Module · Department: Marketing · €34/mo

AI Performance — are you mentioned when the guest asks ChatGPT?

More and more guests ask ChatGPT instead of googling. The module asks the questions your guests ask — every week, several times — and shows whether the hotel is mentioned, who is named instead, and which sources carry the answers. Plus the traffic from AI assistants in your GA4.

30-day trial · No commitment · Measured in ChatGPT every week · the hotel's own town as location · the sources ChatGPT itself lists
Try IO Planner on demo data before you decide Enter your name, hotel and e-mail. We approve your access and send you a link to the module on our demo hotel SiloStay — a short guide takes you through it, exactly as when we present it.
Try it LIVE →
Do you know the answers?

3 questions most hotels cannot answer

Nobody can see what ChatGPT says about you by looking in Google Analytics. The only thing that works is asking the questions — again and again — and counting the answers.

1

Are we mentioned when the guest asks an AI instead of googling?

2

Who is mentioned instead of us — and how often?

3

Which sources carry the answers — and are we there?

Part 1 of 2 · The answers

Ten questions, asked the way the guest would ask them

You write the prompts yourselves — a conference for 60 in North Jutland, a weekend for two by the water, a big birthday with an overnight stay. Each prompt is asked five times a week, and the figure is the share of answers in which the hotel appears. Any answer can be read in full, with its sources alongside.

AI Performance SiloStay · demo data
  • ChatGPT Conference hotel in North Jutland for about 60 people with accommodation? Mentioned 4/5 · rank 1
  • ChatGPT Weekend stay for two in North Jutland with good food and water nearby? Mentioned 3/5 · rank 2
  • ChatGPT A special venue for a wedding with party and accommodation in North Jutland? Mentioned 2/5 · rank 3
  • ChatGPT Hotel with spa in North Jutland for a long weekend? Not mentioned 0/5
Illustration with demo data (SiloStay). The real module shows your own figures.
  • ✓Mention rate — the share of answers in which the hotel is mentioned — per prompt, per category and in total, week by week
  • ✓Who is named instead — the hotels ChatGPT recommends when it is not you — and how often
  • ✓The sources behind the answer — the pages ChatGPT itself cites: is your own website among them?
  • ✓The whole answer — one typical answer per prompt, as the guest would get it — with the places marked
  • ✓Actions from measurement 4 — absence and decline trigger a line with what needs doing
Part 2 of 2 · The AI traffic

How many actually come in from the AI assistants?

GA4 has its own channel for visits from ChatGPT, Gemini and the others. The module shows it week by week, per source, against the weeks before — and whether the visits turn into purchases. If you have a ChatGPT Ads account, the spend sits alongside.

  • ✓Visits per source per week — chatgpt.com, gemini.google.com and the others — from your own GA4
  • ✓Against the weeks before — the same number of weeks before the period, so a rise can be seen for what it is
  • ✓Purchases from AI visits — when GA4 measures purchases on the website, they are shown here
  • ✓ChatGPT Ads — spend, clicks and conversions per campaign — if the account is connected
AI traffic SiloStay · demo data
AI visits (12 weeks) 412 ▲ 38.0 %
Share of all visits 3.3 % ▲ 0.9 pp
From chatgpt.com 96 % chatgpt.com
Purchases from AI visits 7 €2,869
Illustration with demo data (SiloStay). The real module shows your own figures.
How it works

From data to decision

No manual entry and no spreadsheets — the module builds on the data the hotel already has.

1

Write your prompts

The ten questions the guest would ask — in the categories that matter to you. Name, town and website come from master data.

2

The measurement runs by itself

Every Tuesday at 06:00 — or every morning if you choose daily measurement. Nothing to remember.

3

Read the answers and act

From measurement 4 the module points to where you are missing — and which page needs writing.

Try it LIVE

Try IO Planner on demo data

Enter your name, hotel and e-mail. We approve your access and send the link and details to your e-mail — then you can click through AI Performance yourself on our demo hotel SiloStay, with a short guide to take you along.

Try AI Performance on SiloStay

No credit card. No installation.
✓How it works: we approve your request and send the access link and details to your e-mail.
✓

Thank you — we have received your request

We approve your access and send the link and details to your e-mail. You will hear from us as soon as possible.

We only use the information to follow up on your interest — see our privacy policy.
Price

Get started today

Module-based pricing — pay only for what you use, and add more modules whenever you like.

AI Performance
€34 /mo
  • ✓ 30-day trial — we only charge after that
  • ✓ No commitment — cancel any time
  • ✓ Standard: 10 prompts × 5 runs, once a week
  • ✓ Add-ons at the same price per call: €0.68 × prompts × runs per month — weekly, or × 7 for daily measurement
  • ✓ Mention rate, who is named instead, the sources and the whole answer
  • ✓ AI traffic from GA4 and ChatGPT Ads spend alongside
  • ✓ Actions from measurement 4
  • ✓ Up to 50 prompts
  • ✓ Up to 10 runs per prompt
Start trial →
Where the answer is

The answers show where you stand — the modules alongside show what it turns into

AI Performance measures what ChatGPT says. Who then comes to the website, and what the ads in ChatGPT cost, live in the modules that read GA4 and the ad accounts.

Questions and answers

What hotels ask us

Do we need to set anything up ourselves before the measurement starts?
Only the prompts — the questions your guests would ask ChatGPT (10 as standard, up to 50). Name, website and town come from the hotel's master data. The first measurement runs the next Tuesday at 06:00 — or the next morning if you have chosen daily measurement.
Why do the answers change from week to week?
ChatGPT does not answer the same way every time — not even to the same question on the same day. That is why each prompt is asked several times (5 as standard, 10 as an add-on), and the figures are the share of those answers. It is the development over several weeks that counts, not a single run.
Do you measure what a logged-in user sees in ChatGPT?
No. We measure ChatGPT through the API (the same model as the app) with web search and the hotel's town as location. It is the model's answer to an anonymous user — not what a logged-in user sees. The dashboard says so: "API answers, not the app's".
When can we see a trend?
From the second measurement the figures are compared with the weeks before. Actions are shown from measurement 4 — there has to be history to measure a fall against.