IE 11 Not Supported

For optimal browsing, we recommend Chrome, Firefox or Safari browsers.

Hexagon Releases AI Tool for Real-Time Emergency Analysis

A new AI-based tool scans incident reports in real time to notify 911 call centers when they’re being inundated with calls about the same emergency, so they can coordinate the most efficient response.

An emergency dispatcher
Shutterstock/M.Moira
Between smartphones, IoT censors and automated alerts, the volume of data that flows through some 911 call centers on any given day has grown exponentially over the past decade. When a center is inundated with calls about a single incident, the ability to know which calls are related, and then dispatch the nearest and most-needed response units to their location, can be more than any one, manual call taker can handle. With this in mind, the Safety and Infrastructure division of the technology company Hexagon created an AI-based tool for recognizing related incidents in real time and helping call takers coordinate a response.

A news release on Tuesday announced the software as HxGN OnCall Dispatch Smart Advisor, or Smart Advisor for short. The company’s Strategic Product Manager Jack Williams told Government Technology that more than 18 months ago, he and some colleagues in the Safety and Infrastructure division wanted to see what public safety challenges might be solved with recent innovations in AI, machine learning and advanced statistics. After interviewing about 15 customers and 30 internal subject-matter experts and conducting several surveys, they realized there was a gap in what they called “real-time operational intelligence.”

“There were tools out there that do AI and have these reporting and analytics capabilities. We have a product that does that. But nothing was solving that real-time analysis, that ability to monitor things on the fly, something that’s constantly trying to connect the dots,” he said. “Having something continually assessing, looking for those patterns, similarities and anomalies, and then alerting [call takers and dispatchers], can really make a big difference, whether it’s an everyday event, whether somebody’s driving around on a moped snatching purses and phones, or a major emergency where those minutes are very critical.”

From companies such as RapidSOS and Motorola Solutions, there are a growing number of platforms on the market that collect video feeds or IoT inputs and send them directly to first responders, but this isn’t one of those. Williams clarified that this is a set of software “agents” that constantly read incident reports logged by 911 call takers in Hexagon’s CAD product, looking for keywords, similarities, recurring locations, statistical anomalies, weather, etc. If it finds any trends, patterns or links, an informational alert pops up on the 911 call taker’s screen along with a simple recommended action, like informing a relevant detective or tactical squad. The call taker can choose to share the information with a dispatcher or responders in the field, or attach it to the event record for posterity.

Williams said dispatch systems contain search functions already, but sifting through reports manually to discover which ones might be related isn’t usually feasible, especially as a major incident — such as a fire, civil unrest or series of bank robberies — is unfolding. Users can also configure Smart Advisor to watch for incidents according to specific criteria.

“If you had a particular location that you were monitoring where there was an event or planned protest, something of that nature, you might want to set a mission up for the location agent, so you’re monitoring all the activity in that area,” he said.

Andrew Westrope is managing editor of the Center for Digital Education. Before that, he was a staff writer for Government Technology, and previously was a reporter and editor at community newspapers. He has a bachelor’s degree in physiology from Michigan State University and lives in Northern California.