Why your SEO agency might not be talking about ChatGPT yet
If you employ an SEO agency and you are wondering why they have not raised the topic of AI answer engines yet, you are not alone. I have had three different in-house heads of growth ask me the same question this year, almost word for word. Their agency talks about Google. Their CEO talks about ChatGPT. The agency reports say their organic traffic is “stable”. The buyers are asking AI for product comparisons. Nobody in the agency meeting is connecting the two.
It is worth understanding why. The reasons are structural, not malicious, and once you see them you can have a much more useful conversation in your next quarterly review.
Structural reason one: the metrics still look fine
Most agency reporting is built around Google. Sessions from organic, keywords ranking in the top ten, pages crawled, Core Web Vitals, click-through rate from the SERP. Those metrics are mature, well-defined, and instrumented.
The problem is that they can be flat while the underlying business pipeline degrades. If ChatGPT is now answering a third of your would-be top-of-funnel queries without sending the user to your site at all, your sessions can be unchanged and your qualified pipeline can be drying up. The agency report does not show this, because the agency report cannot see it. There is no equivalent of Google Search Console for ChatGPT, no impression count, no click-through rate.
So agencies show what they can measure. The metrics look fine. Nobody is being deceitful. The model is just no longer complete.
Structural reason two: the playbook is genuinely new
SEO as a discipline has thirty years of accumulated practice. There are clear specialisms (technical, content, off-page, local, e-commerce, international) with established certifications, conferences, and reference books. An agency hires for those specialisms and assigns them to clients.
GEO has roughly three years of accumulated practice. There is no certification worth the paper. The reference books are six months out of date by the time they are printed. The senior people doing this work are mostly independent consultants or in-house specialists at large brands, not staff at mid-sized SEO agencies. That is changing, but slowly. Most agencies are figuring out how to retrain their existing staff, and that takes time.
The honest agencies will tell you this. They will say something like “we are building our GEO practice and not ready to charge for it yet”. The less honest ones will rebrand their existing content audit as a “GEO audit” and hope you do not notice.
Structural reason three: the pricing model is awkward
Most agency retainers are built around fixed monthly activity. Twenty pages of content per month, ten backlinks, four technical fixes, one report. The work scales linearly with the retainer. Everyone understands what they are paying for.
GEO does not really fit that. The work is half analytical (figure out what queries the engines are being asked, who they are currently citing, where the gaps are) and half craft (rewrite the pages, restructure the schema, build authority signals across the web). The deliverables are less neatly countable. A single page rewrite that earns six AI citations is worth more than thirty pages of generic content. But agency pricing pages are not built around “one really good page”.
So agencies that try to sell GEO under their existing model end up either underpricing the work and burning out, or padding it with familiar SEO activity to justify the retainer. Neither is what the client actually needs.
What to ask in your next quarterly review
If you take nothing else from this article, take these five questions to your next agency meeting. The quality of the answers will tell you most of what you need to know about whether your current setup is fit for purpose.
1. Where do we currently appear in ChatGPT, Perplexity, and Google AI Overviews for our most important commercial queries? A competent agency, even one without a formal GEO practice yet, should be able to run these manually within a week. If the answer is “we do not track that”, at least you both know where you are.
2. Which of our pages have proper Organization, Person, Article, and Service schema, with cross-referenced @id URIs? This is the foundation of AI citation readiness. If the agency does not have a confident answer, the foundation work has not been done.
3. What is our public-web reputation footprint? Mentions in industry publications, podcasts, expert quotes, conference talks. AI engines corroborate against the broader web. If you have no presence outside your own domain, the engines have no signal to amplify.
4. What proportion of our buyer queries do we estimate are now happening inside AI engines rather than Google? Nobody knows precisely. A good agency will have a defensible estimate based on your category and your buyer profile. A useless agency will deflect entirely.
5. If we wanted to invest in GEO over the next twelve months, what would you recommend, and what would change in our retainer? The answer should be specific. Vague answers (“we would integrate it across all your activity”) are a signal that there is no actual plan.
A few caveats
I do not think most agencies are bad. The structural pressures I have described are real, and the better agencies are working through them seriously. Some are genuinely ahead. The point of this article is not to undermine your existing relationship. It is to help you ask sharper questions.
I also do not think most clients should fire their SEO agency and hire a GEO specialist. The bulk of the foundational work, the technical audit, the content quality, the link profile, is the same in both disciplines. What you may need is to add a GEO-focused conversation alongside the SEO one. Sometimes that comes from the same agency learning the new discipline. Sometimes it comes from a separate specialist. Sometimes the agency relationship simply needs to evolve.
And to be transparent, my own consulting practice sits in exactly that space. So take this article as the perspective of someone with a position, and not as a neutral survey. The questions above are still the right questions.
Where this leaves you
The conversation about AI search is happening at your CEO’s desk whether or not it is happening in your agency meeting. The most useful thing you can do this quarter is bring it into the agency meeting, with the five questions above, and see what comes back. The answers will tell you whether you have a current partner who can evolve with you, or whether you need a different conversation altogether.