Over the last six months, after every initial call with a client, they do the Columbo towards the end of the call and hit me with the “just one more thing…” and I already know what they’re going to ask:
“How do you rank in AI?”
It got to the point that I made a slide deck to answer the question which might not have been what the person on the other end of that question was expecting, but they left entertained if not a little bombarded.
If you’ve somehow skirted around the very useful Reddit posts on this subject and still don’t really know the simple answer to this, then the short answer is this:
You rank in AI search the same way you rank in regular search.
AI search isn’t a replacement for traditional SEO. It’s maybe another branch of SEO, in the same way that Local SEO is, but anyone who talks about Generative Engine Optimisation (GEO) or Answer Engine Optimisation (AEO) or AI Engine Optimisation (AIEO – just a U short of the full vowel set) is probably trying to sell you SEO software that’s positioned in a way to take advantage of the AI bubble.
AI search will look to the same things that established search engines look to. If you’re linked to a lot or mentioned a lot, then AI engines will also acknowledge this.
When you ask these AI assistants questions, they sometimes just outright do a Google search for you and talk you through the results.
For things like the AI Overview in Google, this is even more clearly drawn from the same sources as traditional search. If the AI Overview replaces anything, it’s the Rank 0 answer boxes that give you the very top position in search results. Now you have a more in-depth version of that from multiple sources with more obfuscated attribution and questionable accuracy.
Trust issues with AI
AI has already got a bit of a reputation for making things up and having questionable accuracy. If AI results were also bringing in websites that never appeared in Google’s top results, then it wouldn’t take long for you to question how legitimate its statements were.
You could argue that AI can get you more specific long-tail results from extended conversations, but this is still going to be drawing on the same things a specific long tail Google search would return.
Google has in fact been using AI in its search algorithm since 2015 with its RankBrain update, but it was much more behind the scenes and not as flashy or eye catching as generative AI with its many fingered elephants and abject flattering speech patterns.
If you’ve been doing SEO in line with Google’s guidelines and the SEO community’s consensus over best practices since then, then you’re already doing AI friendly SEO. I’ve read more than one “Everything about the way you do SEO has changed now” articles that goes on to describe almost the exact process I’ve been using for 10 years.
AI Search is evolving, but that’s the internet
The caveat is that this is absolutely true for now, but who knows about tomorrow?
The issue with that argument is that it’s the same for everything on the internet. Google sometimes changes something dramatic without warning and suddenly your business doesn’t rank for something that you didn’t realise was the only reason you’re making money. On a more B2C level, they might one day decide to opt you into a social network linked directly to your GMail account that’s functionally useless but inserted everywhere. Google’s not the only one guilty of this either – I gave up trying to keep up with Facebook’s API that seemed to change in subtle ways that would break everything I was pulling from it every three months. Entire services and systems disappear, merge with other things, become premium, or otherwise just stop being useful.
The internet is evolving, or devolving, or revolving, depending on who you talk to. Nothing stays the same and AI is no different. But this hasn’t been the revolution that has been advertised, at least not for search.