At Fordje, we've started a weekly series putting our platform head-to-head with ChatGPT—using real-world city code scenarios that clean tech and construction companies are likely to run into.
The goal? Not to "win."
But to understand how each tool handles real-world complexity—and where we still need to improve.
Why We're Doing This
While our primary focus is on our Synthesizer Tool, we know our Research Tool still plays an important role for many users. These teardown exercises are our way of pressure-testing the platform, learning where we excel, and identifying what needs refinement.
They're also a reality check: AI tools are everywhere. We wanted to see how general-purpose AI compares to a product purpose-built for jurisdiction-specific code clarity.
The Scenario
In this week's teardown, we gave both Fordje and ChatGPT the same challenge: interpret a zoning and setback rule from a specific U.S. municipality—something a builder or developer might encounter early in the permitting process.
How ChatGPT Performed
ChatGPT did what it does best: deliver a confident, articulate response—fast.
But it made critical mistakes:
- It referenced outdated versions of local code
- It misinterpreted setback rules based on generalized assumptions
- It didn't surface any jurisdictional context or citations
In short: the answer sounded right but wasn't accurate. That's not a flaw of ChatGPT—it's a limitation of trying to apply general-purpose AI to highly localized, fragmented municipal data.
How Fordje Performed
Fordje fared better. Our platform:
- Correctly identified the applicable city, county, and state code layers
- Surfaced the relevant zoning classification
- Pointed directly to the appropriate setback requirement, with source links
- Provided clear, structured data points in seconds
Still, we're not claiming perfection. We saw areas where formatting and usability could be improved—and we're actively working on that. Trust is earned through clarity, not just correctness.
Why This Matters
We built Fordje because we've seen firsthand how messy, inconsistent, and hard-to-navigate city codes can stall clean tech and construction projects. Generic tools can't solve that alone.
These weekly teardowns are our way of staying honest, pressure-testing our assumptions, and improving in the open.
Watch the Teardown
We'll be sharing more of these breakdowns in the weeks ahead. If you're curious how Fordje works—or where it still needs work—this is a good place to start.
Have a code scenario you want us to test? We'd love to hear it.
