Load your city's zoning and building code into ChatGPT, then ask for the unit count on an infill lot by address before you reach the property
Andy Choi of Urban Equity Holdings, a Los Angeles infill multifamily and mixed-use firm, used to wait one to two weeks for an architect's feasibility report to learn how many units a lot could hold. His team first replaced that report with a custom Excel sheet, then moved the code itself into the model: they uploaded thousands of PDFs of Los Angeles City and Los Angeles County building code and zoning code into a ChatGPT model and refresh that code library in the first quarter of each year when new California code editions come out. The query is now an address, asked by voice from the car. The model returns what is buildable by right, what Transit Oriented Communities bonus density adds if the site qualifies, and whether ADUs are available, and an image model adds a rough envelope with height limits, a rough site plan, and schematic floor plans. Choi says being off by one unit is roughly $30,000 of NOI. He claims the unit counts have been accurate across hundreds of case studies; the publication did not audit that claim and notes that the same prompt returns different answers in ChatGPT, in Claude, and in a fresh window. His firm's control is two questions on every AI-assisted deliverable: what prompt produced this, and did you QA the response.
At a glance
- Tools used
- ChatGPT
- Category
- Investing & underwriting
- Audience
- investor
- Difficulty
- No-code setup
- Source
- The Two-Week Feasibility Study That Now Runs From Andy Choi's Car
AI for CRE Collective (Editorial Desk) · Article · Aug 4, 2026 - First step
- Download the current zoning code and building code PDFs for the one city where you buy most often, and note the edition year on each file. Tip (our recommendation, prompted by a reader question on X, not a step in the AI for CRE Collective article): the by-address answer from the zoning code library is a first screen, not a go decision. Before anyone puts money on the lot, check the overlay district map, recorded easements from the title report, and a will-serve letter from the water and sewer utility.
Compliance note: A model's feasibility answer is a screening estimate, not a zoning determination. Confirm the unit count with an architect or the planning department before you price an offer, and keep the prompt and a reviewer's sign-off with every AI-assisted feasibility note.
Suggested prompt
Written by this directory, inspired by the source. Replace the [bracketed] parts.
You have the attached zoning code and building code for [city/county], [edition year]. Answer only from these attached documents, and cite the section number for every rule you apply. For the lot at [address], parcel [APN], zone [zone if known], lot size [square feet]: 1) list the by-right density, height limit, setbacks, floor area ratio, and parking requirement; 2) calculate the maximum by-right unit count and show the arithmetic; 3) state whether any density bonus program in the attached code could apply, what the bonus unit count would be, and which facts I must confirm to qualify; 4) state whether ADUs are allowed and how many. If a needed fact is missing or the code is ambiguous, say so instead of assuming. End with a list of every assumption you made.
Get 50 AI Prompts for Real Estate Agents, free
Subscribe to the Daily 5 and download the 48-page PDF right away. The prompts help with:
- Getting more leads from the contacts already in your CRM
- Faster follow-up, so no text or inquiry goes unanswered
- Less time on listing descriptions, paperwork, and showing instructions
- Copy-paste prompts tested by working agents, brokers, and landlords
- Five new real-world AI use cases every weekday morning in the Daily 5
Confirm by email. Unsubscribe in one click. Review every AI draft for accuracy and fair housing compliance before you send it.