Is "AI visibility" just off-site reputation with a new name?
I keep digging into what actually gets a product mentioned in ChatGPT or Perplexity answers, and the data keeps pointing somewhere that makes me question the whole "GEO/AEO" hype.
Ahrefs looked at 75k brands and found branded mentions correlated with AI visibility way more than backlinks — 0.664 vs 0.218. It's correlation, not causation, so take it with salt. But the gap is big enough that I can't ignore it.
Which makes me wonder if we're overcomplicating this. Maybe AI visibility isn't a new discipline at all. Maybe it's just the old thing: be genuinely talked about in the right specific places.
But then the instability spooks me. Reddit reportedly went from around 60% of ChatGPT citations down to 10% after a single update. If a channel can swing that hard, is any of this stable enough to actually build a strategy on?
So for the founders here — are you doing anything on purpose for AI visibility yet, or ignoring it until it settles down? I'm honestly torn on whether this is a real 2026 priority or just a shiny distraction.
Replies
the Reddit swing is actually the argument for your "not a new discipline" theory, not against it. if AI visibility were its own stable channel with its own rules, a single update wouldn't be able to erase 50 points of it overnight. what actually happened is the model reweighted which real-world sources it trusts, and Reddit got deprioritized. that's not a channel collapsing, that's a proxy for reputation moving, same as it would if Google devalued a link category. the practical answer for founders is you can't budget for "AI visibility" as its own line item because there's no lever attached only to it, you budget for being genuinely discussed somewhere real and the AI mentions are just the current thermometer reading, not the thing you're actually optimizing.
@galdayan Yeah, you've reframed it better than I did. "Thermometer, not the thing you optimize" is exactly right — the mention is the reading, not the lever. I was treating the volatility as a reason to distrust the whole thing, and you're right that it's the opposite: it only swings because it's downstream of something real moving.
Where it gets practical for me is the next layer down. If you can't budget for "AI visibility" directly, you're really budgeting for "be discussed in the specific places this model currently trusts." But that set isn't obvious and it clearly shifts. So the actual work becomes figuring out which real-world spots move the reading for your category — for one business it's a niche forum, for another it's industry review sites, for another it's a couple of "best of" lists that happen to rank.
Which almost makes it sound like old-school PR and community presence with a live feedback loop bolted on. Are you finding the "where" is stable per niche even when the platform weighting isn't? That's the part I can't tell yet.
Building a GEO tool in this space after ~12 years in traditional SEO, here’s my take on whether it's actionable or just a distraction:
While off-site reputation is the foundation, there are very specific, tactical things you can do right now that go beyond traditional PR:
Intersect Backlinks with Cited Links: Comparing where you have traditional links vs. where LLMs actually pull citations gives you a much clearer roadmap for long-term prioritization.
Analyze Fan-Out Queries: Tracking fan-out queries across Search and Shopping yields gold. It helps optimize both on-site content and your product feed descriptions for Google Merchant Center.
LLM Perception Backtracking: If you run enough conversational queries, you can map out the exact "brand perception" an LLM holds about you. Once you see that perception, you can backtrack to where the model learned it and build a strategy to rebalance or correct it.
So is it a priority? It definitely helps, but the actual impact depends on brand maturity and how aligned you want to be with where search is heading long-term.
@voiqu
ok the perception backtracking one is new to me and i like it a lot. thats the part i keep hitting. you can see the model saying something wrong about you but the "where did it even get that" is the hard bit.
so when you backtrack, whats actually the giveaway? is it usually one loud source the model leans on, or more spread out and you just nudge the whole mix till the answer changes? for me the wrong info is always way stickier than the right info, kind of annoying.
and your first point makes sense too, the gap between where you have links and where you actually get cited feels like the real map. most people only looking at half of it.
@saied_alimoradi
I just ran a check on a larger brand we monitor to double-check this exact dynamic, and the data backs your hunch: it’s almost always 2 or 3 "loud" anchor sources rather than a diffuse mix.
When you isolate the conversations with wrong or negative perception, the incorrect info usually zeroes in on a specific product/topic tied to a couple of recurring URLs: an old review, a forum thread, or a legacy comparison post.
Why wrong info is so sticky: It’s an authority trap. Old, inaccurate articles often sit on domains with years of accumulated backlinks and index weight. Your updated official page might have the current truth, but that legacy source has higher retrieval weight.
That’s why "nudging the whole mix" rarely works. The move is to spot the anchor source and focus on creating fresh, highly-citeable counter-content targeted specifically at that exact sub-topic to win the retrieval context.
@voiqu
this is great, thanks for actually running it. the "authority trap" framing clicks for me. the legacy source wins on retrieval weight even when its just plain wrong now, so out-publishing it on that exact sub-topic makes sense. you cant argue with the model, you just have to give it a better thing to grab.
the bit i didnt expect is that its only 2-3 anchors. makes it feel more fixable than i assumed, you dont have to boil the ocean, just find the loud ones and go at those specifically.
anyway this was one of the more useful exchanges ive had on here. following you, would be good to stay in touch as this space keeps moving.