08/26/2026 | Press release | Distributed by Public on 08/27/2026 06:54
When you call 911 today, the operator on the other end is listening to you and typing what they hear. There's no transcript, no summary, and if you don't speak English, you can wait five to seven minutes for a translator. Prepared built the AI that closes that gap by transcribing, summarizing, and translating emergency calls in real time across roughly 1,000 agencies and 80 million calls a year, before Axon acquired the company in 2025.
In this episode of The Flywheel, Prepared co-founder and CEO Michael Chime joins M13 Managing Partner Karl Alomar and ProfG Markets host Ed Elson to talk through the decisions behind it, including why he left Yale after his third year, why Prepared gave its technology to government agencies for free before charging anyone, what changes about managing people when AI does more of the work, and what convinced Karl to write the first check.
Most 911 systems were built on the assumption that every call comes from a landline, so they can't accept a photo, video, or text from a mobile phone.
Michael discusses how a degree starts a career but it doesn't create one, and the decision comes down to whether the conviction and the ability to execute are real enough to justify the risk.
Michael shares how AI changes the org structure but not the underlying job. A 500-person company running 10,000 agents still has 500 people who need to believe in the mission.
When a caller doesn't speak English, the call taker presses a button and waits for a translator, and for a less common language like Vietnamese, that can run five to seven minutes on an emergency call. Michael explains how Prepared translates in real-time and passes the information straight to responders.
Young founders can benefit from the partnership of seasoned operators who offer experience without taking away their unbound vision, speed and energy.
Prepared is a public safety technology company that builds AI for 911 centers. Its software sits alongside a live emergency call and transcribes, summarizes, and translates it in real time, surfacing the details a responder needs.
In most 911 centers, the call taker listens to the caller and types what they hear, with no transcript or summary generated automatically. AI purpose-built for public safety runs alongside the call instead by transcribing it in the background, summarizing it, translating it when the caller doesn't speak English, and extracting the location, hazard, and injury details that matter to the units being dispatched.
Most 911 infrastructure was designed around the assumption that every call originates from a landline, so it has no way to ingest images, video, or text from a mobile device. Replacing those core systems outright has proven difficult, so newer software generally layers on top of the existing stack rather than ripping it out.
When a caller doesn't speak English, the standard process is for the call taker to press a button and wait to be connected to a human interpreter. For a less commonly requested language, that wait can run five to seven minutes on an emergency call. Real-time AI translation removes the wait, so the same information reaches responders at the same speed regardless of the language spoken.
A degree starts a career rather than creating one, and the decision rests on whether conviction and demonstrated ability to execute justify the risk.