AutoSitu Uses AI Agents to Review Development Plans That Take Cities Weeks to Process

The Macro: Getting a Building Permit Should Not Take Months

Every real estate development in the United States starts with a permit. And every permit starts with a plan review. City planning departments need to check that a proposed development complies with zoning codes, setback requirements, height limits, parking ratios, lot coverage rules, stormwater management standards, and dozens of other regulations.

This review process takes weeks to months. Planning staff are overworked and understaffed. The regulations are complex, sometimes contradictory, and different for every jurisdiction. A single site plan might need review from zoning, fire, public works, transportation, and utilities departments, each checking against their own set of codes.

The cost of slow plan review is enormous. Developers pay carrying costs on land and construction loans while waiting for approvals. Housing projects get delayed. Commercial developments miss market windows. The National Association of Home Builders estimates that regulations add over $93,000 to the cost of an average new home.

AutoSitu, backed by Y Combinator, builds AI agents that review development plans against local codes, completing in minutes what takes human reviewers weeks.

The Micro: AI Planners Working Across City Departments

Asher Lin and George Zhai cofounded AutoSitu with complementary backgrounds. Asher studied urban planning and architecture at the University of Michigan, bringing domain expertise in zoning and compliance. George comes from Georgia Tech with a focus on autonomous systems and AI engineering. The combination of someone who understands planning codes and someone who can build the AI to interpret them is exactly right.

The product handles cross-department review workflows. AI agents analyze site plans against local zoning codes and building ordinances, flagging non-compliance issues and providing guidance. The system handles PDFs today and is rolling out support for CAD formats (DWG, DXF) and BIM models.

The speed claims are significant: up to 90% faster reviews with 80% cost reduction. If a review that takes a human planner three weeks can be done in hours, cities can process more applications with existing staff. The system positions itself as a copilot for planners rather than a replacement, which is the right framing for government sales where job displacement concerns can kill deals.

The customer list is impressive for a startup: partnerships with San Jose, Seattle, Boston, Los Angeles, San Francisco, Sterling Heights, and NVIDIA. Having multiple major cities onboard suggests the product delivers real value in municipal workflows.

Competitors include Computronix and Tyler Technologies for government permitting software, and Symbium for AI-assisted zoning analysis. But most existing permitting software digitizes the workflow without automating the actual analysis. AutoSitu’s AI review step is a meaningful advancement.

The Verdict

AutoSitu is tackling one of the most painful bottlenecks in real estate development. Plan review is slow everywhere, and the problem is getting worse as regulations grow more complex and planning staff remain scarce.

At 30 days: how many plan reviews has AutoSitu completed for its city partners, and what is the accuracy rate compared to human reviewers?

At 60 days: are developers proactively using AutoSitu to pre-check plans before submission, reducing the back-and-forth cycle?

At 90 days: is AutoSitu expanding to additional jurisdictions, and how much effort does it take to configure the AI for a new city’s zoning code?

I think AutoSitu has strong product-market fit. Every city needs faster plan review. Every developer wants faster approvals. The challenge is the government sales cycle, which is notoriously slow. But the early traction with major cities suggests they are navigating it well.