As AI adoption gains momentum in the restaurant industry, many quick-service restaurant operators are balancing curiosity with caution. And according to Restaurant365’s How to Use AI Safely in Restaurants webinar, the real challenge is making sure teams use AI tools in ways that protect sensitive financial, employee, and operational data.
“One of the biggest misconceptions is that people believe the AI model itself is the main risk, but that’s not the case,” says David Mortensen, customer success manager for Restaurant365. “The real risk, especially in the restaurant industry, is the sensitive data that people might enter into a public AI system. That information could then be used in uncontrolled ways. The key to using AI responsibly means having guardrails in place.”
The first step Mortensen recommends when implementing AI is approving a small set of paid, secure AI tools. “In other words, don’t let it become the wild West where one person is using ChatGPT, someone else is using Claude, and someone else is using something totally different,” he says. “The second step is to define a clear ‘do share this, don’t share that’ policy. Then you can build a list of the kinds of information that is safe to use.”
Some data Mortensen recommends avoiding are: payroll information, Social Security numbers, wage data, banking and routing information, and employee last names. Mortensen also suggests not mentioning full P&Ls or balance sheets, vendor contract pricing, guest contact information or loyalty data, and proprietary recipe information.
Another important guardrail is assigning one or two internal AI experts. Then when employees have questions or concerns, they know exactly who to go to. “Lastly, don’t share passwords,” Mortensen says. “Everyone should have their own login so activity can be tracked. That way, if there’s a problem, operators know who did what.”
Once guardrails are set and AI is fully implemented into everyday operations, teams can start effectively using it. “I think the first step in using AI on a day-to-day basis is to treat it like a powerful team member,” Mortensen says. “What I mean by that is: you wouldn’t trust any employee with Social Security numbers, salary information, or vendor contract pricing. So, if operators start thinking of AI as if it were another team member, that helps them think clearly about what guardrails need to be in place.”
With this in mind, Mortensen also emphasizes the importance of human decision makers. “AI should only be used as a tool,” he says. “For example, just because someone on staff suggests doing something a certain way, doesn’t automatically mean it’s the right way. The same is true for AI. Operators still need someone who understands the business to validate what it produces.”
Some of the safest and most practical use cases for AI in restaurants today are simple tasks. “I think one of the easiest early wins is drafting emails,” Mortensen says. “Job postings are another great starting point. AI can also be useful for creating training programs or standard operating procedures. Another helpful use case is analyzing non-sensitive operational data like inventory or sales patterns. AI can help surface trends, identify top-performing items, or highlight areas that need improvement.”
Looking ahead, AI will continue to evolve in the restaurant industry and operators need to be prepared. “I think operators need to get ahead of AI now by standardizing their internal framework,” Mortensen says. “That means identifying approved tools, defining guardrails, assigning owners, and asking vendors the right questions. If operators do this now, they’ll be able to take full advantage of AI without scrambling later in response to a data issue or loss of trust.”
To learn more about safe AI implementation, watch Restaurant365’s How to Use AI Safely in Restaurants webinar.
By Abby Winterburn
The post How To Safely Implement AI For Early Wins appeared first on QSR Magazine.