Blue Voice raises $6m to put police rulebooks on officers' phones
A Harvard Law dropout has taken $6m in seed funding for an AI trained on department policies that general models cannot read. The moat is the data nobody has published.
Blue Voice, a Boston company that answers police officers' questions about their own force's rules, has come out of stealth with $6m led by SignalFire and Las Olas VC.
The pitch is a familiar shape. Harvey did it for lawyers, OpenEvidence for doctors, and founder David Lawrence draws the same comparison for policing.
What the three companies share is not the model. It is the corpus.
Blue Voice says it is trained on department-specific laws, local ordinances, protocols and guidelines — material that sits inside police forces rather than on the public internet, and which a general-purpose chatbot therefore cannot retrieve.
Lawrence was a Harvard Law student when an on-campus shooting involving a police officer set off a row about police conduct. He concluded that most policing errors happen when officers cannot reach their department's rules quickly, and left law school.
He founded the company with Amit Patankar, a Harvard MBA and former Google engineer, and Michael Gropman, a retired Boston police deputy chief. That is three years ago; the funding lands now.
Officers at 225 county agencies across 25 states use the tool, and the platform fields a question a minute, according to the company.
The alternative, on Lawrence's account, was flipping through a 15,000-page manual, waking a supervisor in the night, or asking Google or ChatGPT. He says consumer models got it wrong up to 30% of the time.
Blue Voice answers by pointing officers to the original regulation rather than generating an answer of its own, which Lawrence says is why officers trust it. The tool supplies the law; the officer decides.
Lawrence says the customer base grew elevenfold over the past year, and attributes that to departments seeing reduced crime and fewer operational controversies. Those are the company's claims, not published figures.
He offers two examples. A rookie officer, unsure whether he could intervene when a man appeared to be pressuring a girl into a vehicle, checked the app and was told the facts met the criteria for child enticement. In another case the system told a chief he could not return an officer to duty after a shooting without a third-party mental health evaluation.
The company also gives officers school maps during active-shooter calls, and has built a function aimed at cold cases. It competes with Lexipol, which is private equity-backed.
Police AI is not a comfortable category. Flock Safety's licence plate surveillance draws steady criticism, and Lawrence says he wants Blue Voice to show that AI can serve public safety without costing civil rights.
For founders eyeing a vertical AI, the lesson is where the defensibility sits. Blue Voice's advantage is not a better model but access to 225 agencies' unpublished rulebooks — an asset that takes years and a retired deputy chief on the cap table to assemble, and that a competitor cannot scrape.