Most experts aren't invisible because they're unknown. They're invisible because AI can't confirm they exist.

Hard truth time: AI systems don't recommend people they can't verify.

It doesn't matter how good you are at what you do. If the signals that would confirm your expertise don't exist in the right places, in the right formats, you won't appear in the answer. Someone else will.

This is the problem I work on at Radar. And it's more structural than most people realise. 

AI citation patterns are hardening right now. Here's the layer most thought leaders are missing and why it's the one that makes everything else compound. Buckle up, readers!

AI doesn't search. It cross-references

When someone asks ChatGPT, Perplexity, Claude or Google AI Mode who the leading voice on a topic is, the system isn't running a keyword match. It's looking for convergence.

Multiple independent sources, attributing the same expertise, to the same named person, across platforms it can actually read.

One strong website isn't enough. One active LinkedIn profile isn't enough. Even a combination of both often isn't enough because both of those sources are either brand-owned or self-published.

AI discounts self-referential signals the same way a good editor discounts a press release (especially a poorly worded AI one).

What it trusts is third-party confirmation. Your name appearing in someone else's content. A podcast host citing your framework. A journalist quoting your POV. A peer referencing your work without being asked to.

The research backs this up. More than 85% of AI citations come from third-party platforms, not owned content. Brands and experts mentioned positively across four or more non-affiliated platforms are 2.8 times more likely to appear in AI responses.

Your expertise needs to exist in layers, and those layers need to be built in a specific order.

The layer most experts skip is the one that makes everything else compound

Most thought leaders I audit have two things working: a reasonable owned presence (a website, consistent LinkedIn activity) and some executive-level visibility (they're an academic, founder or CEO; they show up when you search their name).

What they're almost always missing is what I call the expert layer — the structured, consistent, independently verifiable record of their specific thinking on a specific topic.

Not just posting. Not just sharing articles. Publishing original perspective, under their own name, with a defensible point of view, at enough volume and consistency that AI systems can build a reliable picture of what they know.

This is the layer that seeds everything above it. Third-party coverage happens when journalists and clients can find your thinking and confirm it's worth referencing. Podcast bookings happen when a host can verify, before they reach out, that you actually have something to say ie they've read your methodology on multiple different platforms. Community mentions happen when your framework has been in circulation long enough for others to use it.

None of that compounds without the expert layer underneath it.

This is an editorial problem, not a tech problem

Most visibility advice in this space defaults to technical fixes like Schema markup. structured data or entity optimisation.

Those things matter but they're downstream of a more fundamental question: does this person have a coherent, documented, citable body of knowledge that AI systems can find and trust?

That question is editorial. It's about what you say, how you frame it, how consistently you say it and whether the record of your thinking is legible to a system that needs to make a fast decision about whether you're a reliable source.

I come from over a decade of editing and the underlying skill in building AI visibility isn't technical. It's the same skill good editors have always used: taking what an expert knows and making it clear, structured and attributable enough that other people (or systems) can use it with confidence.

That's the work. And most experts haven't started it yet.

The window is narrower than it looks

AI citation patterns are hardening. The systems are building entity graphs right now — associating named people with topics, weighting those associations based on signal volume and source quality.

Once your competitor is anchored in those graphs, displacing them costs significantly more than building first.

The experts who move now don't need to outspend anyone. They need to become resolvable or visible before the category settles.

If you want to know where you currently sit — what AI actually knows about you and what's missing — that's exactly what a Radar Authority Audit surfaces.

The question isn't whether AI knows your name. It's whether it knows enough to trust you with the answer. Radar makes sure you're in the answer.