Unmanaged AI use exposes cities to legal liability, service inconsistency, and resident complaints before any policy is in place.

How this work moves from concern to finished city materials.

This page shows the shortest path from AI concern to finished city materials. It answers three questions: who owns the work, what gets produced, and where review happens before anything goes live.

See the starting path See the review path

Starting path: best for a city that needs the first team, first sequence, and first outputs.

Review path: best for a city that already has tools in use and needs legal, procurement, leadership, or resident-impact triggers.

Five stages, five reviewable outputs

Each stage ends with something a city leader, CIO, or CAIO can actually review: a list, draft, review path, guidance package, or named owner.

These five stages describe how the work moves forward. For a complete list of what the finished program includes, see Program Includes.

Stage 1: Gather local inputs

Pull together what is already happening so the work starts from local reality instead of assumptions.

  • Usually led by: a project lead with IT, legal, procurement, and operations input.
  • Output: a working inventory of current tool use, existing policies, vendor questions, and staff concerns.

Stage 2: Draft the core rules

Write the plain-language rules your city is actually willing to use, review, and revise.

  • Usually led by: the project lead with legal and leadership review.
  • Output: a draft policy covering approved uses, prohibited uses, records, oversight, and escalation.

Stage 3: Build the review path

Define what kinds of AI work trigger extra review before anything is approved or deployed.

  • Usually led by: CIO, CAIO, legal, procurement, and leadership together.
  • Output: a review path with triggers for legal, procurement, leadership, and resident-impact checks.

Stage 4: Prepare staff and public materials

Translate the policy into something staff can use and residents can understand.

  • Usually led by: the project lead with HR, communications, and operations support.
  • Output: staff guidance, training material, public explanation, and notices or FAQs if needed.

Stage 5: Confirm ownership and rollout

Approval is not the end of the work. Someone has to keep the policy current once it is live.

  • Usually led by: an executive sponsor and named program owner.
  • Output: a clear owner, rollout plan, update responsibility, and a place for questions or exceptions to go after launch.

Why a city can enter the process in the middle

Not every city starts at the same point. Some need leadership decisions first. Some already have tools in use and need review triggers. Some need public explanation. The site keeps those paths connected so your team can start where the pressure is highest instead of pretending everything begins at step one.

What helps this hold up under review

The point is not just to look organized. The point is to give leadership, legal, and the public something that can hold up when questions come in.

  • Draft before polish
  • Review before approval
  • Training before rollout

After approval: keeping the policy current

The five stages get the city to an approved package. Keeping it current is a separate commitment — not a new project, but a named responsibility the city carries after launch.

  • Review as your city's capacity allows — but treat two years as the outer limit. The AI landscape changes fast enough that a two-year-old policy may not reflect current tools, risks, or practice.
  • The named owner from Stage 5 should know what triggers an earlier review: new tool categories deployed, a vendor change or addition, a public complaint or incident, a significant change in law or regulation, or a service delivery shift that brings new AI uses into scope.
  • When the policy goes back into review, document what changed and why — so the city can show its reasoning, not just its current rules.

What to open next

Once you understand the path, go to the page that matches your immediate problem instead of reading everything in order.

  • Start From Zero if your city needs the first working team, first sequence, and first outputs.
  • Rules and Review if you already have tools in use and need the legal, procurement, leadership, or resident-impact triggers.
  • Toolkit if someone on your team needs the repo entry page, docs, examples, and working files in the bundle.
  • Get Help if the local context is complicated, politically exposed, or cross-departmental.

Once the sequence is clear, move into the repo or use guided help.

This page should explain the flow. The repo is where your city actually drafts the policy package, supporting materials, and review outputs. If the local path is still complicated, politically exposed, or spread across too many departments, use the help path here.

Open the working files repo

Use the repo when your team is ready to move from sequence and explanation into the actual files, prompts, and templates.

Open the working files repo

Get help with the local handoff

If your city understands the stages but still needs help structuring the handoff, the reviewers, or the first sprint, get help here.

Get help