Aurelian deploys AI to route non-emergency 911 calls

The New York Post reports that Seattle-based public safety technology company Aurelian has built an AI call-taker named Ava that answers police non-emergency phone lines, takes reports, and routes complaints to the appropriate department. Aurelian CEO Max Keenan has stated that roughly 70% of calls entering a typical emergency communications center are non-emergencies, including barking-dog complaints, parking disputes, lost property reports, and requests for city services, per Aurelian's publicly reported deployment data corroborated across multiple agency launches. The New York Post cites a March report finding that the LAPD answered just 57.43% of 911 calls within California's 15-second standard in 2024, below the state benchmark. City officials told The New York Post that roughly 100 operators must be on duty to meet minimum staffing requirements. Volusia County, Florida reported a 98.5% successful resolution rate for non-emergency reports handled by Ava since adopting the platform, per the Volusia County Sheriff's Office.
What happened
The New York Post reports that Seattle-based public safety technology company Aurelian has developed an AI call-taker named Ava that answers police non-emergency phone lines, takes reports, and routes complaints to appropriate departments. Aurelian CEO Max Keenan has stated that roughly 70% of calls entering a typical emergency communications center are non-emergencies, a figure Aurelian reports across its deployed agency base. The New York Post cites a March report that found the LAPD answered just 57.43% of 911 calls within California's 15-second standard in 2024, below the state benchmark. City officials told The New York Post that roughly 100 operators must be on duty across a 24-hour period to meet minimum staffing requirements.
Deployments and track record
Aurelian's Ava platform has been adopted by multiple agencies. Volusia County Sheriff in Florida reported a 98.5% successful resolution rate for non-emergency reports handled by Ava since deployment, per the Sheriff's Office. Kitsap 911 in Washington State launched Ava on a dedicated non-emergency line in May 2026, per public reporting. Aurelian raised a $14M Series A in August 2025 to expand its dispatcher-relief AI capabilities.
Technical details
Ava handles low-acuity call types including barking-dog complaints, parking disputes, noise complaints, lost property, abandoned vehicles, and minor traffic crashes with no injuries. These map to mature automation techniques: automatic speech recognition, intent classification, slot-filling workflows, and deterministic routing to municipal service channels or online forms. Ava continuously monitors every interaction for emergency signals; if detected, it immediately transfers the caller to a human telecommunicator. For practitioners, the key engineering and governance concerns are transcription accuracy in noisy environments, emergency-detection precision, and integration with existing CAD (computer-aided dispatch) and records systems.
Context and significance
Emergency dispatch centers nationwide face staffing shortages and high non-emergency call volumes, which public reporting connects to slower emergency response times. The LAPD figure cited by The New York Post illustrates the operational stakes. Aurelian positions Ava as an operational relief tool for strained centers, not a replacement for emergency dispatch.
What to watch
Observers should monitor independent performance evaluations, published accuracy and safety metrics for emergency-detection logic, and any municipal pilot results or procurement records as additional agencies evaluate the platform.
Scoring Rationale
A notable operational AI deployment in public safety with verified performance data from multiple agencies, directly relevant to practitioners building voice AI and dispatch automation. The LAPD response-rate context adds policy stakes, but this is a product deployment story rather than a technical breakthrough or frontier AI development.
Practice interview problems based on real data
1,500+ SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
