Retainers / SEO & GEO
There are two search engines now. Only one shows you a link.
A growing share of buying research never reaches a results page — someone asks an assistant, gets an answer with three brands named in it, and clicks one. If you are not one of the three you were not outranked, you were never retrieved. That is a different discipline from SEO, it is measurable, and almost nobody in ecommerce is doing it deliberately yet.
- What it is
- A monthly SEO and generative engine optimisation retainer — classic organic search work plus the structured-data, entity and retrievability work that determines whether AI assistants can find, trust and cite you.
- Who it's for
- Brands who want compounding organic acquisition rather than a one-off audit, and who have noticed that a share of their category research is now happening inside AI answers.
- What GEO actually means
- Generative engine optimisation. Where SEO competes for a ranked link, GEO competes to be the source an assistant cites when it composes an answer. Different mechanics, overlapping foundations.
- What runs monthly
- Technical crawl and fixes · structured data and entity work · content briefs and optimisation · internal linking · AI visibility testing across assistants · written reporting on both channels.
- How AI visibility is measured
- We run your real buyer questions through the major assistants monthly and record whether you appear, which URL is cited, and whether what is said about you is accurate.
- Investment
- From $3,500 per month. Included from the Growth tier when bundled. Three-month initial term, then rolling monthly.
- What it is not
- Not link buying, not spun content, and not a ranking guarantee. Nobody controls the ranking, and anyone selling one is selling you risk.
The shift
Being unranked and being unretrievable are different problems.
Most SEO retainers still optimise entirely for the first column and quietly ignore the second — which is where an increasing share of high-intent research now happens.
SEO-only, in an AI-answer world
Everything is optimised for a results page fewer people reach
- AI crawlers are blocked or unconsidered in robots.txt, so citation is impossible regardless of content quality
- Key facts exist only in prose, which does not survive extraction into an answer
- Pages never name the tools or entities a question is about, so nothing links them to the query
- Claims appear only on your own site, with no corroboration for a model to weigh
- Assistants describe the business incorrectly and nobody is checking, let alone correcting it
- Reporting is clicks only, so influence that never produced a click is invisible
SEO and GEO together
Optimised for a ranked link and for being the cited source
- Retrieval crawlers explicitly allowed, verified rather than assumed
- Extractable fact blocks on every commercial page — the structure assistants actually lift
- Entity markup naming the real platforms and concepts, connecting the page to the question
- A maintained llms.txt stating the facts you want cited, including the ones you want corrected
- Monthly assistant testing on your real buyer questions, with misattributions tracked and fixed
- Reporting covers both — rankings and traffic, plus citation presence and accuracy
The four pillars
Three you know. One is new.
Classic SEO rests on technical, content and authority. Generative search adds a fourth — how easily a machine can retrieve, parse and trust what you published.
Technical
Crawlability, indexation, site speed, canonical hygiene, redirect health and structured data validity. Shared foundation for both channels — and the reason a GEO programme on a broken site is wasted money.
- Crawl and index control
- Redirect and canonical audits
- Core Web Vitals coordination
Content
Pages that answer one query completely, in the words buyers actually use, with the commercial questions — cost, timeline, what happens next — answered rather than dodged.
- Query-mapped page briefs
- Cannibalisation control
- Commercial intent coverage
Authority
Internal linking that concentrates equity on the pages that earn money, plus the external corroboration that makes a claim credible to both a search engine and a model.
- Internal link architecture
- Orphan page recovery
- Corroboration sources
Retrievability
The new one. Whether an AI system can reach your pages, parse a clean fact from them, connect them to the right entities, and trust the claim enough to repeat it with your name attached.
- AI crawler access
- Extractable fact blocks
- Entity and schema markup
- llms.txt maintenance
Retrievability
Whether an assistant can even reach you.
