Cleveland is using AI to flag property problems and fight blight. Here's how it works.

It doesnt necessarily look like much. A white car with City Support Vehicle written on the doors and two cameras on the roof.

Watch the video below:

But the retrofitted Ford sedan is part of Clevelands escalating push to fight blight. For almost a year, its been traveling city streets, taking pictures of every property it passes.

Artificial intelligence scans the photos for dozens of potential red flags, from out-of-control grass to collapsing porches and illegal dumping. Then city officials review the pictures to see if theres really something worth pursuing.

Were trying to chase down the bad actors. And were trying to work with the good actors. And this makes it way easier, said Dennis Butler, who deals with code enforcement as manager of the demolition bureau in the Department of Building and Housing.

In late 2025, Cleveland teamed up with City Detect, an Alabama-based company that uses visual AI to help local governments monitor property conditions. Their yearlong pilot project, paid for by an $85,000 grant from the Rocket Community Fund, is scheduled to end Feb. 1.

So far, the results are promising, said Doug Meil, Clevelands assistant director of Urban AI.

City Detects data dashboard shows the car, driven by city workers, has snapped multiple images of more than 150,000 properties. Most of those photos were high-quality enough for AI to analyze. The system identified possible concerns at 46% of the parcels.

In some of those cases, theres no reason to follow up. For example, the artificial intelligence recently flagged potential violations, including a missing door, at a construction project on the East Side. But theres no door yet because the building isnt finished.

In other cases, though, the pictures justify sending out an inspector to get a closer look at a house or dispatching a crew to clean up a pile of tires.

Nobody is cited based purely on the submissions from the car, Butler said. Everything is confirmed. This is merely bringing something to our attention.

Cleveland's last old-school citywide property survey took place in 2022, through a partnership between the city and the Western Reserve Land Conservancy. It cost $170,000 and took roughly six months to complete, with support from 40 inspectors.

It was expensive and time-consuming, Meil said of that hands-on approach. The prior one was done in 2015. Every seven years is not a great cadence to be able to keep track of 158,000 parcels.

The pilot contract with City Detect costs half as much. It requires two drivers, who take turns.

They can canvass the whole city several times a year, driving 25 to 35 miles per hour. That gives inspectors access to much more recent images and enough data to track property conditions over time.

Its just more efficient. Its less expensive, Butler said. And the ability to update it continuously, thats really where the value is, in my opinion.

The cameras only take pictures not videos. City Detects system blurs out peoples faces and license plates. And city employees dont see the images until the day after theyre captured. Theres no real-time monitoring.

This is all about buildings. This is all about tires. This is all about detecting graffiti, Meil said. This is not about tracking individuals.

Officials compared the program to the Google Maps Street View feature with a layer of analysis.

This is all taken from the public right-of-way, Butler said of the photos. Just like an inspector would take. Its even less intrusive, because its from the street. An inspector could walk around. He could ask neighbors for access and things like that. Were not doing that here. But we do have a front-facing photo of the property.

The Cleveland 311 team, which deals with non-emergency complaints and service requests, recently started using the City Detect system to track roadside trash. Other cities tap the technology to identify everything from potholes to hurricane damage.

Unlike Dallas, which deployed 100 City Detect cameras on sanitation trucks this year, Cleveland tiptoed into testing artificial intelligence for code enforcement.

The advice that everybody has on AI projects or really any analytic project, for that measure is to start slow and iterate, Meil said. And I think weve done that. I think weve done that well. Lets get the data. Lets validate that it works. Lets make multiple passes to see how consistent it is. And lets be careful about what were saying in public and make sure that its accurate. So were taking it slow.

Its still unclear whether the technology will become a permanent fixture in the citys code-enforcement toolbox.

Weve got a couple more months left in the pilot, Meil said. And we are assessing the year-two plan.

Michelle Jarboe is the business growth and development reporter at News 5 Cleveland. Follow her on X @MJarboe or email her at [email protected].