Relay Raises $36M to Turn Frontline Radio Chatter Into AI-Readable Operational Data

Relay develops a cloud-hosted smart radio communications system designed for workers who spend most of their shifts on their feet, including factory production line operators, hotel housekeepers, caregivers, emergency responders and supermarket workers. For many of these employees, the radios are the primary interface for reporting to supervisors, providing updates and flagging problems. Relay says much of the most valuable operational knowledge at such businesses is communicated over radio waves, and most of it stays there.

The company argues that frontline workers are usually too busy to enter that information into a computer, meaning almost half of it is never recorded and disappears after it is spoken. Co-founder and Chief Executive Chris Chuang calls the problem "signal evaporation." He said it occurs because traditional systems of record ask workers to record details sometimes hours after they have left the task or incident, leaving the information subject to recall and interpretation.

Relay has developed a voice AI system that integrates with its smart radio platform to capture workers' conversations as they use handheld radios, turning speech into structured and queryable data that can feed operational workflows and generate insights. The system is trained to understand domain-specific jargon. Chuang offered the example of a warehouse worker reporting over the radio that "the palletizer has broken down three times this week." While some AI tools might confuse "palletizer" with "appetizer," Relay understands that a machine stacking goods onto pallets has a recurring glitch and automatically creates a work order ticket, he said.

Chuang said that as the problem is fixed, a technician can speak into Relay the details of what was found and the steps taken to address it. That information is then structured and becomes part of the documented record, which the company says streamlines root cause analysis, reduces unplanned downtime and closes the gap between what is spoken on the line and what is stored in a traditional CMMS.

Relay's AI system also includes a TeamTranslate capability that provides real-time voice translation in more than 30 languages. Workers speak in their native language, and the system automatically relays what they say in the receiver's own language to support over-the-radio communication among multilingual teams.

Chuang said many customers never realized how much of their operations were not formally recorded. He said official logs at major manufacturers capture only about one in four equipment stoppages, with the rest remaining in live radio discussions and workers' memories. In a loud, fast-moving manufacturing plant, much of that information is lost before someone is able to record it, he said. The report also described a food processing facility where workers used Relay radios to discuss a broken water pump that was functioning poorly, but no one created a record.

Relay says its goal is to bring AI to workforces that do not sit at desks in front of computers all day. The company's pitch is that capturing what frontline workers say in real time, rather than asking them to record it after the fact, can help businesses avoid missed handoffs, rework and expensive operational delays.