OpenTable Quietly Has the Most Interesting AI Strategy in Reservations

I am watching the Fortune Brainstorm Tech livestream from a coffee shop two blocks off the Strip, because the badge line in Las Vegas this morning was the kind of thing that makes you reconsider your life choices. CES is still technically on — it closes tomorrow — and the AI track at Brainstorm has absorbed half the post-keynote crowd. The panel I came for is the AI agents one, and the name on the slate I am most interested in is not the Salesforce executive or the Land O’Lakes CIO. It is Sagar Mehta, the chief technology officer of OpenTable.

Mehta does not get the marquee treatment that Marc Benioff does. He does not have a Beyoncé-at-Dreamforce moment queued up. He sits, he answers questions, and over the course of roughly twenty minutes he says enough to make me close my laptop and rewrite the post I had been drafting on the flight in. Here is the contrarian read, which I will defend for the rest of this column: OpenTable is quietly the most interesting AI play in restaurant reservations right now, and the people watching Resy and Tablecheck for the next move are watching the wrong table.

What Mehta actually said on stage

The Fortune writeup of the panel hits the headline beats, but the texture matters. Mehta confirmed two things that have been rumored in industry Slacks for months. First, OpenTable is deploying a voice-AI booking experience — an agent that takes phone calls into restaurants, handles the “table for four at seven-thirty” workflow, and writes the reservation back into the OpenTable system of record. Second, OpenTable is rolling out AI-powered review responses for operators: drafted, brand-aware replies to the Tripadvisor-style flood that every general manager has been ignoring since 2019.

Neither feature is, on its own, revolutionary. Voice agents on restaurant phone lines are a crowded space — there is a startup spawning one every time a Y Combinator batch closes. AI-drafted review responses are table stakes for any reputation-management vendor. What is revolutionary is who is shipping them. OpenTable is not a startup. It is the dominant Western reservations platform, embedded in roughly fifty-five thousand restaurants, and owned by Booking Holdings, which also owns Booking.com, Priceline, Kayak, and Agoda.

When Mehta said “agents,” he was not waving at a roadmap deck. He was describing capabilities that are either in production or in late-stage pilots with named restaurant partners. My read: that is a different posture than every reservations competitor I have talked to this quarter.

Why Booking Holdings’ distribution moat matters

Here is the part the AI-pure-play set keeps underrating. OpenTable does not need to win the AI agent race on technical merit alone. It needs to win on distribution and integration, and on that axis the Booking Holdings parentage is a structural advantage that nobody else in reservations can replicate.

Think about what a voice-AI booking agent actually needs to be useful. It needs the inventory. It needs the restaurant’s table configuration, pacing rules, special-occasion notes, dietary tags, and the guest’s history with that specific venue and across the network. OpenTable already has all of that, because the restaurants have been running their hosts on the OpenTable Guest Center for over a decade. Resy has a chunk of it. Tablecheck has it in APAC. Nobody else is even close.

Now layer in Booking. If OpenTable’s eventual concierge can answer “find me a Tuesday-night tasting menu in Lisbon near my hotel,” the hotel side of that query lives one engineering meeting away inside the same parent company. That is the distribution moat I keep gesturing at. The Resy-American Express partnership is real and powerful for cardholder benefits, but it does not solve the “I am in a city for the first time and I want a great meal near my room” problem in the way Booking’s footprint can. As our later coverage of OpenTable’s AI rollout argues, this is why the Booking integration is the load-bearing piece, not the LLM.

My read: the moat is the data plus the cross-vertical distribution, not the model. Anyone can wire up GPT-class inference. Almost nobody can wire it up to fifty-five thousand restaurant inventories and a global hotel graph in the same product surface.

The Concierge tease, decoded

Mehta did not formally announce a Concierge product on stage today. He gestured at one. The OpenTable AI Concierge press page is the public anchor for that conversation, and the framing he used today maps cleanly to it: a diner-facing assistant that handles discovery, booking, modification, and post-meal review prompts in natural language, across the OpenTable network.

I want to be careful about chronology here, because the timeline is the story. The Concierge general-availability launch is forward — on the roadmap, not in production for consumers as I write this. What Mehta confirmed today is the direction and the underlying agent infrastructure. The voice agent for inbound restaurant calls and the review-response tool are the operator-facing wedge. The Concierge is the diner-facing crown jewel that the same agent platform is being built to enable.

If you want the steel-man version of why this is non-trivial: every reservations platform has a “discovery” surface. OpenTable.com has search and filtering. Resy has the editorial guide. Tablecheck has the local-language listings. Concierge is the first credible attempt to collapse discovery, booking, and modification into a single conversational interface at OpenTable’s scale. The bet is that a meaningful share of diners will, within the next product cycle, prefer to type or speak to an agent rather than scroll a list.

My read: the Concierge is the right ambition, the partial-rollout cadence is the right tempo, and the fact that it is forward-dated rather than shipped today is actually a credibility marker. The platforms that announce GA at a panel are the ones that have not built the thing.

What this means for Resy and Tablecheck

This is where my contrarian thesis gets sharp. Resy has been the cool-kid platform of the last five years, and American Express is a wonderful parent for cardholder marketing, but I do not see the Resy AI roadmap that competes with what Mehta described today. The Resy product surface still leads with editorial guides and a notify list. That is a 2019 strategy in a 2025 market.

Tablecheck has a stronger APAC position and a better operator product than most people in the US realize. But its AI story, at least the one in market today, is closer to a workflow assistant inside the back-of-house product than a diner-facing agent layer. The Resy and Tablecheck question is not whether they can ship AI features — they obviously can. The question is whether either has the data graph and the distribution surface to ship a credible Concierge competitor in the same window.

And then there is the SevenRooms angle, because the Crunchbase company timeline for SevenRooms is the cleanest reminder that the high-end / hospitality-CRM segment is consolidating in a different direction. SevenRooms is winning on CRM depth and operator workflows for high-touch venues. That is a different product than OpenTable’s mass-market reservations layer, and a different product than a diner-facing agent. As a piece we later publish on Resy and the reservation stack lays out, the segment is splintering into three competing center-of-gravity bets, and OpenTable’s AI bet is the most defensible of them.

My read: the next twelve months are going to look like OpenTable shipping incremental agent capability, Resy and Tablecheck announcing their own AI features, and SevenRooms doubling down on the operator-CRM moat. The first move was today, on a panel stage in Las Vegas, and most of the industry missed it because Mehta does not do keynotes.

I will be at NRF in New York next week with this thesis in my back pocket. If you are an operator running OpenTable, Resy, Tablecheck, or SevenRooms and you want to tell me what the AI features actually look like inside your back office, the inbox is open.

— Maya covers restaurant tech. Tips: tips@tabletransfers.com.