AI + Leadership

AI Governance for Local Government

AI adoption in local government is accelerating. The governance question has not kept up.

The Governance Gap

Every municipality faces the same tension: the pressure to adopt AI is real, the risks of adopting it without structure are real, and the anxiety about getting it wrong, or falling behind, is real at every stage of adoption.

That tension exists at every stage. Leaders who have not started feel the pressure to begin. Leaders well into implementation feel the pressure of others moving faster. Leaders who built sophisticated AI capability still hold governance questions they have not resolved. The common thread is not the organization’s place on the adoption curve. It is whether a framework exists for the decisions the technology keeps requiring.

Technology vendors have shaped most of the AI conversation in local government. The questions they lead with, which tools, which platforms, which models, matter. They are second-order questions.

The first-order question is the one that governs any organizational change: what is the organization protecting, and do its systems enforce that?

Government leaders face constraints that make governance especially critical. Staff using cloud AI tools may be creating FOIA exposure without knowing it. Vendor-controlled platforms can change models, pricing, and capabilities without notice. Institutional knowledge built on a vendor’s system is knowledge the organization does not fully control. And in many organizations, employees are already using AI tools informally, without clear policy, without visibility, and without alignment to organizational priorities.

These are not technology problems. They are governance problems, and they take two things: a framework for deciding what AI should do, and a technical implementation that enforces those decisions.

Where The Shift That Sticks Meets AI

The Shift That Sticks was not designed for AI. It was designed for leading change when the ground keeps moving, which is exactly what AI governance requires.

Face the Truth asks what is happening with AI in the organization right now. Not the vendor pitch, not the fear, but the reality: which staff are using AI tools, what data is leaving the network, and where the governance gaps are.

Define What Matters asks what the organization protects regardless of which tools emerge. Data sovereignty, staff trust, institutional knowledge, public accountability: which of these is non-negotiable, and do current practices enforce it?

Make It Real builds the governance systems that keep AI inside the boundaries leadership has defined. Not theoretical policies. Working systems with clear rules, appropriate access controls, and accountability structures.

Reinforce the Change keeps those structures standing as the technology evolves. The tools will keep changing. The question of what the organization protects does not.

What Simon Sinek calls the Circle of Safety, the boundary inside which people can focus on the work rather than on protecting themselves, gets harder to hold when AI brings uncertainty about roles, data, and institutional knowledge. AI mirrors the framework wrapped around it. An organization with clear values and strong governance deploys AI in line with those values. An organization without that clarity finds AI amplifying whatever patterns already exist, including the ones its leaders would rather change. Governance holds the boundary as the technology shifts.

This is what Imagine That Performance is building. Not AI tools. A governance framework, grounded in a proven method, that helps local government leaders adopt AI in line with what their organizations value.

Four questions a leader can put to the team this month

Most governance gaps are not visible from the top. They show up in the questions a leader has not yet had a reason to ask. These four are a starting point, and each one belongs on the leadership team’s next agenda.

Which staff members are using AI tools today? In most organizations, employees are already experimenting with AI on their own. Research calls them “secret cyborgs”: people using AI to get more done and not telling leadership, often because no policy makes it safe to say so.

What data is passing through those tools? Every cloud AI interaction sends organizational data to an outside server. In government, that can mean constituent information, internal deliberations, or draft documents flowing through systems the organization does not control and may not be able to account for under a public records request.

Is there a policy that makes AI experimentation safe to talk about? When staff cannot discuss how they are using AI, the organization loses the ability to learn from what works, prevent what is risky, and build shared standards. The silence is the governance gap.

Could the organization answer a FOIA request about its AI use today? An uncertain answer marks the distance between where the organization stands and where its governance needs to be. The condition changed faster than the structures around it.

An unsettled answer to any of the four is the point where the right conversation matters more than the right tool. The next step is to put the four questions to the leadership team, write down the answers, and bring them to peers who hold the same seat.

Join the Conversation

Governance questions do not resolve in isolation. They move faster when thought through out loud with peers who face the same constraints: the same FOIA exposure, the same political dynamics, the same pressure to adopt without a plan.

Imagine That Performance convenes small virtual sessions for city managers, county administrators, and village administrators working through what AI governance means for their organizations. Each session is confidential, facilitated with the method behind The Shift That Sticks, and open at any stage of adoption. Sessions are set as interest comes in, and this page is where that interest is collected.

A working conversation about what matters when the technology keeps changing.

AI Governance Conversation

Send it to Rob. A note is enough to start. Say where the organization stands with AI and what is on the desk, and everyone who writes hears first when a session is set.

A note here goes to Rob Duncan and stays with Imagine That Performance. What comes back is a session date, or a question from Rob.