Over time, a citizen's history of completed missions becomes a portfolio of real civic contribution, tied to the verified identity from Part 2.

That last point matters. Participation stops being an occasional consultation and becomes measurable civic work. A ministry can see not just how many people attended a meeting, but how many missions were completed, in which districts, on which problems. Engagement becomes an operational capability rather than a communications exercise.

Four design principles

Getting this right is mostly a matter of discipline. Four principles keep missions honest.

Missions must be real work that matters. If completing a mission changes nothing, citizens will notice, and they will be right to stop. Every mission should map to a decision, a dataset, or a fix that someone in government actually uses. Engagement theater, where the activity exists to produce participation statistics, is worse than no program at all, because it teaches citizens that their effort is decorative.

Completion must be acknowledged. The single cheapest thing a government can do is close the loop. You reported the pothole; here is the work order; here is the repair. Acknowledgment is not a courtesy. It is the mechanism that makes the next mission worth doing.

Progress must be visible. Citizens should see their own contribution history, and communities should see their collective progress. A district that can watch itself verify its business registry or map its drainage problems has something no consultation ever produced: shared, visible momentum.

Privacy is not negotiable. Participation is opt-in, always. A mission platform is a channel citizens choose to use, never a surveillance instrument. Citizens control what they share, contribution data is used for the stated purpose, and choosing not to participate carries no penalty. The moment participation feels compulsory or watched, the entire model collapses.

The trust flywheel

Here is why this belongs in an AI playbook at all. The scarcest resource in digital government is not budget or technology. It is trust.

A citizen who completes a mission and sees the result has just run a small, personal experiment on their government: I did something, and the state responded. That citizen is measurably more likely to try the next digital service, to register the credential, to use the AI assistant, to believe the queue from Part 1 is actually moving.

Trust compounds. Each closed loop makes the next service launch easier. Governments that skip participation and jump straight to sophisticated AI services are asking citizens to trust a system they have never seen respond to them. Governments that build the flywheel first find that adoption of everything else gets cheaper.

An honest caution about gamification

It is tempting to read this and reach for points, badges, and leaderboards. Be careful. Gamification done cynically backfires badly. Citizens are adults; they can tell the difference between recognition for genuine contribution and a dopamine trick designed to farm their attention.

The rule is simple: recognition must map to genuine contribution. If a badge represents fifty verified reports that improved a public dataset, it means something, and citizens will value it. If it represents opening the app seven days in a row, it is noise, and it cheapens everything around it. Reward the work, not the engagement metric. When in doubt, make the recognition quieter and the impact louder.

Common questions

Does this replace traditional consultation? No. Some decisions genuinely need deliberation, hearings, and formal comment. Missions complement those processes by keeping citizens engaged between the big moments, and by giving governments a continuous signal instead of an annual one.

What stops people from submitting junk? The same things that keep open mapping and community platforms healthy: verified identity from Part 2, cross-checking between contributors, and recognition that rewards accuracy over volume. Missions tied to real identity attract real effort.

Is this expensive to run? The platform cost is modest compared to almost any traditional engagement program. The real investment is organizational: every mission needs an owner in government who will use the results and close the loop. That discipline, not the software, is the hard part.

Participation is not a communications problem to be solved with better posters. It is a product to be designed, and the governments that design it well will find citizens ready to do the work. If you are exploring what civic missions could look like in your country, start the conversation here.