Skip to main content

Rankings get you retrieved. Sources get you cited.

When a buyer asks ChatGPT who to hire, the answer is written from a handful of pages the engine trusts: yours, a competitor’s, a directory, a listicle. We find out which pages it used, then build and fix the sources so the next answer names you and links you.

Your buyer questions, run through three AI engines, with every source each one used.

We run the test on ourselves first.

  • 4 of 11

    buyer questions where ChatGPT cited an Emerald page, in our November 2026 panel

  • 33

    AI answers logged source by source, across ChatGPT, Perplexity and Google AI Mode

  • #2

    nationally for "motorcycle accident lawyer" in six months, for The Law Center

How we audit and build AI search visibility.

Organic rankings decide which pages an engine reads. What those pages say, and whether other sources agree, decides who gets named and linked. We work both halves and measure them separately.

  1. Run the buyer panel

    We write the questions your buyers type, in their words, and run each one through ChatGPT, Perplexity and Google AI Mode. Every answer gets one of three marks: cited (your URL is a source), named (your brand appears with no link) or absent.

  2. Log every source

    The answer is the output. The sources are the lever. We log every URL each engine cited or pulled, and sort it: publisher, competitor, directory, listicle or forum. That list becomes the work plan.

  3. Fix the entity record

    Engines merge what your site, your Business Profile, LinkedIn, directories and old press say about you. When those disagree, the model picks one. We trace each wrong fact to the sentence it came from and correct it at the source.

  4. Get onto the lists engines read

    For "best [service]" questions the engines lean on roundups and directory rankings. We find the ones each engine cites for your query, check which are open to new entries, and make your entry easy to lift: a numbered row, a named result, a current address.

Why ranking in the blue links is not enough anymore.

Each of these came out of our own November panel. The same patterns show up in client audits, because they are how the engines build an answer.

  • The engine read your page and still left you out.

    Make your entry extractable

    ChatGPT fetched our law firm SEO comparison, then built its answer table from the numbered entries on other lists. Our entry sat in an unnumbered section under the eight agencies we reviewed, so it never made the table. We made it a numbered entry. Structure decides what gets copied.

  • A directory answered the question for you.

    Own the third-party record

    When Perplexity recommended us for New Orleans, its stated reason was our score on an Agency List ranking, not anything on our own site. Rankings, roundups and directories carry the "best of" questions. We find the ones your engines cite and get you onto them by hand.

  • The model has your facts wrong.

    Trace it to the sentence

    Google AI Mode listed a Miami office for us, an office we had already closed. We traced it to two of our own posts from 2021 and 2022 that still said it. Wrong answers usually quote something real and out of date. Fix the source, not the answer.

  • Agencies pitch AI ranking guarantees.

    Auditable source work

    Nobody controls model weights or guarantees a ChatGPT slot. You can control retrieval rank, entity facts, on-page proof and the third-party pages the engines read. That is the work, and we rerun the same panel on a fixed schedule so you see which answers moved.

Win the retrieval, then own the record the answer is built from.

AI search runs in two stages. First the engine retrieves pages, mostly from a search index. Then it writes an answer from what those pages agree on. We win the first stage with the agentic SEO system we build and open-source, and the second with entity facts, extractable tables and third-party agreement. The two stages do not move in lockstep: in our November panel, ChatGPT cited two of our articles that Google had not indexed yet.

View our open-source SEO kit

From the insights

Go deeper on the search problems behind this page.

Common questions

What is the difference between AEO, GEO and SEO?

SEO gets your pages indexed and ranked, which decides whether an AI engine retrieves them. Answer Engine Optimization (AEO) and Generative Engine Optimization (GEO) cover the second stage: making your brand, pricing, proof and third-party footprint clear and consistent enough that ChatGPT, Perplexity and Google AI Overviews name you and link you inside the answer.

What is the difference between a mention and a citation in AI search?

A mention means the engine writes your company name in the answer. A citation means it links your URL as a source the buyer can click. If you are named while a directory or competitor holds the link, they get the visit. We track the two separately for every question in your panel.

How do you measure AI search visibility?

With a fixed panel of buyer questions, run through ChatGPT, Perplexity and Google AI Mode. Each answer is marked cited, named or absent, and every source URL the engine used is logged and classified. We rerun the same questions on a schedule and compare cell by cell, so a change in the answer can be traced to a change in the sources.

Do I need to rank on Google before AI engines will cite me?

Ranking helps, because most engines retrieve from a search index first. It is not a hard gate. In our November 2026 panel, ChatGPT cited two of our articles that Google had not indexed yet. A well-built source page can earn citations while it waits on Google.

Which AI engines do you audit?

ChatGPT with search, Perplexity, and Google AI Mode and AI Overviews. Each leans on a different mix of search indices, roundups, directories and site pages, so we run your buyer questions across all three before we prioritize fixes.

Can anyone guarantee a top spot inside ChatGPT?

No, and any agency promising guaranteed AI placement is selling you something it does not control. Answers vary by phrasing, engine and user. What we deliver is a measured baseline across your buyer questions, stronger retrieval rankings, sharper source pages and a broader third-party record.

What does it cost?

AI search work runs inside our SEO programs, not as a separate add-on. Our programs start at $3,500 a month for focused, scoped jobs, and our average starting engagement runs $5,000 to $10,000 a month.

See where AI cites you, names you and skips you.

Get a free AI visibility teardown. We run your buyer questions through ChatGPT, Perplexity and Google AI Mode, log every source each engine used, and show you the first fixes we would make.