Guillaume Delepine

We just raised our seed from @a16z American Dynamism fund and Seven Stars Capital to build Longeye: AI that solves crimes in hours, not months. This became personal when I woke up to burglars running out my door, laughing. Police came but had no capacity to investigate:

This isn't about replacing detectives. It's about restoring their ability to actually investigate. Link to the full Police Chief Magazine feature: policechiefmagazine.org/product-featur…

We're heading for a breaking point where evidence load will overwhelm the system. Agencies that adopt smart AI now will keep pace and keep trust.

Police Chief Magazine just featured @Longeye_ai on how AI is transforming investigations. The opening scene: a detective drowning in PDFs, bulletins, and alerts. Somewhere in that chaos is the clue that cracks the case. This is modern policing.

Attending NSA Winter Conference? Come find us at booth #103.

AI solves the capacity problem. @Longeye_ai processes evidence up to 10x faster so investigators can focus on interviews, suspects, and closing cases.

We'll be at the NSA Winter Conference in DC this weekend discussing how AI helps law enforcement solve mounting digital evidence challenges.

The white paper explores how @Longeye_ai 's CJIS compliant platform helps agencies strengthen investigative capacity while upholding ethical and legal standards. If you're managing this crisis, this research will resonate. linkedin.com/pulse/revoluti…

Chief Carlson's research shows ethical AI can process evidence at 100x+ the speed of manual review. Not by replacing investigators, but by handling tedious work so they can focus on judgment, interviews, and cases only humans can solve.

Chief Denton Carlson just published a white paper on the reality every investigative leader is managing in silence: The evidence exists to solve more crimes. Your team just doesn't have time to review it.

When agencies measure data volume instead of case counts, resource allocation becomes honest. You see who's overloaded. You identify which cases need AI just to be processable. Measuring terabytes isn't about being technical. It's about measuring what actually matters.

Right now, we pretend case counts capture workload, then wonder why detectives are overwhelmed. We assign 30 cases assuming it's manageable, ignoring that 3 cases contain more evidence than the other 27 combined.

Tomorrow's detectives will measure workload in terabytes, not case counts. A detective in 1995 handled 30 cases. A detective in 2025 handles 30 cases too. But that number doesn't capture the real workload anymore. The better question: how much data are you reviewing?