Having AI Isn't Having a Strategy

“A system that predicts a rate is a fundamentally different proposition from a system that decides to change it without asking.” That’s LodgIQ President and CCO Mark Charlinski’s core complaint about how the hotel industry talks about AI: predictive, generative, and autonomous systems get flattened into one undifferentiated buzzword, the same way “cloud” became meaningless marketing shorthand a decade earlier. Writing in a Hospitality Net opinion piece, Charlinski points to the Hotel Data Conference, where he says nearly every panel invoked “AI” but rarely specified which type it meant.

His deeper argument is about sequencing: most hotel organizations start with the technology and search afterward for a problem to apply it to, rather than starting from a real operational problem and working backward to the right tool. He cites an audience example — using AI to solve a specific conference checkout-extension logistics issue — as the model for problem-first thinking, and recommends internal AI hackathons as a way to surface genuine operational problems before any technology gets scaled.

The coherence gap Charlinski is describing shows up concretely elsewhere in hospitality coverage this month. One Hospitality Net piece documents hotels running five separate, disconnected AI systems — one each for pricing, guest messaging, booking, front desk, and upselling — with no shared source of truth between them, illustrated by a New Orleans property where the messaging AI promised something the pricing AI had no knowledge of. And another offers a practical antidote: a decision framework distinguishing Chat, Workflow, Agent, and Orchestration by how much of a task’s path can be flowcharted in advance, arguing “most hotel work is a workflow dressed up as an agent problem” — pre-arrival messaging and standard reporting rarely need the autonomy hotels are buying for them.

Buying the right label isn’t the same as having a plan for what it’s supposed to do.