› TomHow / Services / AI search

Search visibility across Google and AI answers.

This is the umbrella service. TomHow starts with normal search demand and crawl evidence, then checks where AI answers name, omit, or cite the brand. The work joins SEO, page evidence, technical fixes and measured answer checks instead of selling each acronym as a separate product.

Last reviewed 2026-09-06

Understand what is being measured.

TomHow reviews search demand, useful page content and selected AI answers as parts of the same customer journey. The measured panel can include official Claude, ChatGPT, Perplexity and Gemini APIs, with the exact platforms and methods agreed in the scope. Google AI Overviews and Microsoft Copilot are not included in that panel. Each report distinguishes observations from interpretation and keeps qualified enquiries in view as the business goal.

Surface
Operator
What we know
Claude
Anthropic
For an agreed buyer question, record the answer, any named businesses and any cited pages. Keep the model and search mode alongside the observation.
ChatGPT Search
OpenAI
Check whether an answer mentions the business and whether it links to a useful page. A mention, a citation and a visit are different outcomes.
Perplexity
Perplexity AI
Record what each sampled answer actually says. Distinguish a cited source from a business recommended to the buyer.
Gemini
Google
Keep the question, model, search mode and date with each result. Compare equivalent checks rather than treating different answers as a single ranking.
Google AI Overviews
Google
Outside the current measured panel. Google search visibility is reviewed separately; the report does not claim a measured AI Overview share.
Microsoft Copilot
Microsoft
Outside the current measured panel. Copilot tracking is not included in the promised deliverables.

Keep the customer journey in view.

Customers can use search results and AI answers while choosing a provider. The useful work connects that discovery to a clear service and an enquiry, with the limits of each measurement made explicit.

Search
A buyer finds a search result and may visit your website. Relevant visits and qualified enquiries show whether that attention is useful.
Answers
A buyer may also use an AI answer while choosing a provider. Mentions and citations help explain visibility; they are not proof of an enquiry or sale.
Evidence
Record searches, page visits and confirmed enquiries. Keep the date, source and measurement limits visible.
Evidence
Add repeatable buyer-question checks where they are useful. Connect answer observations to visits and enquiries only when the evidence supports it.
Work
Make your service clear, your pages useful and your contact route easy to use.
Work
Use SEO, GEO and AEO together where they support that customer journey. Prioritise the next useful change, not the number of acronyms.

Agree the scope. Measure what matters.

We agree the buyer questions, platforms and checking schedule before the work starts. A dated observation records what a particular system returned for that question. It does not describe every answer or predict future recommendations.

The plan brings useful search and AI-answer findings together with the website and contact path. Success is judged against the agreed business goal. Citations can inform the work; qualified enquiries and sales need their own evidence.

> Free tool

Create your first llms.txt file.

Use the free builder to summarise your business, services and useful public pages. llms.txt is optional guidance, not an access control or a proven ranking signal. The builder runs in your browser.

Open the builder

AI search optimisation, answered plainly.

What is AI search optimisation?

AI search optimisation reviews how a business is represented in selected AI answers and improves the public information those answers may use. It combines useful content, SEO, appropriate structured data and dated answer checks. The aim is to help suitable customers understand the service and enquire, without promising a mention or citation.

Is AI search optimisation different from SEO?

The work overlaps. SEO reviews how customers find and use pages through search. AI search work adds checks of selected generated answers and their sources. Rankings, clicks, mentions and citations are supporting evidence; qualified enquiries and commercial outcomes remain the reason to do the work.

Which AI search tools should B2B brands track?

Start with the systems your customers use. TomHow offers sampled checks through official Claude, ChatGPT, Perplexity and Gemini APIs, with platforms, models and limits agreed in the scope. Google AI Overviews and Copilot are outside the current measured panel. API observations do not reproduce every answer a customer may see.

How do you start marketing a site for AI search?

Start with the customers you want to reach and the questions they ask. Check relevant search and answer evidence, review whether your pages explain the service clearly, then prioritise useful changes and a working contact path. Recheck the same questions and confirmed enquiries after the changes.

What is LLM citation tracking?

LLM citation tracking records the businesses and pages named in sampled AI answers to agreed buyer questions. Each observation keeps its date, platform, model and limits. The scope sets the number of questions and checking frequency. A citation is supporting visibility evidence, not a confirmed lead or proof that the system recommends your business.

What should I do if organic traffic has dropped to zero?

Start by proving whether the drop is real. Check GA4, Search Console, consent, tracking scripts, indexation, sitemap, robots, and the pages that should be earning demand. Then map the buyer prompts and rebuild the pages that AI systems and search engines can use.

How is GEO different from SEO?

GEO focuses on how a business and its pages appear in generated answers. SEO focuses on search visibility and the pages people reach. Clear content, accurate business facts and a usable website support both kinds of work. TomHow agrees one scope around the buyer journey and reports what was actually observed.