AI SEO uses AI tools to speed up tasks like keyword grouping, content briefs and data analysis, while manual SEO relies on a person's judgement and experience. The practical answer is to use both. Let AI do the repetitive work and keep humans responsible for strategy, accuracy and anything that needs real experience.
This debate is often framed as a contest. It isn't. After 14 years in SEO, I see AI as a faster pair of hands. It hasn't changed the question I care about, which is whether the work brings leads.
What AI does well in SEO
- Keyword clustering. Grouping thousands of queries by topic in minutes.
- Drafting metadata. First versions of titles and descriptions, which you then edit.
- Summarising. Competitor pages, reviews, sales call notes.
- Writing formulas and scripts. Spreadsheet formulas, regex for Search Console filters, basic schema markup.
- Finding content gaps. Spotting which questions a page doesn't answer.
- Log and data analysis. Reading big exports quickly. Verify the sums.
What still needs a person
- Strategy. Which keywords actually make money for this particular business?
- Judgement. Should we fix this, or leave it? Should we redirect, merge or delete?
- Real experience. Stories, results and opinions.
- Accuracy. AI can produce confident errors.
- Relationships. Link earning and PR rest on trust between people.
- Diagnosis. When traffic falls, you need someone who can test causes, not repeat likely ones.
- Client honesty. Telling someone SEO isn't right for them yet.
A task-by-task view
- Keyword research: AI suggests and groups. Real search data and a human decide.
- Content brief: AI drafts. A person adds the angle and experience.
- Writing: AI can assist. A subject expert must edit and fact-check.
- Technical audit: crawlers find issues, and AI can explain them. A person sets priorities.
- Link building: AI helps with research. Outreach should be personal. Mass AI emails get ignored.
- Reporting: AI can summarise. A human says what it means and what to do next.
The risks of leaning too far on AI
The biggest is scale without quality. Google's guidance says generating many pages without adding value can break its scaled content abuse policy. It doesn't matter that a tool wrote them. I've written about this in AI content marketing.
Two more risks are worth naming. AI sometimes gives out-of-date SEO advice, because old advice fills the web. And everyone's tools now produce similar content, so similar content gives you no edge at all.
The risks of avoiding AI
Doing it all manually is slower and costs more. If your competitor can produce a good brief in 20 minutes and you take three hours, you'll fall behind. Refusing the tools isn't a virtue. Used well, they free up time for the work that matters.
What about AI search itself?
Google says optimising for its AI features is still SEO. Its guidance says there's no special markup required, and that you don't need an llms.txt file for Search. It warns against AEO and GEO hacks. The basics carry over: indexable pages, helpful original content and a clear site. Treat claims of guaranteed "AI ranking" with suspicion. No third party has inside access to Google's systems.
How to choose an SEO partner today
Don't ask, "Do you use AI?" Almost everyone does. Ask:
- Which parts of my work will use AI?
- Who reviews it, and what are their qualifications?
- How do you add real experience?
- How do you check facts?
Good answers are specific. If you want a broader idea of how to pick someone, read how to hire an SEO expert. If you want to see how I work, take a look at my SEO services. Whatever the mix of tools, I'd judge it the same way: did leads and revenue go up?
Written by Aman Potlia.