Artificial intelligence is rapidly becoming a key component of American law enforcement. Officers today are generally familiar with its usefulness in areas like license plate recognition, data analysis and digital forensics.
These early efforts to harness AI for police work have yielded some helpful advances and innovative products. But what can still be missing is an approach that’s more comprehensive and holistic — a wide-ranging platform that can work across and help unite the many varied tools and activities of an officer’s day.
That’s the idea behind Motorola Solutions’ new Assist AI, a broad-based AI engine designed to help streamline workflows, improve situational awareness and reduce the burden of mundane administrative tasks for law enforcement users. It’s a feature of several of the company’s latest and most advanced products.
Assist “stitches everything together across our suite of applications,” said Todd Piett, Motorola Solutions’ senior vice president of command center and cloud solutions, who discussed it as part of his opening remarks at Motorola Solutions Summit 2025, held in May in Grapevine, Texas.
In the field, Assist can support officers with time-saving help like starting report drafts based on verbal narrations, conducting quick searches based on spoken commands and generating integrated incident timelines that blend information from various sources. It’s a key component of the company’s new SVX, which combines a remote speaker microphone, body camera and AI assistant.
On May 13, alongside the summit’s opening, the company announced it would provide all public safety agencies with a complimentary starter edition of its new Assist Chat, a secure, CJIS-compliant AI chat assistant that connects users conversationally to their agency’s data, procedures and case histories.
“With Assist Chat,” said Mahesh Saptharishi, Motorola Solutions’ executive vice president and chief technology officer in announcing the offer, “any role within a public safety agency – whether on the front lines or in the back office — can now converse with their data in a secure environment.”
“Converse” is the key word here; Assist Chat is based on a large language model (LLM) that lets it interact naturally, ChatGPT-style. It’s secure in its access to agencies’ documents and data, and nothing is shared. It lets users make secure text or voice queries of dispatch, records, evidence data and previous case histories, including searching videos and images.
An AI-powered day
Assist is also expanding to serve a broader range of the law enforcement team, including nonsworn positions who often shoulder much of departments’ daily administrative burden. Call-taking and dispatch in particular represent an area ripe for streamlining; Assist provides those professionals real-time transcription, translation and summaries of calls, with caller data, mapping and facility information integrated in a single interface.
If a caller reports an auto accident, for instance, Assist can support that telecommunicator by producing a prepopulated incident report that facilitates dispatching an ambulance. It can also help officers in the field by checking their narratives against video, radio and CAD data and can power deep research by investigators and analysts.
“AI is not a discrete ingredient but the foundation of our public safety software,” Saptharishi added. “Think of Assist as our products’ central nervous system. It can proactively support each user with role-specific information contextualized to the time, task, person and place.”
Piett shared an example of what an officer’s Assist-integrated day might involve. He can start duty with a verbal command to Motorola’s voice-activated personal assistant, ViQi, which officially begins his shift. While driving, he observes a possible DUI. He can initiate video recording in several ways and search the driver’s plate number by reading it out to ViQi, which is designed for Assist access when typing is inconvenient. His communications with dispatch are transcribed in real time, with key words flagged. An incident report is begun automatically; all the dispatcher must do is accept it.
When the officer initiates a stop, Piett imagined, the driver initially complies, then abruptly backs up, causes an accident and flees. The initial officer can use an emergency button to summon pursuit reinforcements and push the driver’s information to colleagues.
As this unfolds, a second call comes in for someone breaking into a car. To assess the scene, supervisors can quickly launch a drone — a new partnership between Motorola and Brinc announced in April lets Brinc drones integrate with key components of Motorola’s suite, including radios and software for call management and real-time crime centers. Drones can be deployed with the push of a button, respond to sensor alerts and feed real-time intelligence directly into Motorola Solutions devices.
A suspect is apprehended but doesn’t speak English; however, ViQi provides real-time translation. It also allows verbally tagging incidents with specific keywords (e.g., “traffic stop,” “arrest”) that initiate other workflows — a key time-saver since, Piett noted, some 60% of bodycam videos never get tagged.
A report is AI-generated based on the officer’s voice activity and video; the officer only needs to polish and fact-check it. The next day, as an evidence technician prepares everything for court, he can batch-redact faces and other sensitive information automatically.
Connected investigations tell the story
A subsequent presentation from D.J. Seals, a user experience enablement lead with Motorola Solutions and a former detective from Georgia, and Stuart Boutell, the company’s director of evidence product management, explored how Assist can complement SVX in a connected investigation.
Connecting audio and video pieces from disparate sources can be challenging, Seals noted – humans may err, fragments may be decontextualized, batteries can fail, pieces must be obtained from other parties. A lot can go wrong when systems are disconnected, and there can be penalties for mishandling, misplacing and failing to disclose something you missed. Yet the amount of digital evidence involved in investigations keeps growing.
Connected investigations facilitate telling the whole and accurate story, Seals emphasized, in ways juries can understand. They can also condense steps and shorten timelines, Boutell noted. Learn more about connected investigations here.
Humans remain in charge
Additional benefits to Assist include activity logging and automated audit trails, which can be useful for internal oversight and accountability, and support for multiagency sharing and collaboration. Future iterations may become more predictive and proactive, able to surface emerging threats and contextual connections.
Badly designed AI can be more hindrance than help, Motorola notes, and it’s tried to insert Assist naturally and at points where manual processes can bog down or be susceptible to error — areas like call-taking, redaction and narrative-building. The seamless integration of multiple data sources ensures consistency and context, as well as helping prevent errors.
Assist doesn’t replace the human officer, the company emphasizes — it’s a labor-saving complement to augment key duties, but humans must provide oversight and ultimately make their own decisions.
Learn more about Assist here.
