Home/AI search/AI search optimization agency (GEO) for B2B

Get named when buyers ask AI

AI search optimization agency services from Tin include asking ChatGPT, Perplexity and Gemini your buyers' questions and fixing the gaps we find. After a change, we repeat the questions and compare the answers. Our people do this work for less than the cost of a junior marketing hire.

Book a call

Missing from AI answers costs you shortlist places

When a buyer asks an assistant which companies to consider, the names in its answer can become the shortlist.

  • Competitors take the mention. One software startup appeared in 1 of 90 answers; its main competitor appeared in 43.
  • A read page gets rejected. ChatGPT found one startup's site during a search, read it and still left the company out.
  • Checklists miss the obstacle. One client had completed an AI-ready checklist while Google had indexed only two of its pages.

We trace where the company disappears from the answer and choose the work that addresses that gap.

Trackers, SEO agencies and writers each cover part of AI search optimization

A mention count tells you whether the company appeared. Read the table for who investigates a missing name and changes the information behind the answer.

OptionWhat it is good forWhat it leaves the founder to do
AI visibility trackerRecording how often assistants mention companies and how those counts changeInvestigate the missing mentions and arrange for someone to make changes
SEO agencyWorking on pages and technical issues that affect search engine rankingsAdd buyer-question tests where assistant answers fall outside the SEO scope
In-house page writingWriting about the product with knowledge from the people building itDecide which questions to answer and where the company is being lost
AI search with TinAsking the questions, making the fixes and checking the answers againApprove the page changes after our people have prepared and reviewed them

We test the answers, fix your pages and add outside sources

We examine what assistants say and the information behind it, then work on your site and the places they can find your company elsewhere.

Find the missing answers

We ask questions without your company name and examine the sources ChatGPT, Perplexity and Gemini cite.

Make your pages usable

We write pages that answer buyer questions and check indexing, schema, llms.txt and markdown copies for readable information.

Develop sources beyond your site

We work on directory listings, backlinks, Reddit threads and LinkedIn content that can inform answers about your company.

We ask the same buyer questions again after every change

We follow each answer from the assistant's first suggestions to its final choice, so the next edit addresses the point where your company drops out.

Before testing

  1. During the check

    Write the buyer questions

    We phrase questions around the problem and the purchase without naming your company, to see whether the assistant finds it unaided.

  2. Before release

    Trace each missed mention

    We follow memory, search queries, results, page reading, shortlisting and selection across three assistants to locate where the company disappears.

  3. After the change

    Make the relevant fix

    We change the page, copy, schema or indexing issue identified by the check, with our people reviewing the work before it goes live.

  4. Compare the new answers

    We repeat the questions, read the differences and use what changed to decide which page or source needs attention next.

We report findings and finished fixes apart from results

You can see what we learned, what we changed and whether the answers moved. Each tells you something different about the progress of the work.

  • Separate the evidence. Reports distinguish a finding that guides the next action, a completed change and a measured result.
  • Understand the starting position. The 1-of-90 check recorded the starting answers and helped choose the next work; it was not an improvement figure.
  • Address the actual cause. When an assistant reads a page but passes over it, we examine whether its claims give the assistant a reason to recommend it.

It fits B2B companies whose buyers research with AI

If buyers ask an assistant about suppliers before they contact you, we can examine what they are hearing. Existing search work gives us pages to build on.

Reach software buyers researching options

We test the questions people use to compare tools in your software category and examine where your product appears.

Extend existing SEO work

We check what assistants do with the information on your search pages and include the needed changes in that work.

Explain the service behind AI

We explain the technology and the team behind an AI-enabled service so assistants have specific information about both.

Talk to Tin

Let's discuss the questions buyers ask before purchasing and the competitors they compare you with. We can explore the AI search work with no long-term commitment.

Book a call

FAQ

What is generative engine optimization?

Generative engine optimization, or GEO, is work that helps a company appear in AI assistants' answers. AI search optimization and answer engine optimization, or AEO, are names used for the same field. The work concerns the information assistants can find, understand and use when answering a potential buyer.

Do we need good SEO before AI search work?

The work overlaps, and we can address both in the same engagement. Assistants search the web, so indexing and readable pages matter when they look for sources. We check those foundations and the answers together to choose the relevant page and site changes. See how GEO and SEO relate.

Can you guarantee an assistant will recommend us?

No. We can ask the questions and work on the information behind the answers, but we cannot promise what an assistant will recommend. We report the findings, the fixes we completed and the answers from a later check separately, so you can judge what changed.

Do we need an llms.txt file?

We include it in the checks, together with the markdown copies of pages. Tin uses both on its own site to make the information available for assistants to read. What those pages say still matters: we look for specific material an answer could use, alongside the files that present it.

What makes a page worth citing for an assistant?

Specific information gives an assistant something it can use: a clear claim, a relevant number or evidence about the customers a product serves. We found one company's markdown copy contained no customer names, logos, testimonials or numbers. We check what the assistant can read, as well as what the page shows a visitor.

Talk about where AI answers leave you out

Join the founders and engineers behind the experiments to discuss how assistants describe your category. You will leave knowing which questions, pages or sources we would check first to understand where your company is missing.

Book a call