NewsLayer

Install NewsLayer

Get the app experience — one tap from your home screen, instant loads and breaking-news alerts.

NewsLayer.com
NewsLayer PulseLIVEBTC$79,118+0.12%ETH$2,478-1.33%SOL$104.66-1.90%XRP$1.39-2.40%DOGE$0.0835-3.31%ADA$0.1996-2.77%Total Cap$2.78T-0.53%Layer Index45 Neutral

Harvard Law dropout raises $6M for Blue Voice to build a “Harvey for police officers”

The seed round for the app that provides real-time legal and policy guidance to officers was led by SignalFire and Las Olas VC.

Marina Temkin

Publisher TechCrunch AI

Aug 31, 2026 at 6:35 PM UTC · 3 分で読める

Harvard Law dropout raises $6M for Blue Voice to build a “Harvey for police officers”
Image via TechCrunch AI
翻訳中…

David Lawrence was a Harvard Law student when an on-campus shooting involving a police officer sparked intense controversy over police conduct. Realizing that most policing errors occur when officers lack instant access to department rules, Lawrence dropped out to tackle the problem with AI. He teamed up with co-founders Amit Patankar, a Harvard MBA and former Google engineer, and Michael Gropman, a retired Boston police deputy chief, to launch Blue Voice—a Boston-based AI startup that provides real-time policy guidance to officers in the field.

Three years later, Blue Voice is coming out of stealth with $6 million in funding led by SignalFire and Las Olas VC. Officers at 225 county agencies across 25 states now rely on the tool daily to ensure their actions follow department protocol—a platform Lawrence likens to Harvey, the AI tool widely used by lawyers, or OpenEvidence, which is a similar tool for doctors.

“When an officer makes a decision, they do so from memory or best guess, based on thousands of pages of law, policy, municipal ordinances, and state laws that they’ve internalized and keep in their head,” Lawrence told TechCrunch.

Before Blue Voice, an officer who couldn’t recall a specific protocol—like the exact seventh step required at a crime scene—had few good options: flip through a 15,000-page manual, wake up a supervisor in the middle of the night, or search generic tools like Google or ChatGPT that lack police-specific training. The problem was that traditional methods took too much time, while consumer AI models delivered incorrect answers up to 30% of the time, Lawrence said.