IHG Hotels & Resorts Chief Commercial & Marketing Officer Heather Balsley is standardizing property metadata to ensure the group's 860,000 rooms appear correctly when travelers query ChatGPT, Perplexity, or Google's AI Overview—arguing that early structured-data investments will determine which chains survive the shift from click-based search to conversational discovery.
The executive previewed her position ahead of a Skift Global Forum session scheduled for late September, framing AI legibility as infrastructure work rather than speculative innovation. IHG operates 6,400 properties across 19 brands in more than 100 countries, making uniform schema translation nontrivial. Balsley's thesis holds that large language models will surface hotels based on how cleanly their attributes map to natural-language queries—location context, room configuration, accessibility features, brand positioning—and that chains willing to refactor backend systems now will claim disproportionate mindshare when model-mediated bookings scale past 10 percent of direct revenue, a threshold she expects within 24 months.
The strategy accepts that AI agents will intermediate guest relationships without ceding that they will own those relationships. Balsley distinguishes between visibility, which IHG must optimize by feeding models parseable data, and loyalty, which she expects to remain anchored in repeat-stay economics and proprietary program structures that models cannot replicate. The argument acknowledges structural tension: if a traveler asks an LLM for a three-night stay near Singapore's financial district with executive lounge access, the model selects from inventory it can verify in real time—but the guest still books through IHG's direct channel if the brand has constructed that pathway correctly, preserving margin and first-party data capture.
Three implications matter for operators and allocators. First, metadata standardization is now a competitive moat. Chains that can describe their inventory in machine-readable ontologies—down to pillow firmness, minibar contents, and check-in automation—will appear more frequently in LLM-generated itineraries than rivals relying on legacy CMS structures or inconsistent franchise-level inputs. Second, direct-booking infrastructure must anticipate agent-mediated traffic. If a model routes a user to IHG.com but the site requires 12 clicks to confirm a reservation, the guest defaults to an OTA the model also suggested. Third, loyalty programs need legibility layers. IHG's One Rewards has 120 million members; Balsley's task is ensuring models understand tier benefits well enough to surface them as decision variables without requiring the traveler to leave the conversational interface.
Watch whether IHG publicly releases structured-data guidelines for franchisees by year-end, signaling how quickly the parent company believes model-mediated search will affect revenue mix. Also track competitor responses: Marriott Bonvoy and Hilton Honors have comparable member bases and stronger direct-booking funnels in some markets, but neither has articulated a metadata strategy at Balsley's level of specificity. The first major chain to announce a formal partnership with an LLM provider—such as embedding live availability directly into ChatGPT's interface—will set the terms for an industrywide race.
IHG's FY2025 RevPAR growth was 4.2 percent across its system; the question is whether early AI-legibility investments translate to measurable share gains when model-assisted bookings become a tracked channel in 2027 reporting.