Toast IQ Launches — a Conversational AI Assistant Across ~148,000 Locations

I am standing at a coffee bar in the Seaport, two blocks from Toast HQ, watching an operator I have known for three years pull out her phone and ask her POS a question in plain English. The question is unglamorous — something about why beer sales dipped on Tuesdays — and the answer comes back as a short paragraph with two suggested actions, not a dashboard. She looks at me and says, “Okay, this is different.” That is the day Toast formally launches Toast IQ as a conversational AI assistant, and I will say it plainly: this is the most ambitious consumer-grade product Toast has ever shipped to its own customers.

The pitch, as Toast frames it in its launch announcement on Restaurant Technology News, is that IQ now pulls from “real-time and historical data across approximately 148,000 customer locations” and translates that scale into something an owner can actually use during a shift. The two front doors are a “For you” feed of prioritized insights and a plain-language Q&A surface that sits across the platform. The thesis I want to test today is whether those two surfaces, together, change Toast from a transaction layer with reports attached into something that looks and feels like an assistant.

What actually shipped, in operator language

If you have used Toast’s Smart Features over the last year — the AI-suggested menu descriptions, the labor flagging, the inventory variance nudges — you have already seen the substrate IQ is built on. The change is that those features were point tools, surfaced inside the screens where you happened to be. The conversational layer makes them addressable. You can now ask, in a sentence, what the system used to make you click through three menus to assemble. That is a small interaction change with a big behavioral consequence: it lowers the activation energy for asking questions of your own data from “five minutes and a spreadsheet” to “ten seconds and a sentence.”

The “For you” feed is the other half of the bet, and it is the half I find more interesting. Yahoo Finance’s coverage of the expansion frames it as a prioritization layer — the system decides what an operator should look at before they think to look. For a single-unit operator who logs in twice a day between shifts, that is a wildly different relationship with software than a homepage full of widgets. It is closer to a notification surface than a dashboard, and the closest analog I can think of is the way a good GM walks into the office and an assistant says, “Three things before you start.” The product is trying to be that assistant.

I want to be careful here. Feeds are easy to demo and hard to keep useful. The failure mode is not technical — it is editorial. If the system surfaces five “insights” a day and three of them are noise, operators learn to ignore the feed within a week, and the whole interaction model collapses back into a dashboard with extra steps. Toast knows this; the question is whether the prioritization model can earn trust at the pace of restaurant attention spans, which is to say, fast.

The Coca-Cola tell

The detail that I keep coming back to is the Coca-Cola partnership. TotalFood’s write-up of the launch carries a quote from Josh Gurley at Coca-Cola explicitly framing the integration around beverage attach — the idea that IQ can identify checks where a drink should have been sold and wasn’t, and prompt the operator (or eventually the server, or eventually the kiosk) to close that gap. That is not a vanity partnership. Beverage attach is one of the highest-margin levers in a casual operation, and it is one of the levers operators consistently leave on the table because nobody is watching every check in real time.

Read between the lines and the Coca-Cola tie-in tells you what Toast thinks IQ is for. It is not, in the first instance, an analytics product. It is a revenue product dressed as an analytics product. The “insight” is just the payload; the action — sell one more drink per ten checks — is the thing being sold to the operator and, through Toast, to the beverage partner. If you assume other category partners are next (snack, dessert, premium spirits in full-service), the shape of the business model gets a lot more interesting than “AI feature in a POS.”

The forthcoming context, and the international footnote

I have been writing about Toast IQ since earlier in the year — there is a forthcoming May piece that gets into the original Smart Features framing, and an upcoming desk review that walks through how the assistant behaves when you actually try to break it on a slow Tuesday. I will not preview those pieces here, but they are the right places to go if you want the longer arc on how Toast got from “AI-suggested menu descriptions” to “ask the POS a question.”

The other thing worth flagging from today’s announcements is that Toast is planning international expansion of IQ. The U.S. footprint of roughly 148,000 locations is the data substrate that makes the assistant plausible — there is enough signal at that scale to make “For you” non-trivial. International rollout is a harder problem than it looks, because the prioritization model has to be re-tuned per market: a Tuesday beer dip in Boston is a different artifact than a Tuesday lager dip in Manchester or a Tuesday wine dip in Dublin. The announcement gives no date, and I am not going to manufacture one. I will note that the order in which Toast picks its first international IQ markets is going to tell us a lot about where the company thinks the next 148,000 locations live.

Mark the interpretation

Here is how I am marking today, on the record. Toast has crossed a line that most of its peers have not: it has stopped shipping AI as a set of features inside screens and started shipping AI as a surface — a feed and a chat — that sits on top of the platform. That is a consumer-grade product decision, not an enterprise-software decision, and it is the right call for an operator base whose median user is a single GM with a phone. Whether IQ keeps the trust it is being handed depends entirely on the editorial discipline of the “For you” feed over the next two quarters. The Coca-Cola partnership tells me Toast already knows where the money is. The international rollout will tell us whether the model travels.

I will check back in after a quarter of real shifts.

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