News Room

2027 AI Impact Study: Agentic AI use in restaurants more than doubles

Written by Mary Hill Amason | Nov 16, 2026, 3:29:48 PM

New research from Hospitality Technology, powered by Craftable, shows restaurant operators moving from AI that talks to AI that acts. CEO David Cantu explains what that means for your margins.

Download the full 2027 AI Impact Study: Restaurant Edition →

Hospitality Technology's 2027 AI Impact Study: Restaurant Edition surveyed executives across quick-service, fast-casual and full-service brands this summer, and Craftable was proud to power this year's report. The biggest shift: active use of agentic AI in restaurants has more than doubled in a year, from 13% to 32%. Yet only 6% of operators say AI is fully embedded across their core systems and workflows.

In other words, operators are moving fast, but most are still early. Here's what the data shows, and what David had to say about it.

From AI that writes to AI that acts

Generative AI use has leveled off at 59%, down from 65% a year ago. Operator attention has moved to agentic AI: systems that coordinate inventory orders, optimize labor schedules and carry out routine tasks across different software platforms.

Workforce scheduling and real-time labor optimization lead active agentic use at 36%, with another 29% piloting. Automated inventory management is close behind, with 43% of restaurants testing agentic replenishment workflows to control food costs. Still, half of operators are running department-level pilots or exploring informally, so if you feel behind, you have plenty of company.

Back-of-house is where margins hide

Back-of-house use cases take top priority. 62% of operators are evaluating or deploying AI to automate repetitive admin work, and 47% use it for automated inventory replenishment and ordering. David sees this as the start of a longer road, one that moves from reporting to analysis, from analysis to prediction, and from prediction to real intervention.

"Back-of-house is where margins hide, and right now AI is just learning to find them. We're on the road to back-office autopilot, but like autonomous vehicles, it's going to arrive in stages. Lane assist first, then adaptive cruise control, then full self-driving with a human still in the loop." — David Cantu, CEO, Craftable

Trust gets built one stage at a time, he adds, not all at once.

Protect the data you're sitting on

As adoption speeds up, policies are struggling to keep pace. Nearly one in three restaurant companies (29%) let employees use public AI tools without clear tracking or guidelines, and 16% admit they have no concrete risk mitigation guardrails at all. That worries David, who points to the intellectual property most operators hold: recipes, wage data and margin numbers. Frontier models can train on what you feed them, and most people never check the setting that turns it off.

"If you wouldn't post it on the wall in your dining room, don't paste it into a free AI tool." — David Cantu, CEO, Craftable

An insight without a fix is just noise

Where is AI paying off? Faster operational decision-making tops the list of realized benefits at 44%, followed by employee productivity (41%) and better forecasting for inventory and demand (32%). 44% of restaurant leaders say AI has met or exceeded their expectations.

David points to OCR and data capture as a clear win: turning a receipt into general ledger visibility in minutes instead of days, and handing that time back to a GM. But scaling is still held back by data. 38% of operators say their data is fragmented across legacy POS platforms and disconnected systems, which makes AI hard to put to work.

"A problem identified without a path to fix it isn't insight—it's just noise with better formatting." — David Cantu, CEO, Craftable

Keep the GM in the loop

Most brands are keeping people in charge. 56% of restaurants enforce strict human review before AI-generated content or decisions reach a guest, and 38% are only comfortable with guest-facing AI under real-time human oversight. For David, that oversight isn't a cost of AI. It's what makes it work.

"AI has the pattern recognition, but the GM has the context. When a GM catches the AI being wrong, that's not a failure, that's the most valuable data point in the whole system." — David Cantu, CEO, Craftable

Budgets are tighter, and total cost wins

76% of restaurant leaders expect their AI investment to grow over the next 12 to 18 months, a cool-down from 90% a year ago. Few are creating new AI line items. Instead, 38% are reallocating from existing IT and software budgets. When evaluating AI vendors, total cost of ownership is the top factor for 62% of leaders, well ahead of accuracy and security compliance. The takeaway for operators: ask every vendor to show a clear path to lower store-level operating costs.

Get the full study

The full report includes the complete breakdown of agentic AI use cases, the systems AI connects to, ROI by category, vendor selection criteria and the guardrails operators are putting in place.

Download the 2027 AI Impact Study: Restaurant Edition →

Want to see what a connected back office looks like for your restaurants? No generic demo. We'll show you your locations, your numbers, your use case.

See Craftable on your numbers →

The 2027 AI Impact Study: Restaurant Edition is presented by Hospitality Technology and powered by Craftable. Survey fielded July 10 to August 4, 2026.