Found in Google. Recommended by AI.
Technical SEO, content and AI visibility run as one program, on one report, by one team. Fifteen years of search work, applied to a channel nobody has fifteen years in.
Most agencies bolted GEO onto an SEO retainer. It does not work that way.
The same content earns both. A page that answers a buyer’s question properly is the page Google ranks and the page a model retrieves. What differs is the shape of the work: retrieval rewards clean structure and third-party corroboration in ways ranking never did.
Running the two as separate programs means paying twice for overlapping work and receiving two reports that disagree with each other.
One content planBuilt once against real demand, structured to work as a ranking page and as a retrievable passage.
One authority programThe placements that move AI answers are largely the placements that move rankings. We go after them once.
One reportRankings, traffic, citations and share of voice in the same document, on the same cadence, with the same person explaining it.What actually happens when someone asks a model about your category
You cannot optimize for a system you cannot describe.
Retrieval
The model does not recall your page. It runs searches, pulls a set of documents and reads them. If your page is not retrievable for the query the model actually runs, nothing downstream matters. This is where technical work still decides everything.
Selection
From that set the model picks passages. It favours content that answers directly, is structured cleanly, carries specifics a summary cannot invent, and is corroborated somewhere other than your own site. This is why third-party presence does more work here than it ever did in ranking.
Citation
What gets named and what gets linked are different things. A mention puts your brand in the answer. A citation puts a link under it. The work that earns each is different, and a program that counts only one of them is measuring half the channel.
What the retainer covers
One program, one team, one report. The plans differ in pace, not in what is included.
Technical foundations
Crawlability, indexation, speed, internal structure and schema, monitored continuously rather than audited once. If a crawler cannot read the page, nothing else on this list matters.
Content built against real demand
Briefs built from search demand and from the prompts buyers actually use, then written for a person rather than to a word count.
Structure a model can retrieve
Direct answers near the top, clean headings, specifics a summary cannot invent, and schema that says what the page is. The work that earns a featured snippet is the work that earns a citation.
Third-party citation work
The sources models quote are mostly not your website. Editorial placements, comparison and roundup pages, communities and directories, worked deliberately rather than hoped for.
Migrations and releases without traffic loss
Redirect mapping, staging checks and post-launch verification. The cheapest way to lose a year of work is a launch nobody checked.
One report, every month
Rankings, organic traffic, citations, mentions and share of voice in one document, in plain English, with what we did and what we are doing next.
Measured, not asserted
The most common thing a buyer has been told about AI visibility is that it cannot be tracked. It can, imperfectly, and the honest version is worth more than the confident one. Here is exactly what we measure and how.
The prompt set
Every number on this page depends on which questions we ask. So we publish the rules.
What we report
Start with a diagnostic
A fixed-scope engagement that answers whether there is anything here worth paying for. Full technical crawl, a prioritized fix list your developers can work from, a baseline of where you currently appear across an agreed prompt set, and a recorded walkthrough. Delivered in [two weeks]. The fee is credited in full against the first month of a retainer started within 30 days.
Almost every agency in this category gives an audit away. We charge for it because it takes real work, and because a diagnostic somebody paid for is a diagnostic somebody reads.
Small
Full crawl, prioritized fix list, AI visibility baseline across [50] prompts, recorded walkthrough.
Book this diagnostic
Mid-size
Everything in Small, plus crawl budget and indexation analysis, Core Web Vitals, and structured data review.
Book this diagnostic
Large or complex
Everything in Mid-size, plus faceted navigation, international targeting, and developer-ready tickets.
Book this diagnosticWhat is realistic, and when
Search work compounds. Anyone promising otherwise is selling a spike. Here is the shape of a normal engagement, including the part where nothing visible happens.
Baseline
Technical fixes shipped, the prompt set agreed and frozen, tracking live across models and separated in your analytics. You get a number for where you stand today, which most companies have never had.
You get · a baseline you can argue withFoundation
Content and structural work on the pages that already have a chance. Movement usually shows in citations before it shows in rankings, because a model can change its answer in a week and a ranking cannot.
You get · first citation movementCompound
The content plan running at pace and third-party citation work landing. This is where rankings move in a way you would notice without a dashboard.
You get · measurable share of voicePosition
Category-level presence rather than page-level wins. The point at which the program is defensible rather than promising.
You get · a position competitors have to spend to takeWork shipped for
See the work →
Whether this fits
Pricing
Every plan covers technical work, content, AI visibility and reporting. They differ in pace and in how much of the category we go after.
Prices in USD. No setup fees. Diagnostic fees are credited in full against the first month when a retainer starts within 30 days. Third-party placement budget, where a plan calls for it, is quoted and approved separately before any spend.
Questions
If your question is not here, one of us will answer it directly.
Still have questions? →GEO stands for generative engine optimization: the work of getting your brand named and your pages cited when someone asks ChatGPT, Perplexity, Gemini or Google's AI Overviews a question you should win. Some agencies call the same thing answer engine optimization, or AEO. They are the same practice under two names, and choosing between the labels matters far less than whether anyone can measure the result.
Whether there is anything here
worth paying for
Send us your web address. We will look, and reply with what we would do and whether it is worth doing. That is an opinion, not an audit. The audit is the diagnostic, and it costs money because it takes work.