SoundThinking, Inc. (Nasdaq: SSTI) announced the launch of SafetySmart Field Agent, an AI product designed to convert public safety data into operational intelligence for law enforcement agencies.

Field Agent allows users to query large volumes of public safety data through text or voice in natural language, returning responses that can include pre-shift briefings, command staff reports, deployment-planning materials, community summaries, grant narratives, maps, and charts. The company says the tool can generate outputs in minutes that previously required trained analysts and manual processes taking hours.

"For decades, getting actionable intelligence out of public safety data meant trained analysts, manual exports, and hours of effort. Field Agent does that work in minutes," said Ralph Clark, President and CEO of SoundThinking.

The product draws on ShotSpotter gunshot detection data and other SafetySmart platform data. Future versions are expected to incorporate additional proprietary data sets, including vehicle intelligence data, as well as third-party data sources.

Field Agent is currently in beta with more than 20 ShotSpotter agencies. Broader availability is targeted for later in 2026. The product supports English, Spanish, and Portuguese. Pricing and packaging details have not yet been announced.

The product is described as incorporating human-in-the-loop oversight, with outputs intended as decision-support tools subject to review by authorized personnel before action is taken.

Commissioner Andrés Romero, Director of Video Surveillance, Analytics, and Urban Intelligence at the National Police of Uruguay, said in a statement: "Field Agent essentially places an analyst in the pocket of every operator."

SoundThinking states it serves approximately 300 customers and supports approximately 2,100 agencies. The company's SafetySmart platform includes ShotSpotter, CrimeTracer, SafePointe, and PlateRanger, among other products. The information in this article is based on a press release issued by SoundThinking.