Don't focus on 'mention building'. Instead, build verifiable trust.

I get it, you're used to links being important for SEO. So it's a familiar pattern of thinking for you. And then ChatGPT & co. come along, with their 'brand new' (I can tell you it's not, but that aside) ability to recognize your brand mentioned. Suddenly you have the epiphany: we should build mentions!

LLM optimization then becomes a kind of PR, with the goal to be mentioned as much as possible in as much as possible prompts…

Sound familiar? If you ask me, I'd say that this is yesterday's thinking applied to today's challenges.

I'm not saying that PR is unimportant. And I'm also not saying you shouldn't take care of your brand narrative.

I am saying though, that any PR campaign will fall flat on its face, backfire even, if your real reputation isn't up to par. So much so, that you should focus really hard on this. LLMs take notice of this stuff.

What do I mean with 'real reputation'? Let's first talk about what it isn't:

Your reputation is what people say about you when you're not around. This applies to you personally in life, and also to brands on (real) social media, where people (not bots) meet and discuss things. Think about Reddit, forums, review sites.

Sure, mentions matter. But here's the thing: mentions are consequences, not causes. And chasing consequences isn't the right approach.

The things traditional SEO taught us - getting mentioned in the right publications, accumulate the right signals - are good ways to build topical visibility.

But LLMs increasingly see right through this and pay more attention to the underlying causes: the actual experiences people have with your brand. And that matters quite a lot, because ...

Traditional SEO rewards visibility of positive signals. The underlying reality matters less than the surface-level signals you can manufacture or earn through PR campaigns.

LLMs interpret language, conversations, and opinions differently. They're trained on the texture of human discourse about brands - not just counting mentions, but understanding what those mentions actually mean. They're doing something closer to: "What would a knowledgeable person who's read everything say about this brand?"

And knowledgeable people smell bullshit pretty quickly.

LLMs want to direct people as risk-free as possible to the candidate that exactly matches their question and takes into account their (increasingly personalization-based) preferences.

And here's the kicker: LLM recommendations are becoming more and more personalized. They're not just finding "the best brand in category X"—they're finding "the best brand in category X for you, given your specific situation, preferences, budget, location, and context." That's not necessarily the brand that puts a lot of focus on PR.

PR usually has little to do with products and services themselves. Things like honoring agreements, delivery time, service level, pricing transparency, complaint handling - that's the stuff that matters.

That stuff doesn't appear in press releases or media coverage. It comes back on review sites and forums instead. LLMs weight authentic customer experience over promotional content. They notice patterns in complaints. They distinguish between "talked about a lot" and "trusted for this specific use case." They pay attention to whether you deliver on your promises, how you handle problems, what customers actually experience versus what you claim.

In fact, it gets even worse.

I think that external PR focus that always worked well in SEO might actually backfire here. Why? Because a brand with heavy PR presence but weak operational fundamentals creates a gap that's visible in LLM training data. Lots of promotional content and media mentions, but then scattered threads of disappointed customers, unresolved complaints, delivery failures, service issues.

That dissonance is exactly what LLMs can and will detect. Traditional SEO saw: "Brand mentioned in TechCrunch + brand mentioned on industry blogs + brand mentioned in press releases = authority signal." LLMs see: "Brand claims fast delivery in all promotional content, but 47 Reddit threads complain about late shipments and poor customer service responses = trust risk."

And as personalization increases, this gets even more nuanced. An LLM might recommend you to one type of customer while steering another type away - based not on how much you're mentioned, but on whether your actual strengths and weaknesses match what that specific person needs.

LLMs see your brand mentioned, yes. They also see your competitors being mentioned. But more importantly: they see what people say about all of you when those mentions happen. The context. The sentiment. The gap between claims and reality.

So when someone pitches you "LLM mention building" or "AI visibility campaigns" or whatever rebranded link building service they're selling this month - ask yourself: are they helping you fix the underlying causes, or selling their non-useful product?

Because here's what's really happening: the same people who sold you guest posts and "high-quality backlinks" have simply found a new angle. They're betting you'll apply the same mental model (more mentions = better rankings) to a fundamentally different system.

The real challenge isn't getting mentioned more. The real challenge is making sure that when LLMs interpret the totality of discourse about your brand - the reviews, the forum discussions, the customer complaints, the Reddit threads, the actual experiences people share - that they understand you as trustworthy, reliable, and genuinely good at what you claim to do.

You can't buy your way to that. You can't PR-campaign your way to that. You have to be that. Focus on your real reputation. Fix your operational fundamentals. Deliver on your promises. Handle complaints well. Make customers actually happy. Then make sure LLMs see it. That's a different game.

The mentions will follow as a consequence - and more importantly, they'll be the right kind of mentions. The kind that make LLMs recommend you with confidence, not cautionary tales that make them hedge. The kind that work even as recommendations become increasingly personalized to individual needs.

Don't let yesterday's link sellers distract you from today's actual work.