Rising ▲
Gemini briefly dropped citations at a model launch.
For you: every model release can move your visibility overnight, so we re-check your questions straight away.
Source: sAIph knowledge base, Oct 2026 ↗AT A GLANCE
The sAIph Search AI Search Audit tests more than 100 guest questions across five AI engines, benchmarks your hotel against the seven or eight competitors guests compare you with, reads your website the way AI crawlers do and maps the sources AI relies on.
You receive an executive summary, an action plan and a fact sheet for your team, with the full report behind them. It is £4,600, delivered in four to eight weeks, and yours to keep whether or not you work with us.
100+guest questions, tested live
5AI engines, every answer read by a person
£4,600until 31 December 2026
Trusted by world-renowned five-star hotels including The Langham, London and Raffles London at The OWO.
WHAT AUDITS USUALLY FIND
Why hotels that rank on Google still lose the AI answer.
Awards AI cannot attach
A hotel held more than 60 accolades. Its website named one as an award. The engines knew them when asked directly, but could not use them to choose "the best hotel in London".
A product that does not exist online
The spa sold memberships, but "spa membership" appeared nowhere on the site. Competitors took up to 65% of those answers.
No numbers on the pages
Capacities, sizes and prices sat in fact sheets and PDFs. The engines named competitors that published a room and a seat count.
Most words never reach the engine
Pages built from layout modules kept about one word in five after AI extraction. The hotel's own magazine articles, written as prose, did far better.
Suites hidden behind the booking engine
The rooms page linked straight to booking, so 19 suite pages had nothing for AI to read. A group template rule, written up as a ready-to-send request.
What was already right
Every AI crawler could reach every page. We said so, and told the team how to keep it that way.
From real 2026 luxury hotel audits, anonymised.
WHERE AI SENDS THE BOOKING
We check who gets the click, not only who gets named.
Every audit includes the direct-booking layer: what AI quotes, where it links and what stops the booking coming to you.
Which domain gets the link
For every question, whether the answer links to your website, an OTA or Google's own travel results.
Which price is quoted
A rate snapshot against the OTAs, so you see where an undercut sends the guest elsewhere.
Your book-direct facts
Whether your site states a reason to book direct in a form AI can repeat, and where it is missing.
Your own rate on Google
Whether your website rate reaches Google Hotel Center and free booking links, matched to your Business Profile.
WHAT YOU RECEIVE
Six documents, written for the person who will use them
WEEK 3 TO 4
Your team, before the walkthrough
Early findings note
The headline, how you do on each commercial priority, what stands out and the questions to settle together, sent before the full report is finished.
- Results by priority, not just one score
- The questions only you can answer, such as which figures stay off the site on purpose
- Reviews of any page you have in draft, with the changes in order
CORE
General manager, owners, group
Executive summary
Everything the audit found, in short form, written for the person who approves spend.
- Your score, by engine and by commercial priority
- Share of voice against your competitor set
- Where you are strong, where you are absent, and why
- What it adds up to, in plain terms
CORE
Whoever does the work
Action plan
Every opportunity from the audit, sorted into what to do now, what needs support and what can wait, set out in phases.
- Each action: what, why (with the evidence), how, owner and effort
- Flagged where it needs us, or needs your group or brand team
- The questions to watch on the re-test, for every action
- Group requests written ready to send as they stand
CORE
Web, PR, sales, events, spa, restaurants
AI search fact sheet
A short note for the wider hotel team, with examples from your own hotel. Written to be circulated.
- What matters for AI search, and what is oversold
- A checklist for any page that carries a fact
- Why prices, capacities and awards need to be on the page as text
REFERENCE
Marketing lead and digital team
Full audit report
The full diagnosis, with the method and the evidence behind every recommendation.
- Question-by-question results across every engine
- Your site read as AI crawlers read it: access, readable text, missing facts
- Competitor benchmark and the off-site sources behind each answer
- Every exclusion and limitation, stated
REFERENCE
The person doing each task
Action plan appendix
Every task in detail: why it matters, how to do it, the phrases to use and what to watch for on the re-test.
- Step-by-step instructions per task
- What not to do, with the evidence
- Before-and-after examples from your own pages
FOLLOW-UP
Your team
Walkthrough and re-tests
We walk you through every finding and agree owners for the first month. Then we re-run the same questions on the same engines at day 30 and day 90.
- In person or on a call
- Owners agreed for month one
- Re-tests at day 30 and day 90, question by question, so the movement is measured and yours to claim
You keep everything, including the question set and the raw results. Hand the action plan to your own team or agency, or ask us to deliver it.
WHAT HAPPENS AFTER
From audit to shortlist.
We say "should", because nobody controls a generated answer. What we can do is measure before and after, on the same questions, engines and scoring.
Not sure where to start? Ask us anything.
Advice, a second opinion or a quick question. We reply within one working day.
FREE · TWO WORKING DAYS · NO CALL
Find out who AI recommends instead of you.
Ten guest questions, your competitors side by side and up to five quick wins your own team can make this month.
Two emails: your Snapshot and one check-in a week later. Nothing else unless you reply. We never sell your details. Privacy notice →