This insight summarizes Mark Charlinski's Hospitality Net opinion on separating AI capability from business strategy, prompted by discussions at the Hotel Data Conference.
“AI” is hiding three different propositions. Predictive systems forecast outcomes, generative tools create reports or content, and autonomous agents take action. Treating them as interchangeable prevents leaders from asking the questions that matter—especially the difference between software recommending a rate and software changing it without approval.
The industry has seen this hype cycle before. “The cloud” became so broadly applied that it stopped communicating anything useful. Adoption moved faster than understanding, and problems around misuse, security and data integrity followed. AI risks the same fate when vendor language replaces operational clarity.
Start with the friction, not the feature. Charlinski highlights a practical conference example: identify attendees in a room block, extend their checkout and resequence housekeeping so they can attend an 8 a.m. session without checking bags. The value of the idea is its order of operations—a specific guest and workflow problem first, then a search for the right tool.
Build capability close to the work. An internal AI hackathon can give employees time to select real problems, test solutions and reveal who can use the technology effectively. Starting small produces faster evidence than an abstract transformation program and creates room to expand only after a use case proves itself.
Takeaway for hotel leaders. Name the business problem, the decision owner and the intended outcome before discussing models. Be explicit about whether the system predicts, generates or acts. Then measure whether recommendations are trusted and used at scale; sophistication under the hood creates no return when the operation does not change. AI can support a strategy, but it cannot substitute for one.