Denver approves $4.6M AI-powered permit review contract
The Denver City Council approved a $4.6 million contract with ComplyAI on Monday for a new online tool intended to speed up the city’s permitting process.
Officials say the platform, CivCheck, uses artificial intelligence automation, machine learning and “intelligent document analysis” to streamline application intake and code-compliance review.
The council approved the five-year contract on a 10-1 vote.
District 7 Councilmember Flor Alvidrez cast the sole dissenting vote, citing concerns about the system’s reliability.
“I appreciate the department’s efforts on how technology can streamline the process,” Alvidrez said. “At the same time, I am growing more concerned because we’re seeing more reporting regarding inaccuracies and limitations of AI systems, especially when they are also asked to interpret complex rules and regulations.”
Alvidrez added that she has also heard from local architects and professionals in the development community who say that current AI tools “may not be ready to handle the full complexity of codes, building codes, and site-specific conditions.”
At-large Councilmember Sarah Parady suggested the city evaluate the program in a year.
“I hope we will check in in a year and make sure that we’re seeing if it’s delivering the kind of results that we’re hoping that it will,” Parady said. “Because this is a lot of money that otherwise could go towards staff, and we know that Community Planning and Development lost a lot of staff in layoffs.”
Robert Peek, development systems performance director for the Denver Permitting Office, has previously described the new platform as a tool to guide applicants through a process and “works like a smart-checklist or validation engine for our customers, and as they submit their construction drawings, it’ll surface any deficiencies.”
As for being ready to deal with the complexity of the city’s codes and regulations, Peek said it would be a gradual process, adding subsequent modules over the life of the contract.
“I think we realize that, as well, and that we need to build up the tool, see the value, and then start to expand from there,” Peek said. “Right now, we see about 37% of our applications get accepted on the first round of submittal, which is quite low.”
Peek said his department aims to increase the number of successful first-round submittals to 80% using the new AI tool.

