AI in Government Permitting

Ask a building official where their time goes, and it's rarely the part of the job they trained for. Before an application is even opened, staff are answering the same questions, chasing down missing documents, and sorting submissions that arrived incomplete.

That work is piling up as the pressure to build only grows. The country is short more than 4.7 million homes, and states are passing new laws to move permits faster. Departments are being asked to do more on the same deadlines. At the same time, for the first time in two decades of surveying state CIOs, AI ranked as the top priority — ahead of cybersecurity, cloud, and budget.

This guide is written for the people doing the work. It covers where community development departments are already seeing results with AI, what early deployments are teaching us, and what to look for before evaluating anything new.

‍

Where AI is working in permitting

The most practical place to start is at the front of the process. Every permit begins with an application, and when that application arrives incomplete, every step after it gets slower. The guide walks through the three points where departments see the clearest difference. At intake, AI can catch common errors before a file reaches the queue. In plan and application review, it can surface the relevant code sections, flags inconsistencies, and draft comments for staff to approve. At the front desk, it can answer routine public questions around the clock, so staff time goes to the work that requires their judgment.

This guide also covers what matters most for public trust: staying in control. The strongest deployments support professional judgment rather than replacing it. They're trained on local code and requirements, and they leave an audit trail on every decision so staff stay the decision-makers.

‍

Inside the guide

In this guide, you'll learn:

• Where the bottlenecks in building live, and why most delays happen before technical review even begins
• How AI is being applied at three stages of permitting: application intake, plan review, and the front desk
• What early adopters are seeing in practice, from California to New York
• The four principles for staying in control: know your community, audit every decision, keep humans in charge, and prioritize security
• How to deploy the right way, starting with a specific problem, your own local context, and the full applicant journey
• The exact questions to ask your own department, and any vendor, before adopting AI