I thought SEO had another two years before AI really cooked it.
I might have been very optimistic.
The launch of Meta’s Muse, combined with tools such as OpenSEO, gives us a pretty good preview of where the SEO industry could be heading.
Not in five years.
Potentially much sooner.
And the biggest problem isn’t that AI can write an article.
We’ve had that for years.
The difference now is that AI is becoming an agent that can actually do the work.
Muse Changes the Conversation
Meta launched Muse in September 2026 as a personal AI agent designed to go beyond answering questions.
Muse can use its own browser, navigate websites, fill out forms and carry out tasks on behalf of the user. Meta says it can even continue working on longer tasks after the user closes the app.
That is an important distinction.
ChatGPT originally gave you information.
AI agents increasingly take the information and do something with it.
Now combine that idea with specialized SEO infrastructure.
Then OpenSEO Arrived
OpenSEO is basically giving AI agents an SEO toolbox.
It connects agents such as Codex and Claude to real SEO data through MCP.
Instead of asking an AI:
“What keywords should I rank for?”
and hoping the model guesses correctly, the agent can access actual keyword information, SERPs, competitor data, backlinks, Search Console information, rank tracking and site audits.
OpenSEO has already created repeatable workflows for:
🔎 Keyword research
🩺 SEO audits
📊 Competitor analysis
🗂️ Keyword clustering
📍 Local SEO
🔗 Link prospecting
📈 SEO reporting
And you can run many of them simply by asking the AI agent to do the job.
That is where things get interesting.
What Are Businesses Actually Paying SEO Agencies For?
Take a typical small or medium-sized company’s monthly SEO package.
A significant portion of it often consists of:
Keyword research.
Competitor research.
Technical audits.
On-page optimization.
Internal linking.
Content briefs.
Blog articles.
Rank tracking.
Search Console analysis.
Monthly reporting.
Maybe some backlink outreach.
Ten years ago, performing all of this required several different tools and quite a lot of manual work.
Even two years ago, someone still needed to move between Ahrefs, Semrush, Search Console, spreadsheets, WordPress and ChatGPT.
Now we are moving toward:
“Analyze my website, compare it with my competitors, find the opportunities, tell me what needs fixing and create the content.”
Then the agent goes and does most of it.
That fundamentally changes the economics of SEO.
The Problem Isn’t That SEO Is Dying
Google isn’t disappearing tomorrow.
Search isn’t disappearing.
Businesses will still want organic traffic.
They will still want to appear in Google.
And increasingly, they will also want to appear in ChatGPT, Gemini, Claude, Perplexity and whatever AI discovery platform comes next.
What could disappear is the amount of manual SEO work you can reasonably charge for.
If something previously took an SEO specialist five hours and an agent can produce a strong first version in five minutes, the value of those five hours changes.
That’s the uncomfortable part.
Articles Are Already Mostly Cooked
Content is probably the easiest example.
Creating a decent SEO article used to involve:
Researching the topic.
Analyzing competitors.
Finding keywords.
Building an outline.
Writing the article.
Creating metadata.
Adding FAQs.
Finding internal links.
Optimizing headings.
Today an AI agent can already handle a large percentage of that workflow.
The human becomes the editor and strategist rather than the person physically producing every word.
And as agents gain better access to CMS platforms, analytics and SEO data, even more of that workflow will happen automatically.
On-Page SEO Is Heading the Same Way
On-page SEO is also extremely automatable.
An agent can inspect:
Titles.
Meta descriptions.
H1s and H2s.
Internal links.
Schema.
Image alt text.
Canonical URLs.
Thin pages.
Duplicate content.
Content gaps.
Broken links.
Search Console performance.
Then it can generate recommendations.
The next obvious step is allowing it to implement those recommendations.
That turns:
SEO audit → spreadsheet → developer → implementation
into:
SEO audit → AI → approval → implementation
That’s a completely different workflow.
Authority Is the Part AI Hasn’t Fully Cooked Yet
This is where I still see a meaningful moat.
You can optimize your own website all day.
But you cannot simply decide that Forbes, TechCrunch, Reddit, an industry publication or another respected website will mention you.
Real authority comes from outside your website.
Backlinks.
PR.
Editorial mentions.
Partnerships.
Communities.
Directories.
Industry relationships.
Original research.
Brand searches.
People actually talking about your company.
OpenSEO already includes link prospecting, meaning an agent can identify potential websites and outreach opportunities.
But finding a prospect and earning a legitimate mention are two very different things.
That still requires distribution, relationships, negotiation, creativity and often something genuinely worth talking about.
For now.
But Give Agents Another 3–6 Months
This is my prediction, not something Meta or OpenSEO has announced.
I think the next big step will be agents becoming dramatically better at the authority side of SEO.
Imagine telling an agent:
Find 100 relevant websites in my niche, identify the right contact, understand what each publication writes about, create a personalized outreach angle and prepare the campaign.
Parts of this are already possible.
The question is how reliable and autonomous these workflows become.
If agents get genuinely good at authority building, digital PR and outreach, another large portion of traditional SEO work gets compressed.
Maybe that happens in three months.
Maybe six.
Maybe longer.
But I don’t think we’re talking about years anymore.
The €1,000 SEO Retainer Has a Problem
This doesn’t mean SEO consultants disappear.
The good ones probably become considerably more productive.
But the agency selling:
4 articles.
Some metadata changes.
A few internal links.
A monthly ranking report.
And an automated audit.
For €1,000–€2,000 every month?
That business model is going to become increasingly difficult to justify.
Especially when a business owner can increasingly ask an AI agent to perform much of that work directly.
SEO Agencies Will Need to Move Up the Value Chain
SEO agencies will have to sell something AI cannot easily reproduce.
Strategy.
Original data.
Brand positioning.
Authority.
PR.
Relationships.
Creative campaigns.
Digital assets people actually want to reference.
Technical implementation involving complex systems.
Real-world expertise.
And perhaps most importantly:
knowing what is actually worth doing.
Even OpenSEO makes an important distinction here. Its own documentation recommends automating predictable processes and letting agents help with drafts and analysis, while keeping strategic decisions with humans.
That’s probably what the next version of SEO looks like.
Not humans doing 100 tasks manually.
Humans deciding which 10 tasks matter while agents execute them.