Milwaukee City Plan Commissioner Uses AI to Turn 700 Pages of Zoning Code Into a Single Conversation — Built Without a Development Team
Non-technical civic leader builds functioning prototype with Claude Code and Google Gemini 3, enters Google's global hackathon to demonstrate what's possible for city services.
MILWAUKEE, WI — February 10, 2026 — Tarik Moody, a City Plan Commissioner for the City of Milwaukee and Director of Innovation and Strategy at Radio Milwaukee, has developed MKEdev, an AI-powered city development intelligence platform that consolidates Milwaukee’s zoning codes, permit forms, GIS data, and planning documents into a single interface. The functional prototype was built in four weeks using Anthropic’s Claude Code—an AI-powered coding tool—and entered into the Google Gemini 3 Hackathon, a global competition showcasing applications built on Google’s latest artificial intelligence models.
[inarticleaad]What makes the project remarkable is that Moody is not a professional software developer. He describes himself as someone with “just enough technical knowledge to be dangerous” — a non-technical builder who leveraged AI tools to bring a civic vision to life without a development team.
“Serving on the plan commission, I see how complex the zoning process is — even for experienced developers,” said Moody. “For everyday residents trying to figure out what they can build on their property, it’s even harder. The information is public, but it’s buried across multiple city webpages, hundreds of pages of code, and a GIS system not designed for everyday users. I knew AI could make this more accessible. I just needed the tools to build it.”
Built With Claude Code
Moody built the entire platform using Claude Code, Anthropic’s AI-powered development environment, enabling users to build complex applications through natural language conversation. Rather than writing every line of code by hand, Moody directed the architecture, design decisions, and product vision while Claude Code handled the implementation — a workflow that enabled a single non-technical builder to produce a platform that would typically require a full development team.
“Claude Code changed what’s possible for people like me,” said Moody. “I understand the problem deeply. I understand what the solution needs to look like. Claude Code let me bridge the gap between vision and execution.”
The platform runs on Next.js, Convex, Mapbox, and Google’s Gemini 3 AI models, with a codebase spanning thousands of lines of TypeScript across a professional-grade monorepo architecture.
How the AI Works — In Plain English
MKEdev uses two types of artificial intelligence from Google, both under the Gemini 3 family:
- An AI assistant that reads and reasons. The platform has been “trained” on Milwaukee’s entire zoning code, 13 neighborhood plans, and dozens of city policy documents. When a user types a question — like “Can I open a restaurant at this address?” — the AI doesn’t just search for keywords. It reads the question, figures out what information it needs, and then takes a series of steps on its own: it looks up the address on a map, checks the zoning district from the city’s GIS system, finds the relevant parking requirements in the zoning code, and pulls up the correct permit forms. It does in seconds what would take a person hours of clicking through websites and reading PDFs.
- An AI that generates images. When a user wants to see what could be built on a vacant or parking lot, the platform uses a separate AI model to generate photorealistic images. What sets MKEdev apart from generic AI image generators is that it feeds the AI the actual zoning rules for that specific parcel—how tall the building can be, how far it must sit from the property line, and how much of the lot it can cover. The result is an architectural rendering that respects the real regulations, not just a pretty picture.
What It Does
A user can ask “I want to open a restaurant at 2000 N Dr. Martin Luther King Jr. Drive — what’s the zoning, how many parking spaces do I need, and what permits should I file?” and the system will look up the address, check the zoning district, calculate parking requirements, search the actual zoning code, and recommend the appropriate permit forms — all in a single interaction.
The platform also features:
- Document Analyzer — Upload site plans, floor plans, or variance applications and receive a plain-language compliance report: does this project meet the zoning requirements, where does it fall short, and what approvals would be needed.
- Site Visualizer — Generates photorealistic images of proposed buildings on any parcel in the city, constrained by that parcel’s actual zoning rules.
- Interactive 3D Map — Displays eight layers of Milwaukee city data covering zoning districts, parcels, TIF districts, opportunity zones, historic districts, and city-owned properties, with over 313,000 features.
- 24 Interactive Cards — Instead of walls of text, answers appear as visual cards showing zoning info, property details, homes for sale, permit forms, compliance reports, and more.
The platform has ingested 42 city planning documents, including all 12 subchapters of Milwaukee’s zoning code (Chapter 295), 13 neighborhood area plans, and housing incentive program documentation.
A Prototype Illustrating What’s Possible
Moody emphasizes that MKEdev is a functioning prototype designed to illustrate the potential of AI-powered civic services, not a replacement for the city’s existing systems.
“This is about showing what’s possible,” said Moody. “The data is already public. The zoning code is already public. The GIS layers are already public. AI gives us the ability to make all of that information genuinely accessible to the people it affects most — homeowners, small developers, and residents who can’t afford to hire a zoning consultant.”
Hackathon Track Record
Moody is no stranger to competitive hackathons. He previously won the Grand Prize at the LiquidAI Championship Hackathon for Hakivo, a congressional tracking application. MKEdev represents his latest effort to apply emerging AI technology to real-world civic challenges.
About the Google Gemini 3 Hackathon
The Gemini 3 Hackathon is a global competition hosted by Google, challenging developers to build innovative applications using Gemini 3, Google’s most advanced family of AI models. Projects are judged on technical execution, potential impact, innovation, and presentation.
About Tarik Moody
Tarik Moody serves as a City Plan Commissioner for the City of Milwaukee, where he reviews and votes on development proposals and zoning changes submitted to the commission. He is the Director of Innovation and Strategy at Radio Milwaukee and a recent graduate of the 2024-2025 cohort of LISC Milwaukee‘s ACRE program (Associates in Commercial Real Estate), which trains emerging leaders in commercial real estate development. His experience on the commission — observing how difficult it is for applicants and residents to navigate zoning information — inspired MKEdev.
Links:
NOTE: This press release was submitted to Urban Milwaukee and was not written by an Urban Milwaukee writer. While it is believed to be reliable, Urban Milwaukee does not guarantee its accuracy or completeness.
