City managers who underestimate what the self-serve path requires often lose their executive sponsor before the first review cycle is complete.
This page is for city teams deciding whether to self-serve with the repo or ask for guided help. If you can assemble 5 to 10 existing staff, dedicate a project lead, and get an executive sponsor to back the work, you can use the repo directly. If those basics are not in place, expert help is the better starting point. The repo gives you the framework. Claude Code does the drafting. Your city still has to supply the buy-in.
Already partway through? Use Program Includes as a gap checklist to find what's still missing from your city's package.
Most cities do not fail at AI policy because they lacked good intentions. They fail because the work got spread across too many people, nobody owned the output, and the process dragged until the political window closed or the champion moved on.
The repo and Claude Code change the math on the drafting side. A risk framework, a review process, a workforce plan, and a community trust layer can be produced through focused project lead time with targeted input from a small team. The quality comes from the built-in city models. The speed comes from the fact that the agent drafts and your team decides.
What the repo cannot do is build the relationships. It cannot make your city attorney feel heard, get your union rep on board, or earn council trust. That work is human work. The repo helps you show up to those conversations with a credible draft instead of a vague concept and a long timeline.
These are not new hires. They are existing staff who need dedicated time and a clear ask. If you cannot name these roles, you probably want expert help before trying to run the repo on your own.
Sets direction, clears roadblocks, and makes sure the work has enough authority to move. Without this person, every review cycle becomes a negotiation and every decision gets deferred. This is the City Manager, CMO, or CIO who has made AI policy a named priority, not a side project. See strategic leadership for the decision templates and leadership working files.
Coordinates the team, runs the Claude Code sessions, and drives the repo workflow from intake to final draft. This is often the CAIO, CIO, or another designated project lead who opens the folder, runs the prompts, and turns stakeholder feedback into revised drafts. The work lives here. Everything else supports this role.
Reviews every draft for legal exposure, public records obligations, and local authority alignment. The repo flags the sections that need legal input, and this person resolves them. Without legal sign-off, the policy does not move. The rules and review framework describes the risk tiers and reviewer lanes legal sign-off is built around.
Keeps the work tied to how the city actually runs, including real service delivery, vendor relationships, and frontline practice. The operations module in the repo is built around this input. If this person is not at the table, the policy will not survive contact with the departments that have to follow it.
Translates the policy into staff expectations, manager guidance, and training requirements. Union notification obligations, employee review concerns, and HR policy alignment all run through this role. The repo's workforce module is built to surface exactly what this person needs to weigh in on.
Makes sure the city can explain what it is doing to residents, council, and staff who are nervous. The community trust module in the repo produces a resident-facing framework, but this person has to own the message and the timing. Public buy-in does not happen without a real communications strategy.
Key consideration
Reviews resident-facing and service delivery uses for discriminatory risk. The repo's risk classification framework flags high-risk uses that require equity review, and this person makes that call. A policy without this review is a liability waiting to surface.
This is not 80 hours spread across six months of bi-weekly meetings. It is focused, dedicated project lead time, with supporting roles stepping in for several hours each when decisions, review, or sign-off are needed. If that kind of commitment is not realistic, expert help will be a better fit.
Hours 1-20
The project lead forks the repo, opens Claude Code, and runs the strategic leadership and governance prompt packs. The output is a scoped policy draft covering risk tiers, review paths, and prohibited uses. Legal sees a first version during this first block, and the sponsor knows what decisions are coming.
Hours 21-55
The remaining three module families are drafted in working sessions. The workforce plan, operations playbook, and community trust framework come out of the repo's prompt packs. Stakeholders review at the points where their input matters. Revisions go back through Claude Code.
Hours 56-80
Multi-stakeholder review is routed using the review sprint kit. Feedback is incorporated. The council review packet is assembled. The project lead presents to leadership. The draft is ready for final review.
Calendar time will be longer. Public comment, council scheduling, and stakeholder availability all add time. The up-to-80-hours figure is the project lead's build time, not the waiting, and not the several hours other roles will contribute.
These are the deliverables the repo is built to produce. Each one comes from a specific module, template, and prompt pack.
Cities at 40–70% completion usually have a sponsor named and a draft started — but are missing one or more pieces from the list above. Use the checklist to identify what your city has, then go directly to the page that covers the missing work.
If your city has a draft but has not built the risk tiers, reviewer lanes, or escalation process yet.
Open rules and review →If the policy exists but managers and staff do not have a clear path for daily use, escalation, or uncertain requests.
Open workforce training →If the city cannot yet explain its AI use to residents, council, or the press in plain language.
Open public transparency →If the policy will launch without a clear owner, update trigger, or plan for keeping it current after approval.
Open strategic leadership →Use the Program Includes checklist to compare what your city has against the full six-part package.
Open program includes →The repo gets you to a credible draft faster than any other starting point, built from best-practice city models and reviewed layer by layer. But the buy-in is a different kind of work. Getting the city attorney invested. Getting the union rep's questions answered before they become objections. Getting the City Manager to name this as a priority in a budget season full of priorities. No repo does that. Your team does that.
The toolkit speeds up the build. Your city still has to carry the buy-in.