Motorola · Edge AI Voice Tagging
Shipped voice tagging on body-worn cameras to 20+ enterprise police departments. 75% less data entry, even when the radio drops.
The problem.
Officers were spending up to 20 seconds per incident manually tagging video metadata—often after the fact, often inaccurately. Connectivity is unreliable in the field. Cloud-only solutions were a non-starter.
The approach.
Together with hardware, ML, and compliance teams, we scoped an offline-first voice-user interface running on the camera itself—hotword, intent recognition, and disambiguation that all work with the radio off. My job was mostly to keep the scope honest: ship the smallest version that actually saved officers time on day one.
The outcome.
Time-to-tag dropped from 20s to 5s (75% reduction). The feature became a competitive differentiator in 20+ enterprise deals (each $1M–$20M). In parallel, I led a strategic End-of-Life migration retaining $50M in ARR across 100+ customers (90% retention), and integrated C2PA content provenance to harden the product against deepfake manipulation.