These are distinct crawlers with distinct jobs, and they are routinely blocked by a robots.txt written before any of them existed. Blocking the retrieval bots removes you from AI answers entirely.
| Crawler | Belongs to | What blocking it costs you |
|---|---|---|
| OAI-SearchBot | ChatGPT search index | You cannot appear in ChatGPT search results at all. |
| ChatGPT-User | ChatGPT live browsing | A user asking about you gets an answer built without your site. |
| PerplexityBot | Perplexity index | No citation in a product built entirely around citing sources. |
| Claude-SearchBot | Claude search | Excluded from answers where your page was the best source available. |
| Google-Extended | Gemini and AI Overviews | Removed from AI Overviews while still ranking normally in classic results. |
| GPTBot / ClaudeBot | Model training | A separate decision. Blocking these does not remove you from AI search, and is a legitimate privacy choice. |
The distinction between retrieval crawlers and training crawlers is the single most misunderstood point in this area. Blocking training crawlers is a defensible choice with no effect on AI search visibility. Blocking retrieval crawlers makes citation impossible.
The month
What actually ships each month.
New 404s, redirect chains, orphaned pages, broken canonicals and invalidated structured data caught and fixed before they compound.
One or two pages briefed, written or optimised against a mapped query, with schema and entity markup shipped alongside.
Your real buyer questions run through the major assistants. Presence, cited URL and factual accuracy recorded and tracked over time.
Rankings and organic traffic, plus AI citation presence and accuracy, plus what shipped and what is next. Sent before the call.
Investment
One retainer, both channels.
SEO & GEO retainer
Weekly technical crawl and fixes, monthly content and entity work, structured data maintenance, internal link architecture, monthly AI assistant visibility testing, and reporting across both organic search and AI citation.
A starting point for scoping rather than a quote — it depends on site size, catalogue depth, competitiveness and how much content production is in scope. Content written by us is included at the volumes stated in your scope; larger programmes are priced accordingly.
Questions we get
Straight answers.
What is generative engine optimisation, in plain terms?
Getting your brand named and cited when someone asks an AI assistant a question instead of typing it into a search box. Classic SEO competes for a position in a list of links. GEO competes to be the source the assistant uses when it writes an answer. The foundations overlap heavily — technical health, clear content, credible corroboration — but GEO adds retrievability: whether a machine can reach your page, extract a clean fact from it and trust it enough to repeat it with your name attached.
Can you guarantee we will rank first, or be cited?
No, and treat anyone who does with suspicion. Nobody controls a search ranking and nobody controls what a language model composes. What we commit to is the work that measurably improves the odds — technical health, correct entity and schema markup, retrievable content structure, crawler access, and monthly measurement of both channels so you can see movement rather than take our word for it.
How do you measure AI visibility when there is no rank tracker?
We run a fixed set of your real buyer questions through the major assistants each month and record three things: whether you appear at all, which URL is cited, and whether what is said about you is factually accurate. Tracked over time that shows movement clearly. The accuracy column matters as much as the presence one — an assistant confidently describing you incorrectly is a problem you can only fix on-page.
Should we block AI crawlers to protect our content?
That is two decisions, and conflating them is the most common mistake. Training crawlers like GPTBot and ClaudeBot collect data to train models — blocking them is a legitimate choice and has no effect on whether you appear in AI search. Retrieval crawlers like OAI-SearchBot, PerplexityBot and Claude-SearchBot fetch a page to answer a live question. Blocking those removes you from AI answers entirely, which is almost never what a brand actually wants.
How is this different from your strategy and analytics service?
The strategy, SEO and analytics service is fixed-scope project work — a full audit, a migration plan, an analytics rebuild. This retainer is the ongoing programme: monthly crawling, content, entity work and AI visibility testing. Brands often do the audit first and move onto the retainer to execute against it.
How long before we see results?
Technical fixes can move things within weeks. Content and authority work realistically takes three to six months to show clearly in organic traffic, which is why the initial term is three months and why we report leading indicators — indexation, crawl health, citation presence — rather than asking you to wait for revenue to confirm it. AI citation can move faster than rankings, because retrievability is often a fixable blocker rather than a competitive one.
Next step
Find out whether AI can even see you.
The free audit includes a retrievability check — crawler access, structured data validity, and your real buyer questions run through the assistants to see who gets named instead of you.
