What Hilton's Planner Is Actually Pulling From — the AEO/GEO Arms Race

I spent an hour this morning typing the same eight prompts into ChatGPT, Gemini, Perplexity and Copilot from a hotel lobby in Midtown. “Best family-friendly hotel in Austin for spring break.” “Boutique hotel near the Las Vegas Strip with a pool.” “Where should I stay in Nashville under $400 a night.” The point of the exercise was not to find a hotel. It was to see whose name the engines volunteered before I asked them to.

The pattern was obvious before I finished the second prompt. ChatGPT named Hilton properties in almost every answer, often twice. Gemini named Moxy and 1 Hotel as often as it named Hilton. Perplexity cited the chain brands by name and the lifestyle brands by neighbourhood. Copilot was the outlier — it volunteered loyalty programmes more than properties. (Interpretation flag: eight prompts in one lobby is not a study. The study landed in my inbox while I was running them. More on that below.)

The contrarian thesis, after yesterday’s Hilton AI Planner audit: the Planner on Hilton.com is the smaller half of the chain’s 2026 AI bet. The larger half is the work that gets a Hilton property named first when somebody asks ChatGPT where to stay in Austin. The on-site Planner is defence. The model-layer work is offence. The category has a name in the SEO trade press now — AEO, for Answer Engine Optimization, and its sibling GEO, Generative Engine Optimization — and the spring-break numbers say the chains are already winning it.

The AIVO numbers, in operator language

AIVO Research’s spring-break 2026 hotel AI visibility study, released earlier this week, ran 660 spring-break-shaped travel prompts across four engines — ChatGPT, Gemini, Perplexity, Copilot — and tracked mentions and positions for sixteen hotel brands. Three numbers carry the story.

Hilton gets roughly ten times more ChatGPT mentions than any single lifestyle competitor. Across the 660 prompts, Hilton’s surface-area share inside ChatGPT answers is in a different weight class. If a traveller is asking ChatGPT a generic spring-break question, the chain name surfaces almost by default. That is the visibility number Hilton’s brand team is selling internally.

Moxy and 1 Hotel rank higher in position despite far fewer mentions. When a lifestyle brand does appear in an answer, it tends to appear earlier in the response, in a recommendation slot rather than a backgrounder. The chains win share-of-voice. The lifestyle brands win share-of-recommendation. Operators who only read the mentions chart miss the second story.

ChatGPT skews 4.8:1 toward chains. Gemini skews 1.6:1 toward lifestyle. This is the chart I keep re-reading. ChatGPT is a chain-friendly engine. Gemini is the engine where a Moxy or a 1 Hotel actually has a fair fight. Perplexity and Copilot sit between them. If you are a lifestyle operator and your AEO budget is going entirely into ChatGPT prompt-tests, you are optimising on the surface where you lose by design.

AEO and GEO are the new RevPAR-adjacent metric

Two acronyms worth learning before your next budget meeting. AEO — Answer Engine Optimization — is the work of getting your brand named correctly inside an AI answer. GEO — Generative Engine Optimization — is the work of shaping the upstream content the engines read, so that the answers come out the way you want. AEO is the audit. GEO is the lever.

The reason both matter in March 2026, and did not in March 2025, is the demand-side shift the AIVO study quantifies. When ChatGPT, Gemini, Perplexity and Copilot become the primary surface where travellers ask “where should I stay,” the brand surface area inside those answers is the new top-of-funnel. It is not OTA share. It is not metasearch share. It is engine share. And it is not auctioned — it is earned.

Hilton’s on-site Planner is a separate surface — a conversion tool for a guest already on Hilton.com. The model-layer work is the acquisition tool for a guest who has not opened a hotel site yet. Forthcoming coverage from Business Traveller will, I expect, cover the Planner UX in detail. The AEO/GEO question — what feeds the engines themselves — is the one that does not get answered by a UX review.

What I want operators to take from this

Three things, in budget order.

One: your AEO baseline is a Tuesday-morning exercise, not a vendor procurement. Run the same eight prompts in the four engines once a week. Log the mentions, log the positions, log the engine. You will have a defensible chart in six weeks. Do not buy a tool for this until you have the chart.

Two: lifestyle and independent brands should over-index on Gemini. The 1.6:1 split inside Gemini is the closest thing to a fair fight in the visibility landscape. The content moves — neighbourhood-anchored, design-led, named-chef — that surface inside Gemini do not surface the same way inside ChatGPT. Pick the engine where you can win, not the engine where the chains already won.

Three: the OTAs are watching the same chart. The forthcoming OpenTable–Booking AI piece and the Marriott AI replatforming column both sit on the demand-side half of this story. If the engines become the top-of-funnel, the OTAs lose the link they used to own, and the chains regain a margin point. That is the trade the AEO/GEO work is really making.

The Planner on Hilton.com is the announcement. The model-layer work is the strategy. Watch which one shows up in the next earnings call.

— Naomi covers hotel F&B and operator tech for TableTransfers. Tips: hotels@tabletransfers.com.