IHG's Chief Commercial & Marketing Officer Heather Balsley is rebuilding how the company's 6,000-plus properties appear in large language model outputs, positioning the hotel group for what she frames as an inevitable migration from traditional OTA-driven search to AI-mediated booking flows by 2027. The restructuring focuses on data legibility rather than distribution partnerships, a distinction worth tracking.

Balsley outlined the framework at Skift Global Forum, arguing that hotel operators have a narrow window to shape how AI models surface their inventory before consumer behavior calcifies around early algorithmic patterns. The effort centers on making property attributes machine-readable in ways that current SEO and metasearch optimization do not address. IHG is mapping amenity sets, location context, and guest-experience variables into structured formats that LLMs can parse without hallucination or generic categorization. The company is running parallel tests with proprietary models and third-party platforms to measure which structural changes increase accurate mention rates in conversational search results.

The strategic rationale turns on two assumptions. First, that travelers will default to natural-language queries for lodging within 18 to 24 months, the same way enterprise buyers now use ChatGPT for vendor discovery. Second, that visibility in AI-generated recommendations will confer pricing power that OTA placement fees currently erode. If those assumptions hold, the hotel groups that standardize data architecture earliest will capture disproportionate share in a market where algorithmic opacity makes traditional paid placement less reliable. Balsley is explicitly not pursuing co-branded AI booking assistants or white-label chatbot partnerships, which positions IHG's approach as infrastructure investment rather than channel experimentation.

The tell is in the timeframe. 2027 is close enough that Balsley's team is already reallocating budget from metasearch and display retargeting, but far enough out that competitive movement remains limited. Marriott and Hilton have discussed AI visibility in earnings calls but have not disclosed comparable restructuring efforts. Accor's technology leadership has mentioned machine learning for dynamic pricing but not model legibility. The gap suggests IHG is either early or isolated in its assumptions about booking-pattern migration speed.

Operators and allocators should watch three follow-on signals through Q2 2025. First, whether IHG's direct-booking share increases measurably in markets where it has completed data restructuring, which would validate the legibility thesis before broader AI adoption. Second, whether competing hotel groups announce similar initiatives, which would confirm industry consensus on the migration timeframe. Third, whether Google or OpenAI formalize hotel data standards for model training, which would create a de facto compliance requirement and advantage early movers.

The structural question is whether AI models will surface hotel inventory at all, or whether they will route queries to aggregators who have already optimized for algorithmic legibility. Balsley's bet assumes the former. The alternative is that AI platforms default to OTA APIs because those systems already deliver structured, commissionable inventory at scale. IHG's restructuring costs are unknown, but the opportunity cost of mistiming the shift is measurable in basis points of direct-channel share.