Headlines this year make it sound like AI has already taken over the order counter: Papa Johns launched a Gemini-powered ordering agent, and McDonald's is piloting an AI voice assistant at the drive-thru. But new industry data tells a very different story about how far AI ordering has actually spread beyond a handful of global chains. Here's what's real, what's still a pilot, and what it means for restaurants and cafés that aren't running a billion-dollar tech budget.
Announced at NRF 2026 in January, Papa Johns became the first restaurant chain to deploy Google's "Gemini Enterprise for Customer Experience" ordering agent — a single AI system meant to take orders consistently across the app, website, phone, kiosk, and even in-car voice systems. Google frames it as solving practical friction points, like helping a customer figure out how much pizza to order for a group. A systemwide rollout is targeted for the end of 2026.
McDonald's has been piloting an AI voice drive-thru assistant, built with Google, at five US locations since June 2026. One franchisee reported that 90% of orders were completed without needing a human to step in. The system handles English and Spanish and can recognize a repeat customer asking for "the usual." McDonald's hasn't committed to a timeline for a wider rollout.
Zoom out from the headline chains and the picture changes quickly. The US National Restaurant Association's 2026 State of the Industry report found that 26% of restaurant operators use some form of AI — but only 6% use it for customer ordering specifically. The rest of that AI adoption is going into marketing (19% of full-service operators), scheduling and labor management, inventory forecasting, and menu optimization — the back office, not the counter.
At the same time, customer appetite is there: roughly six in ten millennial and Gen Z diners say they'd be willing to order from an AI agent. The gap isn't demand — it's that voice/agentic AI ordering is still expensive and complex to build well, which is why it's showing up first at chains with the scale (and the tech partners) to justify it.
A voice or chat ordering agent that actually works well needs to be trained on a specific, well-structured menu; integrated cleanly with the POS and kitchen systems; and tested across thousands of orders to catch edge cases before it embarrasses the brand. That's realistic for McDonald's or Papa Johns, who can dedicate engineering teams and negotiate directly with Google. It's not yet realistic — or necessary — for most independent restaurants and cafés, including the vast majority of Kuwait's F&B market.
The good news is that most of what customers actually want from AI ordering — speed, no friction, no app to download, instant updates — is already achievable today with a well-built digital menu. A QR-code menu that opens straight in the browser, updates prices and availability in real time, supports multiple languages, and lets customers order and pay without waiting for a server gets you most of the customer-experience benefit that AI ordering is chasing, without needing enterprise-grade AI infrastructure.
Think of it less as "AI ordering vs. digital menus" and more as sequencing: the digital menu is the foundation any future ordering agent — AI-powered or not — will eventually sit on top of. Restaurants that already have clean, structured digital menus and direct ordering in place today will be the ones best positioned to adopt AI ordering tools cheaply and quickly whenever they become affordable for businesses their size. Restaurants still relying on printed menus and phone orders will have to build that foundation first, AI or no AI.
