AI is everywhere. In your workflow, in agency pitches, and for the past few months in commercial proposals promising to “multiply your organic traffic tenfold thanks to artificial intelligence.” Spoiler: it’s more complicated than that.
Here’s what I actually observe in the field, after testing, breaking, fixing, and occasionally crying in front of Google Search Console.
What AI genuinely does well in SEO
Let’s be honest: there are tasks where AI is genuinely impressive.
Generating topical clusters? Effective. In a few minutes, you have a content architecture that would have taken half a day to build manually. Content briefs? Clean, structured, usable. Analyzing search intent across a batch of keywords? Fast and often accurate.
AI excels at repetitive, structured tasks. It doesn’t complain, it doesn’t watch Netflix on the side, and it doesn’t ask for a coffee break. On that front, it beats any intern hands down.
What AI does less well (and what nobody tells you)
This is where it gets interesting.
AI hallucinates. Cheerfully. It will generate convincing statistics that don’t exist, cite studies that were never published, and assure you with total confidence that “according to a 2023 Stanford University study…” when said study came entirely from its imagination. Always verify your sources. Always.
Content generated without supervision? Google is getting better at detecting it. And more importantly, your readers feel it. That slightly flat, too-polished thing that says everything without really saying anything. The famous “quality content optimized to answer user questions.” We’ve all read that sentence a thousand times. It means nothing.
And then there’s the technical side. AI doesn’t understand your site. It doesn’t know that your cache plugin has been acting up since the last update, that your server is in Germany but your audience is Canadian, or that your 301 redirects have a complicated history since 2019. Context is something you bring. AI just responds to what it’s given.
The mistake everyone makes
People confuse speed with efficiency.
Yes, AI produces 40 articles in a day. But 40 empty articles means 40 pages diluting your domain authority, burning Googlebot’s crawl budget for nothing, and bringing zero value to your visitors. Low-quality mass SEO has always existed. AI just industrialized the stupidity.
The real question is: does this content deserve to exist? Does it answer a genuine question with genuine expertise? Will a human searching for this information leave satisfied?
How I actually use AI in my SEO
I use it as an acceleration tool, not a replacement.
For semantic research, to structure content plans, to find angles I wouldn’t have thought of naturally, to generate first drafts that I then rework. A bit like a first pass from a competent colleague who just doesn’t know your client yet.
What I never delegate to it: strategic analysis, in-depth technical audits, personalized recommendations. Because SEO, at its core, is about understanding a business, an audience, a positioning. That’s still human work.
What this means for you
If you manage your SEO in-house, AI can save you time on production. But don’t skip the “human brain that rereads, adjusts, and brings lived experience” step.
If you’re working with an agency selling you “AI SEO” at bargain prices, ask questions. Who’s actually writing? Who’s reviewing? Who’s doing the strategy? Because AI without a pilot is a self-driving car in a Parisian parking garage. Theoretically capable. Practically risky.
AI in SEO is a great Swiss army knife. But like any Swiss army knife, if you don’t know how to use it, you end up with a cut where you thought you’d make a clean slice.
Use it. Really. But keep your hands on the wheel.