_{Welcome to Transit Intelligence. Embrace discovery!}

I was catching up with a buddy who works in municipal government last week, and he mentioned they're rolling out a new AI policy. "Everyone has to use Copilot," he said, like this was a win. I nearly knocked my coffee over.

This is exactly backwards.

I've been deep in these tools for a few years now, and if there's one thing I've learned, it's that no single tool stays good for long. Gemini was incredible six months ago — now it's inconsistent. I stopped using ChatGPT seven months ago because the quality tanked. Claude has been my go-to lately, but that could change next month. The models are evolving so fast that any policy mandating a specific tool is obsolete before the ink dries.

Besides: your people are already finding their own paths. You know how urban planners design these beautiful paved sidewalks, and then everyone just walks across the grass anyway? Those worn dirt trails are called desire paths — and they're exactly what's happening with AI in your organization right now. Your frontier people, your curious ones, they're already exploring tools that aren't on your approved list. They're going around your policy because the policy doesn't serve them.

So you have a choice. You can keep paving sidewalks nobody uses and pretend you're in control. Or you can follow the desire paths and build something real.

The AI policy that actually works isn't a mandate. It's a green light with two guardrails.

That's it. That's the policy. Everything else is encouragement: go explore, try multiple tools, report back what you're learning, make mistakes, share them.

This matters more than you think. The culture you're building right now, in this very critical moment, is going to compound. If you build a culture of curiosity and permission, your staff will lean in. They'll follow your AI leaders. They'll talk to each other. That's a positive feedback loop that will carry you for years and launch you WAY ahead of the curve.

But if you lock everyone into one tool and call it a strategy, you're going to wake up two years from now wondering why you're behind. Your best people will have worked around you. Your cautious people will have learned nothing. And you'll have missed the one window where public transit could actually get ahead of the curve instead of chasing it.

For once, the cost of experimentation is nearly zero.

Now go exploring.

Transit Intelligence is where I share practical AI insights for transit leaders. Subscribe if you haven’t. Forward to a colleague who needs it. And hit reply — I read everything.