The AI Bookability Check
The public scoring framework for one question the whole market is about to care about: can a local business be found, understood and booked, by customers, by search, and by the AI assistants people now ask first. Here is exactly how it works.
A real report, scored
One number out of 100, four dimensions underneath it. This is what a typical local business looks like today: findable, but nearly invisible to AI.
Overall presence
out of 100
Online Presence
Customer Actions
AI Readiness
Agent-Ready Booking
Why this matters
Discovery already moved to AI. Booking is next
A year ago, 6% of consumers used AI to find a local business. This year it’s 45%.AI is now the third way people find local places, past Yelp and TripAdvisor. More and more, people just ask an assistant “find me a good place tonight” and go with what it hands back.
A year ago
This year
Share of consumers who used AI to find a local business, 2025 vs 2026. Source: BrightLocal Local Consumer Review Survey 2026.
83%
of restaurant locations are entirely invisible in AI-generated recommendations
Uberall QSR benchmark, May 202684%
of hotels don't appear in AI-generated recommendations at all
HotelWorld AI Visibility Index, Q1 202689%
of Americans are open to an AI agent handling a restaurant booking
PureSpectrum, 2025
They’re invisible not because they block AI, but because their sites can’t be read by it. And booking is becoming action too: Google switched on AI restaurant booking in 2026. The businesses that get ready first get chosen. The rest get quietly skipped.
Being early is the whole advantage. When an independent place is visible, most AI recommendations go to it, direct and commission-free. The score is how you get there before it’s a problem.
The four dimensions
Each is scored on its own, then combined. What each one checks, why it decides whether you get chosen, and where a typical business lands today.
Online Presence
What it checks. Whether a real person, and Google, can find you at all: a reachable website, a complete Google/Maps profile, and social profiles that actually resolve.
Why it matters. If you can't be found, nothing downstream matters. This is the floor everything else stands on.
Customer Actions
What it checks. Whether a visitor can act without hitting a dead end: a clear way to book, call, open the menu or get directions, on a phone, in seconds.
Why it matters. Most local business is lost between interest and action. Friction here is revenue walking away.
AI Readiness
What it checks. Whether your business facts are structured so a machine can read them: clear, consistent hours, location, services and offer, in a form assistants and search AI can parse.
Why it matters. AI answers are now the first screen. An assistant recommends what it can read and verify, and quietly skips the rest.
Agent-Ready Booking
What it checks. Whether an AI assistant, acting for a customer, could actually complete a booking on your site: a findable booking path, readable date and time fields, and no wall that stops an agent cold.
Why it matters. People are starting to let AI book for them. This is the one score almost nobody else measures, and the one that will decide who gets chosen.
The diagnostic, in numbers
scores in your report — presence, actions, AI readiness, agent-ready booking
seconds to your first score. Six fields, nothing else to fill in
for the full report — free, with a concrete action plan by email
What we deliberately don't do
A score is only worth something if you can trust it. So the rules are strict, on purpose.
We never invent data.
If something can't be verified from what's publicly there, we mark it unknown. We do not guess a number to fill a box.
We don't scrape private or paid data.
The score reads what is openly published. No back doors, no data a business didn't choose to make public.
We publish what and why, not the exact how.
The four dimensions and their reasoning are open, so you can trust and check them. The precise weighting and the live agent test are our engine. What matters to you is the result, and the fix.
We never sell you a fix you don't need.
The diagnostic is free and stays free. If the score says you're fine, we say so. We only quote work the number actually calls for.
How to read it
Your number, in plain terms
The only way to know your number is to run it.
Sources
- BrightLocal, Local Consumer Review Survey 2026. Reported usage of generative AI for local recommendations rising from 6% in its previous survey to 45% in 2026.
- Uberall, “Fast Food, Faster Discovery”, 2026 QSR benchmark. 83% of restaurant locations in the benchmark did not appear in AI-generated recommendations; 17% did.
- HotelWorld AI, Q1 2026 Visibility Index (reported by Hospitality Net). Only 16% of hotels appear in AI-generated recommendations at all.
- PureSpectrum for Popmenu, 2025. Consumer openness to AI agents handling restaurant tasks.
- Structured data makes business facts easier for machines to parse and verify, which is why the AI Readiness dimension weighs it. We do not publish a single multiplier here: the figures circulating for that effect lack a traceable primary study.
- G-Lab aggregate diagnostic data. Per-dimension distributions across businesses scanned, updated continuously.
We publish only figures we can stand behind, with the source named and dated. Sample report values are illustrative of a typical result and are not a specific business.
Methodology v1.0
Published 21 Jul 2026
This is a public scoring framework, so it is versioned like one. v1.0 defines the four dimensions above, what each one checks, and how a result is banded. When the method changes, the version number changes and the change is recorded here. Scores produced under an earlier version stay labelled with that version.
- v1.0 · 21 Jul 2026
- First public version. Four dimensions (Online Presence, Customer Actions, AI Readiness, Agent-Ready Booking), static-analysis based. Weights and the live-agent test remain internal.
This method is applied market by market in G-Lab Research, starting with 200 Amsterdam restaurants. We study the gap between being online and being genuinely bookable by AI. If you are a researcher, journalist or agency and want the detail behind a number, write to hello@g-lab.studio.
Data handling in research
- What we look at. Publicly available business information only: the website, the public Google business profile and the public booking path. We do not collect data about individual people, and we do not access anything behind a login.
- We never submit a booking. The agent walks a public booking path to the confirmation step and stops. No business receives a reservation it did not agree to.
- Public means aggregate. Studies publish distributions and market-level findings. We do not publish league tables of named businesses with low scores.
- Your own result is yours. A business can request its full individual result privately at any time, and can ask to be excluded from a study or from any future contact. One email is enough and it takes effect immediately.
- Corrections. If a result is wrong because the site changed or our check misread it, tell us and we re-run it. Material corrections are recorded in the changelog above.
Requests and corrections: hello@g-lab.studio.
Free Online Presence Diagnostic
Not sure where to start? Get your score first
A 60-second check of your website, Google presence and socials — then a concrete plan by email.