Intro
Every booking made through an intermediary costs a hotel group between 15% and 25% in commission. That margin is the single biggest lever a revenue manager and a CMO share — and the one most often left to loyalty programmes and parity rules alone.
AI answer engines have quietly become the upstream layer of the traveller journey. When a guest asks ChatGPT, Perplexity, Gemini or Google AI Overviews for “best boutique hotels in Lisbon with a rooftop” or “five-star resorts in the Maldives for honeymoon”, the answer either names the hotel’s own domain — or it names Booking, Expedia, or a metasearch aggregator. That choice is now made before any pricing page is seen.
This playbook explains how to measure that choice, change it, and read the results quarter over quarter.
Score your direct booking exposure
Why AI visibility is upstream of direct booking
Direct booking strategy is usually framed as a downstream problem: parity rules, member rates, loyalty perks, retargeting. All of it works only once the traveller already knows the hotel and reaches the brand.com page.
AI answer engines sit one step earlier. They are now where shortlists are built. If the hotel’s own domain is absent from those answers, the only way back into the funnel is paid: meta-search bidding, OTA placements, or branded search defence. Each route reintroduces commission, ad spend, or both.
Winning AI citation share for high-intent prompts pulls the traveller directly onto own-domain. It does not replace loyalty, parity, or retargeting — it widens the top of the direct funnel so those downstream tools have more volume to convert. The cost of staying invisible is paid every quarter in the OTA mix.
The six-phase playbook
The work runs in six phases. Each one is measurable, each one has an owner, and each one feeds the next.
1. Measure current AI citation share
Run a baseline across the AI engines that matter for the hotel’s source markets. Capture which prompts surface the hotel’s own domain, which surface intermediaries (Booking, Expedia, Tripadvisor, Kayak, regional aggregators), and which surface editorial titles. The baseline is the only honest answer to “how exposed are we today” — and the only thing later quarters can be compared against. The Capston Hospitality Scorecard frames the eight dimensions used here.
2. Prioritise the top 10–20 commercial prompts
Not all prompts carry the same revenue intent. A “best of” query in a feeder market converts very differently from a generic discovery query. Rank the prompt library by commercial weight: source-market size, average length of stay implied, package value, season. Lock the top 10 to 20 as the priority set. Everything else is monitored, but not actioned this quarter.
3. Close own-domain authority gaps
For each priority prompt, audit the hotel’s own domain on three layers: entity (does the brand resolve as a clean entity across knowledge graphs), schema (Hotel, LodgingBusiness, FAQPage, Review aggregate, Offer where appropriate), and content depth (does the page actually answer the prompt or simply list the property’s features). Most hotel sites are built for brochure, not for answer extraction. Closing the gap means rewriting key pages so AI engines can lift a clean passage. See the Capston Core methodology for the five-stage process.
4. Earn editorial coverage in 4–6 travel media titles per market
AI engines reuse trusted sources. Four to six travel and lifestyle titles per source market — Condé Nast Traveler, Travel + Leisure, National Geographic Traveller, Suitcase, Monocle, regional broadsheets — carry disproportionate weight. The work here is earned, not paid: a press programme tied to the priority prompts, not a generic media push. One citation in the right title shifts more answers than ten citations in low-trust outlets.
5. Build the answer-shaped page for each priority prompt
For every prompt in the priority set, the hotel’s own site must have one page that answers it directly, in extractable form: a clear summary, factual specifics (location, rooms, distinctive feature, price band, season), structured schema, and an offer or booking module above the fold. This is not landing-page copywriting — it is answer engineering. The page exists so AI engines have something canonical to cite.
6. Measure the shift in citation share quarter over quarter
Rerun the baseline every 90 days against the same prompt set, the same engines, the same competitor set. Track three numbers: own-domain citation share (target: rising), intermediary citation share (target: falling), editorial citation share (target: stable, with the hotel named in the cited piece). The shift in citation share is the leading indicator. The shift in OTA mix on the P&L follows by one to two quarters.
How this complements existing direct booking work
This playbook does not replace anything a strong commercial team already runs. Loyalty programmes, member rates, best-rate guarantees, parity enforcement, retargeting, owned-channel CRM, app-only perks — all of that continues. It widens the funnel, it does not redesign it.
Specifically, it pairs with parity and rate strategy: if the hotel’s own page is the answer cited by the AI engine, the traveller lands on brand.com before comparing OTA rates, and the parity rule does its job. It pairs with loyalty: more direct first-touch means more enrollable guests. It pairs with the OTA capture defence playbook, which addresses the downstream half of the same problem — what happens when the AI engine routes the traveller to an intermediary anyway.
Revenue management still owns inventory, pricing and channel mix. Marketing still owns brand and demand. This playbook gives both functions a shared metric — citation share — that sits cleanly upstream of the variables they already steer.
What 8-week and 16-week milestones look like
Week 1–8 (Foundation): baseline complete, priority prompt set locked, own-domain audit delivered for the priority 10–20 prompts, first three answer-shaped pages live, editorial programme briefed to PR. Internal stakeholders aligned on the metric.
Week 9–16 (First measurable shift): remaining answer-shaped pages live, first two to four editorial placements earned, second baseline rerun complete. First read of citation share movement. By week 16, a hotel group that started below industry average on own-domain citation share should see a measurable rise on the priority prompts — not on every prompt, and not in every engine, but on enough of the priority set to validate the investment and shape the next quarter.
P&L impact is not the week-16 read. The week-16 read is citation share. P&L follows, one to two quarters behind.
How this fits into Capston Core
Direct booking recovery is one of the commercial playbooks Capston Core supports for the hospitality vertical. It uses the Capston Core methodology for the five-stage process, the Capston Hospitality Scorecard for measurement, and the OTA capture defence playbook for the downstream half of the problem.
→ Back to Capston Core
FAQ
How long before this shows up in the booking mix?
Citation share moves within one quarter of disciplined work on the priority prompts. The shift in OTA-versus-direct mix on the P&L typically follows one to two quarters behind, because traveller decision cycles for higher-value stays are not instant.
Does this replace our loyalty programme or parity rules?
No. It widens the top of the direct funnel. Loyalty, parity and retargeting still do the conversion and retention work downstream. Without AI visibility upstream, those tools operate on a smaller pool.
Which AI engines should we measure?
The engines that matter for the hotel’s source markets. Default set: ChatGPT, Perplexity, Google AI Overviews, Gemini. Mandarin-language engines are added for groups with material China inbound; regional engines are added per market.
Can a single property run this, or does it need a group?
A single property can run it. The economics are stronger at group level because the editorial programme and own-domain authority work compound across properties, but the playbook is the same.
Final CTA block
Recover bookings from OTAs by winning AI citation share.
Score your direct booking exposure
Read the methodology