Every so often you meet a company that has already died once. Kea is one of those. The version of the business we invested in is, in the best possible way, not the business Adam Ahmad set out to build back in 2018. It is a far better one, because the world finally caught up to the idea.
Kea builds voice AI for restaurants. In plain terms, it answers the phone. It takes the order, fields the questions ("are you open Sunday?", "do you do gluten free?"), pushes the order straight into the point-of-sale system, keeps the menu and stock levels in sync, and collects payment through a secure text link while the customer is still on the line. For the owner there is a layer of tooling to monitor every call, train the agent and step in whenever they want to. Call the demo line and it picks up on the first ring. It is genuinely uncanny.
The comeback
Adam started Kea with broadly the vision he is executing on today. He was simply too early. The first incarnation ran headlong into the wall that defeated a lot of ambitious voice startups: the AI was not accurate enough to be trusted with a paying customer, and the cost of running it made the unit economics impossible. Through 2024 the business came close to folding altogether.
Then the ground shifted. As large language models leapt forward, the team rebuilt the product from the ground up. Kea v2, the product in market today, runs at north of 99% call accuracy on a healthy gross margin, two numbers that simply did not exist in the old world. From a standing start in the first half of 2025 the business has won 30 restaurant customers and built a real, recurring revenue base, every one of them a brand new logo rather than a holdover from the old days. There is not much in venture more compelling than a proven founder getting a second swing at the same pitch with a much better bat.
How we met
We were introduced to Adam by our friends at Marbruck, who co-invest alongside us in Mutinex and EatClub and who have backed Kea several times over the years. When an insider who knows a business this intimately, through the good times and the very bad, offers you a seat at the table, you take the call. This was access offered by a believer, not a process being run because the company had exhausted its options closer to home. That distinction matters a great deal to us. Our best investments have almost always been proprietary, relationship-led deals rather than competitive auctions.
Why now
Hospitality is being reshaped by AI faster than almost any industry we look at, and for good reason. Wages are up nearly 40% over the past decade, labour is chronically short, and yet roughly one in five restaurant orders still arrives over the phone, often during the exact peak windows when there is no human free to pick up. Every one of those missed calls is a lost order. A tool that recaptures that revenue while taking a job nobody enjoys off the roster more or less sells itself. The drive-thru lane is already almost entirely automated. The phone line is next.
The thesis, and the obvious question
Kea's go-to-market is as novel as the product. The team has pointed AI at its own funnel, using agents to work through a database of hundreds of thousands of restaurants, auto-building a customised demo for each prospect and onboarding a new venue inside a day. When a quarter of demos turn into customers, you are not really selling, you are revealing something the customer already wanted. The market is enormous, with more than 400,000 restaurants in the US alone, and we see genuine defensibility compounding in the integration layer, the menu-ingestion know-how, and a product that gets more accurate with every single call it handles.
The obvious question, and the one we asked ourselves over and over, is "what if Toast just gives this away for free?". It is a fair challenge. Our comfort comes from the depth of what Kea has built, the real switching cost once a restaurant comes to rely on it, and a point-of-sale market so fragmented that the incumbents have so far chosen to partner with Kea rather than compete. We do not need to win the whole market. We need a sliver of a very large one.
Why we invested
We think very highly of Adam and his CTO John. They have been through the fire, they have learned the lessons that only a near-death experience can teach, and they run the business with a discipline that shows in every metric. We also believe EVP has something to add here beyond capital, with obvious cross-sell into Deputy, EatClub and Ordermentum across our portfolio. Ultimately this is a free swing at a genuine land-grab, at a moment in time when voice AI has just become good enough to win, alongside a founder we trust to make the most of it. That is why we invested in Kea.