If an AI assistant does not recommend your business, the answer may be as simple as the request not matching your location or services. It may also lack reliable information about you. Treat the missing mention as a clue to investigate, not proof of a hidden penalty.
See the actual answers: Use Loop's free AI Visibility Checker to test relevant local questions. It shows which businesses were named so you can compare the information a customer receives.
1. The question is not a good fit
A city-wide request for a late-night vegan meal should not return a breakfast café across town. Test prompts that reflect what you genuinely offer. Change one factor at a time: suburb, opening time, service, budget or accessibility. If you appear for close matches but not broad ones, your business may be accurately understood but less relevant to the broader request.
2. Your core facts are missing or inconsistent
Check your website and major business profiles for the same name, address, phone, hours and booking route. Make menus, room types, services and policies easy to read as text. Google says local results use relevance, distance and prominence, and recommends complete, accurate Business Profile information. Those factors are not a disclosed formula for every AI assistant, but they are a sound place to start.
3. Your website does not answer the customer's question
“Great food and a warm welcome” does not explain whether a restaurant handles gluten-free diners, takes bookings after 9pm or has parking. Add pages and FAQs based on real customer questions, with current details and clear contact or reservation links. Google's guidance for AI Search emphasizes unique, useful content and ordinary SEO foundations rather than special AI files.
4. The information is hard to access
A search assistant may not use facts hidden behind a login, blocked from crawling, or only shown in an image. Check that important pages load, expose readable text and are linked from your navigation. For ChatGPT search, OpenAI documents the role of OAI-SearchBot. For Google AI Search, pages must meet normal indexing and snippet requirements. Neither rule promises inclusion.
5. Customers cannot complete the next step
Even when AI names you, the experience fails if a customer cannot see current availability or book. Test your reservation, order or enquiry path on a phone. Give buttons and forms clear labels. The AI-ready website checklist covers that journey, and Loop's free website assistant can answer from approved details and connect to booking options.
What to fix first
- Correct factual errors on your site and business profiles.
- Write one useful page or section for a high-value customer question.
- Make the booking or enquiry path work from start to finish.
- Repeat the same tests after the information has had time to be discovered.
Do not buy fake mentions or write reviews for customers. A genuine, current business presence is more useful than a manufactured signal. See how to run a repeatable check and how to measure change. The full guide connects this work to a more usable website.
Use a diagnosis matrix before making changes
| What you observe | Likely area to inspect | First check |
|---|---|---|
| Known by name, absent from discovery prompts | Relevance to the specific need | Does your site clearly describe that service or use case? |
| Named with wrong hours or services | Conflicting or stale sources | Compare your site, profiles and booking pages |
| Recommended but no booking path | On-site conversion | Complete a reservation on a phone |
| Absent across all reasonable prompts | Basic discoverability and crawlability | Check whether key pages and profiles are public and current |
This table is a way to choose the next investigation. It is not a diagnostic result from an AI provider. Run several prompts before deciding which pattern you have. The testing guide shows how to do that consistently.
Compare your business with the actual customer request
It is tempting to search for “best restaurant” and expect your venue to appear. But a useful assistant is trying to answer a person's full question. “Quiet dinner near the theatre at 7pm” combines atmosphere, location, time and possibly availability. A venue across town or closed on the requested day may be a poor match even if it has excellent reviews.
Write down five questions your business is truly equipped to answer. For each, identify the page or profile that proves it. If you offer private dining, publish group size, room type, meal options and an enquiry route. If you have accessible rooms, explain the actual access features rather than relying on a generic “accessible” badge. Specific, verifiable details help customers decide and make the business easier to describe accurately.
Audit the information chain
Most local businesses have several versions of their facts online. A restaurant might have hours on its own site, a menu on a delivery platform, availability on a reservation service and an old holiday notice on a directory. When those conflict, a customer cannot know which to trust. Use your own current website as the source of truth and update the major profiles you control.
Build a simple record with the page URL, owner, last checked date and fields that matter. Give one person responsibility for seasonal and holiday changes. When a booking provider changes, update the website link and test it. Google says complete and accurate Business Profile information supports local relevance. Its AI Search guidance also recommends keeping profile information current.
Check whether important pages can be found
Ensure the menu, services, location and booking pages are linked from the site navigation or relevant service pages. Put essential facts in text. A beautiful image of a menu is less useful than an accessible menu page with dish names and prices. A PDF can be a download, but it should not be the only place a customer can find critical details.
For Google AI Search, a page must meet normal indexing and snippet requirements; Google says there is no separate AI-specific markup requirement. For ChatGPT search, OpenAI's OAI-SearchBot documentation explains the crawler used for search discovery. If your site blocks a crawler or hides key pages behind login, ask your site provider to inspect it. Technical access is necessary for some search experiences, but it does not guarantee inclusion.
Use genuine evidence, not manufactured signals
Customers often describe a business more specifically than its own marketing copy. A review might mention a calm setting, a helpful host or a well-handled dietary need. Read these comments for service insight and respond when a concern can be resolved. Do not ask customers to insert search phrases or offer a reward for a Google review. Google's policy prohibits incentivized and biased reviews.
Local partnerships can also clarify what a business is for. A hotel listed by the convention bureau, a restaurant in a theatre's visitor guide or a salon included in a neighborhood event page may help people find the service in context. Ask for accurate descriptions and links where there is a real relationship. Buying artificial mentions is unlikely to create a dependable customer experience.
Prioritize fixes by customer impact
Fix incorrect facts first, especially hours, location, price, accessibility and booking links. Next, make the most common high-intent questions answerable on your site. Then improve your booking path. Only after these basics work should you spend time on niche formatting or speculative AI tactics. Google's official guidance explicitly says special AI text files and “GEO hacks” are not required for Google Search.
Do not judge a fix by an immediate single re-test. Record when you made the change and repeat the same question set after systems have had time to discover it. Compare accuracy and actionability as well as mentions. The measurement guide explains how.
What if a competitor appears instead?
Inspect why that business suits the question before treating its inclusion as a problem. It may be closer, open at the requested time, better described for the particular service or connected to a working booking platform. Record the difference and decide whether your business actually offers a comparable experience. If it does, make that fact clear on your site. If it does not, focus on prompts that reflect your strengths. A useful AI recommendation should match the customer, not rotate every local venue equally.
Frequently asked questions
Is my business being penalized by AI?
An absent recommendation does not prove a penalty. The request may not fit your business, the assistant may use different sources, or important facts may be missing. Diagnose with several realistic prompts and check your public information before drawing a conclusion.
Do I need an llms.txt file?
Google says no special AI text file is required for its generative Search features and that it does not use llms.txt for ranking. Another system may have its own requirements, but a file should not displace useful pages, accurate listings or a working booking route.
Will more reviews guarantee inclusion?
No. Reviews can be useful evidence for customers and can contribute to local prominence in Google, but a recommendation depends on the question and many other factors. Seek honest reviews from real customers without incentives or requested wording.
What if my booking provider has the wrong details?
Correct the information with that provider and test the resulting customer path. If a reservation is unavailable, do not imply it is. Keep your own site clear about the correct booking route and contact method.
Next step: Check your current AI visibility.