Editor’s note: This article is part of Police1’s Emergency Communications Week, which looks at how dispatch is changing — from smarter tools and automated routine tasks to new approaches that reduce unnecessary 911 demand. Thanks to our Emergency Communications Week sponsor, Autura.
In 2025, the Sparks (Nev.) Police Department received 160,000 calls. A staggering 71% of those calls were non-emergent. For an under-staffed department, operating at minimum staffing, their call takers were spending too much time on record requests, answering administrative questions about filing reports and complaints, and transferring calls to other city departments.
Like most departments in the United States, Sparks Police couldn’t easily or quickly solve the high volume of non-emergency calls by hiring more staff.
Instead, they hired an AI assistant.
“Each of our dispatchers is working a radio channel while they’re answering the calls from the public,” Connie Shepperd, the police services manager for Sparks Police, said.
“We were looking to improve call handling by helping manage call surges and providing faster service to the community with reduced hold times and wait times for non-emergency.”
Shepperd, who supervises the 911 Communications Center, explained that the recently implemented system was trained on the specific call types and routing needs Sparks Police dispatchers regularly handled through the department’s non-emergency line.
The Sparks Police AI assistant, powered by Desk Officer, answers all calls to the department’s non-emergency number. The AI assistant answers callers’ questions when it can, transfers non-emergency callers to the correct person or department, and routes high-priority or potentially life-threatening calls to a 911 dispatcher.
“We originally considered implementing a traditional phone tree, but when we saw some of the AI possibilities, we knew that the phone tree would decrease community satisfaction and still probably result in the majority, if not all, calls going to a dispatcher,” Shepperd said.
| RELATED: Chatbots: The next technology revolution in 911 dispatch centers
Training the AI assistant on local call-handling rules
Shepperd’s dispatcher team worked with the AI assistant vendor to set up rules for call answering and transferring. The rules, specific to Sparks Police, were tested for four months.
“Our whole dispatch team was instrumental in creating the transfer rules for the Desk Officer,” Shepperd said. “Their main concern was that it was going to handle calls appropriately and route them to the right areas.”
During the training phase, calls to the AI assistant tested the rules and assessed the effectiveness of the call handling. The test results were used to fine-tune the AI assistant rules to accomplish these goals:
- Quickly route actual emergency calls to dispatchers who can determine the nature of the emergency, give the caller instructions, and dispatch response resources.
- Answer questions from non-emergency callers or route them to the correct city department, person or resource.
- Immediately answer incoming calls with no wait times or holds.
An additional benefit: the Sparks Police AI assistant is fluent in more than 30 languages, allowing it to understand callers and respond in the language they use.
| RELATED: Smart dispatching: How artificial intelligence is reshaping emergency response
Early results: 1,500 calls handled in two weeks
After two weeks of use, the Sparks Police AI assistant has been able to answer 25% of its non-emergent calls. That’s about 1,500 calls and 50 hours of talk time.
Dispatchers, who were eager for the system to come online, are already noticing the impact of the AI assistant during busy periods or call surges.
“We’re looking at the faster answer times and how many routine or non-relevant calls are handled and whether or not our staff gets to spend more time on higher priority situations,” Shepperd said.
Over time, Sparks Police will monitor the AI assistant’s performance, specifically in these areas:
- Where are calls being routed by the AI assistant and are those calls being routed correctly?
- What types of non-emergency calls are still reaching dispatchers, and could additional transfer rules further reduce that volume?
- How many fewer non-emergency calls are being answered by dispatchers?
- Are callers, the public Sparks Police serves, experiencing shorter wait times and faster resolution of their reason for calling?
In addition, Shepperd’s team will continue to review the routing and transfer rules, assess how calls are being triaged and ensure they are striking the right balance between the AI assistant and the human dispatcher.
One unexpected benefit of the AI assistant is a significant drop in the number of robocalls being answered by dispatchers.
“A lot of the nuisance calls have been eliminated from our center,” Shepperd said.
| RELATED NEWS: Calif. sheriff’s department to answer non-emergency calls with AI
Questions to ask before adding AI to your dispatch center
Shepperd encouraged other call centers to evaluate the AI solutions available for assisting dispatchers, especially by reducing non-emergency call volume.
Begin by asking your current CAD vendor if there is an AI assistant for call triage, transfer rules for non-emergency calls, automated responses to commonly asked questions and instantaneous language translation.
When considering solutions, assess the capability to integrate with existing software and processes, the ability to set department- and community-specific triage and transfer rules, and the potential for the system to both serve the public better and lighten the burden on the department’s call takers and dispatchers.
Because AI technology is evolving quickly, it may be tempting to wait six months or a year for more refined capabilities. But waiting means missing out on tools already available to better serve the community and reduce daily non-emergency call volume.
It also makes sense to involve dispatchers in scoping the problem, reviewing products, attending demos, and selecting a solution for the department, as they will play a large role in writing the rules and testing the system before it goes live.
“(Dispatchers) were very excited about its implementation in the hopes that it was going to alleviate some of the pressure on them,” Shepperd said.
Getting dispatchers involved early will help the system succeed and free them to focus on the work that matters most: answering time-sensitive, life-threatening calls and getting the right resources to the right place at the right time — the mission that called them to public safety in the first place.
NEXT: How is AI revolutionizing policing? Dive into our Policing Matters Podcast episode, where we explore the role of AI in modern law enforcement, real-world applications and benefits, and ethical and operational challenges. Listen to the full podcast here.
