The eight stages of building AI search visibility.
We score where the assistants place you today, then clear what stops a model trusting you: crawl access, entity resolution, schema, the coverage your category expects. Eight stages. Every one names what it produces and which side of the table owns the work.
every stage ships something you can hold us to.
What each stage produces.
Benchmark
We agree the buyer prompts and the competitors, run them across the assistants, and hand back a scored baseline. Those are the numbers you judge us on later.
Technical foundation
Crawl access, rendering, indexation, page speed. AI crawlers hit the same walls Googlebot does. You get a prioritized defect list. We sit with your developers while they clear it.
Entity foundation
Your brand has to resolve as one thing across schema, Wikidata, your Google Business Profile, and the platforms models read. You get an entity map and the markup deployed against it.
First-party authority
Topic Ownership Strategy: a hub-and-spoke build that covers a query space end to end. You see the content map before anyone writes a word.
Third-party validation
Links, citations, and mentions from sources the models already trust. We name the targets up front and report what landed.
Social utility
Reviews, community, and the signals that show real people use you. A model checks these before it puts your brand in front of a buyer.
Total Graph Authority
The point where your pages, the third parties who validate you, and your social signals describe the brand the same way. Stages two through six build toward it.
Measurement loop
The dashboard goes live at stage one and never turns off. Mention rate, citation share, recommendation rate, and share of voice against named competitors, every month.
Topic Ownership Strategy and Total Graph Authority.
Topic Ownership Strategy is how we build the first-party half. A hub page owns the subject, spokes cover the questions around it, and every page carries one job: to inspire, to educate, or to convert. A model landing on any one of them meets the same account of your brand. The full build and the three content groups sit on the Topic Ownership Strategy page.
Total Graph Authority is where the eight stages point. Your own content, the third parties who validate you, and the social signals around you all describe the brand the same way. That agreement is what holds when a model answers the same question fifty times and phrases it differently each time. We set the model out in Total Graph Authority, and the scoreboard that tracks the move toward it lives on the measurement page.
Results from AI search clients.
Where your team is on the hook.
A release path
We write the schema, the redirects, and the fixes. Someone on your side ships them. More engagements stall here than anywhere else, which is why we scope the release path in week one.
Subject-matter access
Topic ownership needs someone who knows the category. An hour a fortnight with your product or clinical lead decides whether a model cites the content or skims past it.
Someone who can decide
The monthly readout ends in a decision about next month. When the only attendee has to take it back to a committee, the loop stretches from four weeks to twelve and the trend line stops meaning anything.
The prompt set, once locked
We build the prompt set with you, then freeze it. Swapping prompts mid-engagement resets the trend line, and a trend line you cannot read tells you nothing.
Analytics plumbing
Broken GA4 and half-fired tags are a different job, and Awesonatics handles it. We will tell you when your tracking cannot carry the reporting you are asking for.
A position inside the organic answer
OpenAI has sold ads in ChatGPT since February 2026, with a self-serve Ads Manager from that May. Those run as a labeled Sponsored block under the answer, separate from it. The organic answer is still earned and no one sells a position inside it. If an agency offers you guaranteed placement there, ask for the mechanism in writing.
Common questions
We are evaluating AI search agencies. What should we look for?
Ask whether the prompt set is frozen between reports. Ask them to show you the weighting behind any composite score. Ask to see a client month where the numbers went backwards and what they did about it. Ask who is expected to ship the code, because an agency that has never thought about your release process will stall in week three.
Can we run these stages in-house?
Stages two and three, yes, if you have a technical SEO who can get schema and crawl fixes deployed. Stage four needs a content team that will hold a hub-and-spoke map for months rather than chasing individual keywords. Stages five and six are the hard ones to staff, because earning third-party validation is relationship work that does not fit neatly into a marketing calendar.
Which stage takes the longest?
Stage five, third-party validation. Technical and entity work ships in weeks and registers with crawlers quickly. Earning citations and mentions from sources a model already trusts runs on other people's timelines: editors, reviewers, and industry bodies who owe you nothing. We name the targets at the start so progress is visible while it is slow.
What happens if the numbers do not move?
We show you which stage is blocking and what it would take to clear it. Usually the answer is unglamorous: fixes written but never deployed, or content signed off but never published. When the blocker sits on your side we say so directly. When the category itself has no meaningful AI answer volume, we will tell you the budget belongs somewhere else.
Does ChatGPT selling ads replace this work?
No, and the two sit in different places on the screen. OpenAI has run ads in ChatGPT since February 2026, with a self-serve Ads Manager from that May. They appear as a labeled Sponsored block beneath the answer. The answer itself is still assembled from what the model retrieves and trusts, which is what these eight stages work on. Buying the ad slot does not put you inside the response.
See which stage your brand is stuck at.
Request an AI visibility audit. We benchmark you and three competitors across the major assistants, then show you which of the eight stages is holding the rest back. A 15-page report and a 30-minute walkthrough, at no cost.