What Makes AI Actually Work in a Planning Department

There’s a lot of energy right now around AI in local government. And honestly, most of it is well-placed. The technology is genuinely impressive, the potential applications for planning and community development are real, and cities are paying attention. 

But there’s a question that doesn’t get asked enough in those conversations: what has to be there before AI can do its job?

Because the planning departments we see getting the most out of new technology aren’t necessarily the largest or the best-funded. They’re the ones that got clear on something first - what information they actually have, where it lives, and where it needs to go to do what it’s supposed to do.

That foundation may be less exciting than the tools sitting within it. But the system is the whole game.

The question underneath the question

When a resident calls to ask what they can build on their property, the answer exists somewhere. It’s in the municipal code, in the zoning map, in the overlay districts the city adopted three years ago. The information is there.

The problem, almost always, is access. Getting from the question to the answer requires someone who knows where to look, how to read what they find, and how to communicate it clearly. 

That’s not an AI problem. It’s an infrastructure problem.

We presented this analogy at the League of Oregon Cities Conference two weeks ago: imagine a modern engine dropped into 1800s America. Obviously it’s fascinating and powerful technology. But it can’t be utilized safely and efficiently unless it: a) gets the right fuel, b) is hooked up to a transmission, c) is supported on a chassis with wheels, steering, and brakes, and d) is travelling on well-maintained roads. 

What we’ve found working with Oregon cities is that when you solve the infrastructure problem first -  when zoning data is digitized, parcel-level, and tied directly to the adopted code, something shifts. Staff spend less time on routine lookups. Residents get answers faster. And AI has the infrastructure to safely and efficiently do a good job.

What this looks like in practice

In Coos Bay, North Bend, and Bandon, UrbanForm worked with the Southern Oregon Coast Regional Housing Collaborative to bring parcel-specific zoning information online across the region. That work is now expanding county-wide in partnership with Coos County and the Oregon Community Foundation. The goal was never technology for its own sake, it was finding tools that remove the friction.

In Lafayette, a Community Development Clerk opens UrbanForm every morning to answer resident questions and review site plans. That’s not a dramatic transformation story. It’s just a better workday - for her and for the residents she serves.

Those outcomes didn’t come from adopting AI. They came from finding partners like UrbanForm that can build the technological infrastructure to make things happen. AI is just one component of a well-run system; it can’t (or shouldn’t) work unilaterally.

The optimistic version of this story

Here’s what we find encouraging: the foundation isn’t that hard to build. Oregon cities don’t need enterprise budgets or dedicated innovation staff. They need accurate, well-structured zoning data and a clear sense of what questions they’re trying to answer.

When that’s in place, everything else gets easier- the AI tools, the public-facing services, the housing capacity analyses, the permit counter conversations. The technology works because the infrastructure beneath it works.