Artificial intelligence will soon be listening to emergency and non-emergency calls made to police and fire department dispatchers in Long Beach.
City officials confirmed last Thursday that they are in the early stages of implementing a new program called CommsCoach AI, developed by GovWorx.
The AI system will review calls solely for quality assurance purposes, according to the Long Beach Post News. It will flag interactions where it believes dispatchers may not have responded appropriately.
Once fully operational, CommsCoach AI will monitor all 600,000 annual calls to the city’s dispatch center, automatically generating audio files, transcriptions, and overall scores for each call. This technology replaces the previous manual review process, where supervisors aimed to spot-check about 12,000 calls per year—a goal they’ve struggled to meet due to staffing shortages.
“It’s a training tool to make sure dispatchers are gathering the details that are necessary so that first responders are clearly aware of the incident upon their arrival,” explained **Reginald Harrison**, Director of the city’s Department for Disaster Preparedness.
The AI system will flag calls based on various factors, including a dispatcher’s tone of voice, use of specific keywords or phrases, arguments, or failure to request essential information. For example, if a caller mentions the word “gun,” the AI will analyze whether the dispatcher asked appropriate follow-up questions. Flagged calls will then be reviewed by a human supervisor.
Harrison emphasized there is “no goal of replacing dispatchers with AI,” noting the city is not looking to automate emergency call handling or remove human judgment from critical decisions.
City officials consulted with the dispatchers’ union before presenting the contract to the Long Beach City Council in August. The CommsCoach contract costs approximately $68,000 per year through 2028, and all data and recordings will remain on city servers.
The program is currently in a testing phase, reviewing select calls to ensure proper functionality. It is expected to be fully operational by the end of March 2026.
Recent Comments