I’ve said it plainly before: my agency is already dead, people just haven’t started covering up the body yet. This post is the practical version of that sentence — what actually replaces the flat monthly retainer, and why AEO and GEO are an expansion of SEO rather than something that makes it obsolete.
The old flat-fee SEO retainer is being replaced by a hybrid model: strategy, judgment, systems design, editorial oversight, and implementation accountability. AI can now do much of the production work agencies used to sell by the hour — but it doesn’t replace the need for someone who knows what to build, what to ignore, what good looks like, and how the work connects to business outcomes.
Why Is the Flat-Fee SEO Retainer Actually Dying?
For years, the retainer model worked because the deliverables inside it — fixed word counts, monthly audits, manual link-building — were expensive enough to justify a recurring fee. Commodity content production and basic technical audits have gotten cheap enough that clients notice when they’re paying agency rates for work a tool now does in minutes — and once they notice that, the reporting call that used to justify the retainer starts to feel like the thing they’re actually paying for, not the work.
Old Retainer vs. AI-Assisted Agency Model
Laid side by side, it’s less “AI killed SEO” and more that the parts of the job that were always production got cheap, and the parts that were always judgment didn’t.
| Dimension | Old flat-fee retainer | AI-assisted agency model |
|---|---|---|
| Core value | Access to recurring SEO labour and deliverables | Strategic judgment, systems, and accountable implementation |
| Content production | Briefs, drafts, and edits sold as time-consuming work | AI accelerates drafts and variants; I own the standards, evidence, and voice |
| Technical SEO | Periodic audits and ticket lists | Continuous diagnosis, prioritization, and implementation support |
| Reporting | Monthly activity summaries and ranking snapshots | Decision-focused measurement tied to visibility, leads, and revenue |
| Pricing logic | Hours, deliverables, and retained capacity | Outcomes, systems, advisory depth, and speed of implementation |
| Risk | Looks replaceable the moment AI can produce the same artifacts faster | Stays valuable as long as it owns judgment, integration, and results |
What Do Clients Actually Ask For Instead?
The agencies that survive this aren’t selling deliverables anymore — they’re selling systems and outcomes.
In practice, that’s the shift behind the three offers I actually run now: an AI Workflow Audit that maps what’s automatable in a business before anyone commits to a build; a System Build for the dashboards, content pipelines, and integrations a business already knows it needs — the same kind of build-out I documented in how a well-intentioned cron job nearly broke my AI assistant; and Franchise & Multi-Location Programs for the site architecture and governance that keeps forty locations from publishing forty different versions of the truth. None of those are a retainer for standardized monthly activity — each one is scoped to an outcome, priced against the work it actually takes, and done when it’s done (or renewed on purpose, not by default).
Is AEO and GEO Replacing SEO, or Expanding It?
This is where I part ways with both the doomer take and the hype take. The agency model — the flat retainer for standardized deliverables — is what’s dying. The discipline underneath it, understanding how information gets found and trusted, is expanding into every surface a client’s customer might use: a traditional SERP, an AI answer, a social platform’s own search.
That’s not a contradiction with “my agency is already dead” — it’s the same sentence, said twice, about two different things. I laid out the fuller argument in My Agency Is Already Dead — and it’s the same pattern Meta is running internally, recording how its own people work so it can automate the job.
What Comes After the Retainer?
Strategic partnership, not a deliverable list. If you’re not sure which of the three offers above fits what you’re actually dealing with, that’s exactly what the first one — the audit — is for: you get the map either way, whether or not you go further with me.
Frequently Asked Questions
Is the SEO retainer really going away?
The flat-fee, fixed-deliverable version of it is. Clients are increasingly unwilling to pay for standardized activity that AI tools now do cheaply — but the underlying discipline isn’t disappearing, it’s spreading into new surfaces.
What’s the difference between AEO, GEO, and traditional SEO?
They’re expansions of the same discipline, not separate fields — SEO for traditional search rankings, AEO for how answer engines surface a direct response, and GEO for how generative AI systems represent and cite a brand.
What should a business actually do instead of signing a retainer?
Start with a workflow audit — a fixed-scope session that maps what’s automatable now, what needs data cleanup first, and what should stay human — before committing to any build or retainer.
Does this mean agencies are obsolete?
The agency model built around recurring, standardized deliverables is what’s obsolete — not the expertise. The businesses that survive shift from selling activity to selling systems and outcomes.
How does this connect to “My Agency Is Already Dead”?
It’s the same argument, extended: the agency is dead in its old form, and what replaces it is exactly what this post describes.
Not sure which one you need? That’s what the audit is for. Book a call.
About the Author
David Henderson is an SEO and digital marketing practitioner with more than 25 years of experience across search, websites, content, analytics, and digital strategy — including enterprise and franchise SEO for brands like AutoTrader Canada, Honda, Canadian Black Book, Skedaddle, and dozens of franchise systems across North America. His current work focuses on practical AI integration for SEO agencies and client-facing marketing systems: where automation improves delivery, where human judgment still matters, and what agencies need to change as search evolves.