A series on the oldest problem in restaurants: getting the right information to the right person. (Part 3 of 3, view part 1 here and part 2 here)
Thanks to AI, the technology to fix team communication in restaurants is finally within reach. The only question left is who builds it, and for whom.
Every major technology wave over the last twenty years built tools for desk workers and left everyone else behind. Anthropic, OpenAI, and Google will build AI that works for teams, and they'll build it for enterprise teams, Fortune 500 teams, corporate teams. It's the easiest target: corporate work looks the same everywhere, and it's the work they do themselves, so it's what they understand best. The 20%, all over again.
They won't build it for the embodied shift worker. The companies building one model for everyone can't spend the time it takes to truly understand work this nuanced, across this many different jobs. The GM running a dinner service, the surgeon handing off a patient, the site foreman briefing a crew at 6am all run on workflows you have to understand from the inside.
My wife works in one of those industries. She is a surgery resident, and I watch her change shifts every day. When she signs out, exhausted, her beeper goes off. It's her cue to call the attending surgeon taking over and share an update on every patient still in the bays waiting to be discharged. You'd think it would all be in the digital patient records, but whether it's time, exhaustion, or just that it's easier to tell someone, there's a part the notes never capture. This isn't just a restaurant problem.
Far from edge cases, they're the majority of the people who make the world run, doing work that takes care and attention.
So who builds it? The teams that understand the work from the inside. That means knowing which three things on a Tuesday dinner shift actually matter, and who needs to hear them before the moment passes. It lives with the people who've worked the floor, and the only way to build for them is to go deep. Forward-deployed engineers build alongside the operators who've spent careers in every role in the restaurant industry, until the software understands the shift the way the people on it do.
AI is going to commoditize the plumbing: connecting systems, moving data between them, orchestrating it through clean APIs. That's useful work, but it's not the hard part. The hard problems are the human ones, because people are hard and complex. In a restaurant, a team runs on how well it communicates. We started Colette to apply AI to the hardest problems in the restaurant industry, and after years of working alongside and building for restaurant operators, I've come to believe that team communication is at the root of what makes a restaurant successful or shuts it down.
Danny Meyer named it in 2006. Back then there was no tool that could solve it, but now there is. I'm confident that this time we get it right, and that the people who've always been left behind finally get tools built for how they actually work.