Methodology

How it works, and where the limits are.

Hotels act on what a report says, so here is exactly how it is made, what it relies on, and what it cannot tell you.

How a report is made

From a web address to a plan.

No integrations, no tags on your site, no access to your systems.

1. Tell us your hotel

Give us your hotel's website. Its name and town sharpen the outside checks.

2. We read what AI reads

Your pages, open map and data sources, public search results, and an assistant asked real traveller questions.

3. You get a plan

A readiness score shown with how much it rests on, the few actions worth doing first, and PDFs to forward to your team.

4. We re-test monthly

The same questions, word for word, so you can see what changed, and only when the change is bigger than the noise.

AI assistants

How we sample AI answers.

The sampled answers come from one AI assistant with live web search. Each question is asked more than once, because the answers vary, and the report shows a plausible range rather than a single number.

Results are always labelled as one assistant on one day. They are a sample, never a market ranking. ChatGPT, Gemini, Copilot, Perplexity and others are not tested by Presence Pro yet; the report gives you a short checklist to test them by hand.

Data sources

Public sources, and your own export.

Your public website, open sources such as OpenStreetMap and Wikidata, public search results (to identify your listings and the pages the assistant cited), and the AI assistant's answers.

We do not read Booking.com or Tripadvisor pages directly, because their terms prohibit automated collection. You can upload your own review export (CSV or Excel from any platform) for a full review and reply analysis. Review text is analysed for that run only and is not saved.

What is included

Available today, and still to come.

We separate what exists now from what is planned. Planned items have no dates; this page changes when they go live.

Available today

  • Monthly management report (PDF) and an evidence appendix, by email
  • A sample of what AI says about your area, and who it names
  • Website readiness checks, quoting the exact words and page
  • Listings and wider-web checks across 30 points
  • Competitors named, and the pages the assistant cited
  • Review and reply analysis from your own export
  • Change since your last report
  • A free one-page snapshot, prepared by us and emailed

Still to come

  • Online dashboards for each hotel and across a portfolio: being rolled out to customers
  • Re-checks more often than monthly
  • Other AI assistants tested automatically (today: a checklist to run by hand)

Never

  • Guaranteed rankings or "become number one on ChatGPT"
  • Fake or incentivised reviews, or paid placement passed off as editorial
  • Posting anywhere on your behalf

Our principles

A thin score dressed up as certainty is worse than no score.

Hotels will act on this, so Presence Pro is built to be right, and to say when it can't be sure.

The score shows what it rests on

"72, across 60% of the full model." Anything we couldn't measure is marked not assessed, never scored as zero and never quietly left out.

Not found is not the same as missing

If we read some of your pages and found nothing, we say not found on the pages checked. A page we couldn't open is never counted against you.

AI answers are a labelled sample

Asked more than once, on the day. We never turn it into a market ranking.

We respect other sites' rules

Booking and review platforms prohibit automated reading, so we don't do it.

No promises we can't keep

No guaranteed rankings. Advice never involves fake reviews, paid placement as editorial, or posting as a guest.

Every recommendation shows its evidence

What we found, the page or quotation behind it, why it matters and what to do. A recommendation with no evidence is withheld.

Limits

What it cannot tell you.

  • Why an assistant chose a hotel. We show what it cited and what your pages leave unanswered. Reasons we suggest are hypotheses to check, not proven causes.
  • That one action caused a change. What an assistant says shifts for many reasons.
  • How you rank across the market. A sample of answers on one day isn't a ranking.
  • What a booking site's pages say behind its block. We identify listings from public search results only.

See it on a real hotel

A free one-page snapshot shows the method on your own hotel, with its limits stated on the page.

Get your free snapshot