AI citations might be the new ranking factor but not in the way we thought....

Hey, it’s Amit here.

Over the past few weeks, a few separate threads have been quietly converging.

Individually, each is interesting.

Together, they suggest something bigger:

AI visibility might not be a new game.
It might just be SEO… observed through a different lens.

Let me explain.

Joshua Hardwick teamed up with SurferSEO to analyse:

They found a moderately strong Spearman correlation (0.41) between:

In plain English: If your brand appears more frequently on the pages AI already cites, you’re more likely to be recommended  -  and higher up in the response.

That’s not causation.

But it’s not random either.

And here’s the question that matters: How do you get mentioned on frequently cited pages in the first place?

That’s where this stops being about “AI optimisation” and starts looking suspiciously like… SEO.

Let’s follow the logic carefully.

Pages tend to get cited by AI systems because they:

If your brand is mentioned on those pages, then your probability of being recommended increases.

So the chain looks like this:

Good SEO + relevant links + brand authority
→ stronger rankings and authority
→ higher likelihood of being cited
→ higher likelihood of being recommended

That’s not a hack.

That’s reinforcement.

And if citation is correlated with recommendation strength…

AI citations might effectively be a new visibility layer sitting on top of traditional ranking signals.

This is where John-Henry Scherck made a good point on LinkedIn.

When AI citations come up, the conversation immediately becomes:

“But CTR is only ~1%. What’s the value?”

We’ve been conditioned to prove everything through clicks.

But if you’re cited:

Clicks are downstream.

Narrative control sits upstream.

If AI systems increasingly act as an interface layer between user and web…

Then citation is not just a traffic event.

It’s a positioning event.

And that changes how we think about value.

Here’s where this gets uncomfortable for some SEOs.

Over the past year, I’ve seen a strong swing toward:

I understand the reaction.

AI Overviews reduced clicks.
Top-of-funnel traffic became less predictable.
Revenue pressure increased.

But informational content does something revenue pages rarely do:

Revenue pages convert.

Informational pages distribute.

And distribution is upstream of:

If AI citations correlate with recommendation strength…

And cited pages tend to be informational reference surfaces…

Then abandoning informational content entirely may indirectly weaken your AI visibility.

Not because you lost traffic.

But because you lost distribution infrastructure.

That’s the second-order effect.

We may have overcorrected.

Not all informational content was vanity.

When it sits within your topical sphere and builds real authority, it compounds.

We might be going full circle.

I attended the Glasgow SEO Gathering meetup (shout out to Andrew Bond from Skyscanner!)  in February. 

Jérôme Salomon from OnCrawl walked through Bing’s new AI Performance report in Webmaster Tools.

It shows:

Now we all know Bing is not driving that much traffic anyway - but it’s the beginning of something important.

It’s the first mainstream sign that: AI citation is becoming measurable.

It looks a lot like an early-stage “AI Search Console.”

Jerome also shared an interesting point about analyzing your own server logs - that this is most likely, the most accurate way to actually understand your AI visibility. 

More bots visiting your page from ChatGPT, Gemini etc - shows your page is likely being cited. I’ll share a resource later.

Jerome also highlighted something fascinating at the meetup and on LinkedIn -  a leaked-looking ChatGPT publisher report structure.

The fields included things like:

That looks very similar to:

Impressions → citation visibility → interaction → click modelling.

We don’t know how real or final that structure is, and what exactly ChatGPT will release to everyone else.

But it suggests infrastructure is forming.

And when infrastructure forms, behaviour follows.

If citation visibility becomes a tracked metric…

Teams will optimise for it.

The question is whether they optimise intelligently  - or tactically.

I’m not saying:

This is correlation.
It’s directional.
It’s emerging.

But the logic aligns with how search systems have historically behaved:

Authority compounds.
Relevance matters.
Distribution reinforces visibility.
Presence across trusted surfaces increases probability.

AI visibility may not be a new game.

It may be the same game:

Just observed through a different interface.

And if that’s true…

The teams that invested in:

…may discover they were building AI visibility long before the dashboards arrived.

Are you seeing citation visibility in Bing yet?

And have you deprioritised informational content over the past year  -  or doubled down on it?

Hit reply. I read every response.

—

Amit Raj
The Links Guy

P.S.
If AI citations are becoming measurable, don’t optimise for the screenshot. Optimise for the probability.