What Calgary Police learned by moving early on technology

Brent Dyer’s office at Calgary Police headquarters was mostly packed when we spoke late last month. He was a day away from retiring from Calgary Police Service (CPS) after more than 32 years, most recently as executive director of IT, and he walked me through some of the work he’d done over his career.

The AI steering committee he set up early on included lawyers, and Dyer’s team worked through AI questions with legal, privacy, and access-to-information staff. Dyer laughs when he says they’re the “people that typically like to slow you down.” He put them in the room early, and says it’s what let the rest of the process move.

Dyer says CPS has been an early adopter of AI in Canadian policing. His work over the past five years focused heavily on governance, including rules around data, disclosure, privacy, and employee-built tools.

Dyer is a 2026 finalist for the CanadianCIO of the Year Awards in the public sector category. He led a team of 100 across a service of 3,300 people, the third-largest municipal police service in Canada. Winners will be announced in Toronto on Oct. 1, 2026. The award recognizes public-sector technology leaders modernizing government and public institutions.

This fall, he’s starting over. 

The province of Alberta is forming a new provincial police force under legislation introduced earlier this year. Dyer is taking over as executive director of IT and information management at the Alberta Sheriffs Police Service, which means he’s about to build a technology function from zero.

“I’m going to build them an IT shop out of LEGO, and I’m going to build it from the ground up. Build the team, build the technology, build the outcomes, build all this stuff. And then we’ll see where it goes,” says Dyer. “See what I do when I grow up.”

Governance came before the rollout

Dyer built the CPS AI steering committee years ago because he wanted questions of legal, privacy, and access worked through while the technology was still being shaped.

“Because they have been an active member of that particular steering committee, we’ve been able to kind of get around some of the friction and actually encourage them to take some chances,” says Dyer.

The approach predates the current generative AI push. 

CPS has had a Technology Ethics Steering Committee with citizen members since 2022, responding to public questions about how the service uses algorithmic policing technologies. Dyer’s AI committee is internal, and the two bodies work in parallel. By 2024, CPS had formalized its AI structure with usage guidelines, governance processes, and oversight around Microsoft Copilot.

Dyer says Chief Katie McLellan gives the team room to make mistakes as long as the intent is honest and the transparency is there.

“She’s OK if we make mistakes,” says Dyer. “The fact that we’re trying, we’re being transparent about it, and our hearts are in it.”

CPS sorts its data by sensitivity, and Dyer says its security systems can audit searches and email activity. If someone makes an inappropriate database lookup, he says the service can identify it quickly.

The service has used both facial recognition and Palantir Gotham, putting some of its technology choices under civil-liberties scrutiny. Citizen Lab says CPS uses Palantir to bring different data sources together for analysis, but not the platform’s predictive-policing functions.

Calgary City Council approves CPS’s total budget, and the Calgary Police Commission approves how that money is spent. Dyer says proposed technologies can also go to public advisory groups before rollout. He used drones as an example.

If an officer uses Copilot to think through a case, does every prompt go into their notes? 

Dyer’s team worked it through with legal and privacy staff. Instead of recording every Copilot prompt, Dyer says officers disclose the use once at the end and verify the information.

The goal, he says, is to free officers from the machine work of reviewing hours of video and going through evidence, so they can spend more time on the human work of interviewing witnesses and being in the community.

What happens when tech hits the real world

CPS runs body-worn cameras with live translation, so officers working a call in a language they don’t speak can communicate on scene. 

Drones head to certain calls before officers arrive, giving incoming responders a picture of what they’re walking into. Officers are using Copilot across the service, and some are building their own AI agents to handle parts of the job that are still slow or manual.

CPS moved on body-worn cameras early. The service was planning full deployment to frontline officers by 2017, but technical problems derailed the first rollout. When CPS came back in January 2019 with Axon cameras, Dyer says officers refused to wear them. They didn’t want their managers watching their work.

“We’ve gone from that to officers not wanting to go out on shift without a body-worn camera,” says Dyer.

The lesson from the first attempt, he says, was that tech needed operational buy-in.

The work was reducing complaints, building public trust, running the legal side, and communicating with a resident at the worst moment of their life. 

CPS became the first Canadian force to fully deploy AI translation on body-worn cameras this past June, with the system covering more than 60 languages in real time. 

“To be able to go out to somebody in the time of need, when they need us, they need somebody the most, probably the lowest experience of their life,” says Dyer. “And to be able to communicate with a police officer through live translation services has been an absolute game changer.”

The drone program is on the same trajectory, with CPS having between five to 10 drones in its fleet and plans for more to come. Dyer says CPS is on track to be one of the first three Canadian agencies to run drones as first responders.

Beyond what the technology team ships, some officers at CPS are writing their own AI agents to solve problems they run into on shift.

“It could be the wild west a little bit,” says Dyer.

The audit trail applies whether an officer runs a search themselves or has an agent do it for them.

What AI gave officers back

Dyer spent years trying to answer one question. What is the value of AI in policing?

“It costs money,” says Dyer. “It’s going to cost a lot of money at the end of the day. It’s never going to save anybody any money. I’m telling you that right now, especially in public sector.”

His answer showed up in the summer of 2026, in a search that took over the city.

Parker Wells went missing on July 16 from a specialized day facility in Calgary’s Thorncliffe neighbourhood. He was 11 years old, non-verbal, and autistic. 

Calgary Police led a massive search operation where hundreds of officers, trained searchers, and thousands of volunteers looked for him for nearly two weeks. On July 29, police announced they had found his body in a culvert system under the city.

“I’ve been here 32 and a half years, I have never ever seen the response for something that’s happened in our city like this one,” says Dyer. “Behind the scenes, there’s literally a thousand people that had been working on this at one time.”

Dyer says the technology his team had built over the previous five years was part of what kept the search going. Drones, mapping tools, and video analytics carried some of the manual load. AI worked through material that, he says, would otherwise have taken officers away from the search.

“It allows officers to be more human rather than to do the machine work behind the scenes and do the investigation,” says Dyer. “It allows them to have some emotion and give them the motivation to keep searching.”

Dyer keeps coming back to time. If technology can take some of the manual work off officers, he says they can spend more of it on the case.

“If we can get to a case faster, if we can solve a case faster for citizens, we earn their trust, and that is the goal at the end of the day,” says Dyer.

Final shots

• Bring legal, privacy, and access teams in before rollout. The earlier they help write the rules, the less cleanup there is after a complaint, disclosure issue, or court challenge.
• AI does not have to cut costs to be useful. Dyer measures it by the time it gives back to officers and investigators.
• Employee-built AI gets messy fast. Set the rules for data, logging, and review before everyone starts making their own tools.

This article is part of a series profiling the finalists for the 2026 CanadianCIO of the Year Awards, public sector category, presented by the CIO Association of Canada. The winner will be announced Oct. 1 in Toronto. Digital Journal is the national media partner for the CIO Association of Canada.