RLC 2025 Opens in Phoenix With 1,700 Operators and a Voice AI Showcase

I got into Phoenix Sunday afternoon, dropped my bag at the JW Marriott Desert Ridge, and walked the show floor while the crews were still uncoiling cables and zip-tying signage to truss. There is a particular kind of quiet you only get on a Sunday-night exhibit hall — the carpet smell, the half-built booths, the rep in a polo shirt frowning at a laptop because the demo loop isn’t looping. By the time I’d done a slow lap, I’d counted six booths with “Voice AI” in the headline copy and three more where it was the subhead. That count is up from last year. The pitch is changing too, which is the actual story.

The Restaurant Leadership Conference runs April 13 through April 16 this week. Roughly 1,700 attendees on the operator side, 160-plus exhibitors, 65-plus sessions across AI, cybersecurity, and labor cost tracks. Informa reports 84% of registered attendees are VP-or-above, which tracks with what I saw at the welcome reception — fewer lanyards reading “Sales Engineer,” more reading “Chief Operating Officer.” Vendelux is tracking 3,416 professionals across attendees and exhibitors combined, which is the number I’d quote if a vendor asked me whether it’s worth flying out. It is.

The floor on Sunday night

The thing you notice walking the hall before it opens is what booths chose to put on the back wall — the giant printed claim that has to read from across the carpet. Two years ago those back walls said “AI-Powered.” Last year they said “Generative AI for Restaurants.” This year, at the booths I cared about, the back walls say things like “92% order accuracy at 14 locations” and “Pilot to rollout in 11 weeks.” That is a different document. That is a procurement document.

I bumped into a multi-unit franchisee from the Southeast — sixty-some QSR units across two brands — who was doing his own pre-walk. He’d flown in a day early specifically to talk to three voice-AI vendors he’s been emailing with for a year, and he had a single sheet of paper in his hand with five questions on it. I asked if I could see it. He laughed and showed me the top one: “Who else in my MSA is live with you, and can I call them tomorrow?” That is question one. Not features. Not roadmap. References, in his metro, that he can phone before he leaves Phoenix.

A VP of digital at a top-25 burger chain told me over a beer that her team’s RLC mandate this year is “no decks, only customer calls.” Her exact line: “If a vendor pulls up a slide deck I’m walking. Show me a live drive-thru lane in production. Otherwise we’ll talk in 2026.”

What the voice-AI vendors are leading with this year

So here’s the contrarian read, the reason I think this RLC matters more than the last two: the voice-AI category has finally stopped pitching the product and started pitching the deployment math. The good booths this morning aren’t demoing the model — every operator at this conference already believes the model works well enough at the speaker. What they don’t know is what it costs to integrate, who owns the menu-keeping, how the AI hands off to a human when it fails, and how long the pilot-to-rollout curve actually is once you account for the franchisee meetings.

The vendors who get that are leading with case studies and unit economics. The vendors who don’t are still showing waveforms. The waveform booths will not be here in 2026.

Two specific shifts I saw walking Sunday. First, almost every voice-AI booth I passed had a printed handout — not a tchotchke, a handout — with a “deployment timeline” graphic on it. Eight weeks, eleven weeks, fourteen weeks. That number used to be hidden. Now it’s the headline. Second, the integration story has gotten honest. The booths are naming their POS partners on the back wall, which means they’ve stopped trying to be everyone’s friend and started picking sides. We’ll cover the POS-side angle — what Toast and the ToastIQ stack are doing here — in a later Pass piece we publish on POS-side AI. For RLC, the relevant fact is which voice vendors are showing up co-marketed with which POS, and how confidently.

Cybersecurity is the sleeper track. A CISO from a casual-dining chain told me Sunday night that her real RLC reason is the cyber track, not AI — because the AI vendors all want access to her order data and her PCI scope, and she has to decide which ones get it. That conversation is happening at this conference, quietly, in hotel-bar booths, and it will shape who closes deals in Q3.

Why Hochman as Leader of the Year signals where the conversation is

Brinker CEO Kevin Hochman is being named Restaurant Leader of the Year here this week. That choice is not random. Hochman ran the Chili’s turnaround on operational basics — labor model, menu discipline, throughput at the table — which is exactly the lens RLC’s program is pushing the AI conversation into this year. The award isn’t about the flashiest tech adopter. It’s about the operator who fixed the business and then layered tech onto a healthy P&L, in that order.

Read the program through that lens and the AI track makes more sense. The sessions aren’t “Will AI replace the GM.” They’re “What’s the actual labor-hour reduction at the speaker, and what does it cost to get there.” The pre-conference coverage from SaveFryOil framed it the same way — labor cost is the headline track, and AI is one of the tools inside it, not the topic above it.

Wendy’s Q1 lands at the end of the month and I expect AI commentary; the Yelp Spring Product Release on April 29 and the next Hilton AI Planner update are also forward of this week. I’ll be filing again from the floor Tuesday and Wednesday — there’s a session I want to sit in on about voice-AI accuracy benchmarking that I have a strong suspicion will be the most-photographed slide of the week.

For now: Phoenix is hot, the carpet smells like new carpet, and the vendors who brought references are going to leave with deals.

— Samuel hosts the Service podcast and files occasional Pass pieces from the road. Tips: [email protected].