Can AI Decide Your SEO Strategy, or Only Carry It Out?

Short answer: No. AI can do most SEO tasks, but it can’t decide which of them matter for your business. We use AI every day for research, audits and drafts. Our strategists still decide what gets targeted, what gets fixed first and what never gets published.

We get asked this a lot, usually by founders comparing a $99/month AI SEO tool with an agency retainer. It’s a fair question. The tool produces more, faster.

But from what we see in client accounts, AI is great at doing SEO work and poor at choosing which work is worth doing. Here are four examples of the difference.

1. AI Finds Keywords. It Doesn’t Know Which Ones Sell.

What we saw: An AI tool recommended a B2B software client target “free project management templates”, with high volume and low competition. It looked like an easy win. But the client sells enterprise software from $20k a year. Those searchers were students and freelancers, not buyers. We targeted lower-volume comparison terms instead. They brought far less traffic and far more demo requests.

The data was correct and the recommendation was wrong. Search volume measures demand, not business value. Fifty visits from people comparing vendors beat a thousand from people who will never buy.

2. AI Flags Every Issue. It Doesn’t Know Which Ones Matter.

 

What we saw: An automated audit for an ecommerce client flagged 150+ issues, all marked “high priority”. Once we reviewed them, only five really mattered, including a canonical error that was hiding a key category page from Google. Most of the rest, like missing alt text on old blog images, could wait for routine updates.

Development time is limited. Prioritisation is what turns an audit into a strategy. Without it, you just have a longer to-do list.

3. AI Says “Publish More.” Sometimes the Answer Is Less.

 

AI tools are built to find content gaps. If a competitor has a page and you don’t, they tell you to write one.

What we saw: A client’s AI content plan proposed 40 new articles. We found that 12 existing posts were already competing for the same searches. Merging them into three strong guides did more for their rankings than the 40 new posts would have, and it avoided years of content debt.

Being able to say “we shouldn’t create this” is one of the most valuable skills in SEO, and no tool is built to say it.

4. AI Doesn’t Sit in Your Sales Meetings

The best SEO decisions we’ve made came from conversations, not dashboards. For example:

  • Sales told us prospects kept asking about integrations, so we built pages around that instead of generic features.
  • Legal ruled out a comparison page the AI had recommended.
  • The product team was retiring a service the tool kept telling us to rank for.

AI works with what it’s given and can’t tell when it’s missing something. And when a strategy doesn’t work, a tool doesn’t explain it to leadership or decide what to change. A tool produces outputs. A person owns the decisions.

How We Split the Work at Thrillax

AI handles Our strategists decide
Keyword discovery and clustering Which keywords bring customers
Crawling and flagging technical issues Which fixes matter now
Content briefs and first drafts What to publish, merge or skip
Reports and monitoring What the results mean and what to do next

AI saves us hours on repetitive work. We spend that time on what clients actually pay for: judgment.

Quick Answers

Can AI replace an SEO expert?

No. It can replace much of the manual work, such as research, audits and drafts, but not the judgment that connects SEO to your business goals.

Should I use AI for SEO?

Yes, for speed. Just keep an experienced person in charge of strategy and priorities.

Is AI-generated content bad for SEO?

Not by itself. The risk is publishing large amounts of generic content with no real experience behind it.

Our view: AI gives us speed. Experience gives us judgment. Your SEO needs both.

Editorial note: Written by the Thrillax SEO team, based on our own work using AI in client SEO. Inspired by Nick LeRoy’s “The Dehumanization of SEO: Displaced by Code, Ignored by Communication“.