Fast-food giant Burger King is launching an AI chatbot, “Patty,” to monitor employee interactions at hundreds of U.S. locations. Part of the BK Assistant platform powered by OpenAI, it helps managers understand service patterns by detecting polite phrases like “please” and “thank you.” It also assists with operational tasks, such as updating digital menus and providing coaching insights for drive-thru accuracy. Despite social media criticism, Burger King clarified that the chatbot is not meant to evaluate individual employees but to improve service quality and support teams. The BK Assistant will be deployed nationwide by the end of 2026, after a pilot in 500 restaurants.
ServiceNow claims it resolves 90% of employee IT requests automatically, with case resolution 99% faster than humans. Its new product, EmployeeWorks, lets employees describe issues in plain language without submitting tickets. The challenge isn’t AI capability but governance and workflow continuity. ServiceNow’s Autonomous Workforce framework and ‘role automation’ ensure AI inherits permissions and governance from human workers, preventing scope creep and boosting trust. Alan Rosa of CVS Health stresses embedding governance in AI deployment to maintain operational integrity.
A YouGov survey for Pega shows 68% of consumers lack confidence in businesses’ use of generative AI for customer service, with over half concerned about responsible use. Simon Thorpe, Pega’s director, says declining trust results from frustrating AI experiences, with 50% reporting unsuccessful interactions with AI-only systems. While 80% prefer human contact for better results, only 2% want to engage exclusively with AI chatbots. Despite concerns, 91% of leaders feel compelled to adopt AI to boost customer satisfaction and efficiency.
Why do we care?
The ServiceNow and YouGov numbers aren’t contradictory. ServiceNow claims 90% autonomous resolution. Half of consumers report unsuccessful AI-only interactions. Both can be true — ServiceNow counts closed tickets; consumers measure whether the problem was actually resolved acceptably. MSPs who benchmark their AI service desk performance against vendor metrics rather than client satisfaction scores are optimizing for the wrong outcome, and they won’t know it until the client is already looking for a replacement.
Burger King’s Patty is the clearest illustration of what happens when AI governance is treated as a labeling problem rather than a design problem. Call it coaching, deploy it in headsets, analyze audio in real time — and then be surprised when the legal exposure surfaces. Many states require disclosure for continuous workplace audio monitoring. “The vendor framed it as support” is not a compliance defense, and any MSP recommending similar employee-facing AI tools without a state-specific legal review is carrying that liability on behalf of their client.
And if that liability triggers a workplace privacy claim, the first question from the carrier will be whether monitoring disclosures and consent requirements were reviewed and documented. If the answer is no, coverage becomes a negotiation — not a guarantee.
The through-line across all three stories is that AI is being deployed at the human interaction layer faster than the trust, consent, and governance infrastructure can support it. The 91% of leaders who feel compelled to adopt AI regardless are the ones producing the failed experiences that erode the trust they’re trying to build. MSPs who help clients deploy AI as a trust architecture — not just a cost reduction — own a defensible position. That defensibility isn’t philosophical. It shows up in contracts, audit trails, and insurability. AI without documented governance doesn’t just erode trust — it increases the probability of uncovered claims. Everyone else is accelerating the problem.