Ask ChatGPT who matters in your category. Not "what are the best tools" — ask it who. Who explains this well. Who do people trust. Who is building the interesting thing.
It will name people. Names you half recognize, names you've never heard, and occasionally your competitor's founder. It is citing humans.
Now ask the harder question: when AI explains your category to a buyer, does it name you?
If the answer is no, you've found the blind spot in your entire AI visibility strategy. Because you've been optimizing pages for an engine that increasingly cites people.
TL;DR: Generative AI has quietly changed the unit of trust from the page to the person. LLMs aggregate human consensus, and they name the humans whose opinions they've absorbed. This creates a new GEO discipline — Creator SEO: map the people AI already trusts in your category, then become part of what those people say. It inverts influencer marketing completely, because the metric that matters is no longer reach. It's citation weight inside AI answers. We've already built the audit machinery for this at Mercury, and every brand should be running it.
I am James, CEO of Mercury Technology Solutions. I run generative engine optimization audits for a living — hundreds of queries across ChatGPT, Perplexity, Gemini and Claude, measuring who gets cited and who gets ignored. And over the past year, one pattern has become impossible to miss: named humans outperform branded pages.
The Shift: From Page Rank to Person Rank
Classic SEO was a graph of pages. Google's PageRank flowed between URLs, and authority lived in domains. That world optimized for a crawler.
Generative AI runs on a different substrate: human consensus, compressed. These models learned from people arguing, explaining, reviewing, recommending. When they generate an answer, they reconstruct that consensus — and they reconstruct it with names attached. "As [analyst] noted." "According to [founder]." "[Expert] argues that..."
We see it in every audit we run. Ask about a software category and the answer cites three review sites and two named individuals. Ask about an industry trend and the person shows up before the publication does. The model didn't invent this hierarchy — it absorbed it. Somewhere in the training distribution, those humans carry disproportionate weight for that topic, and the model reproduces that weight every time it answers.
並非是品牌被引用，而是人被引用 — it's not the brand being cited, it's the person.
Here's the uncomfortable implication: you can have flawless technical GEO and still lose the answer, because the answer is being assembled from human voices and you aren't one of them.
What Is Creator SEO?
Creator SEO is the discipline of managing person-level citations inside AI answers. Three moves:
1. Map the voices. Run your category's query set across the major LLMs and extract every human named in every answer. Weight by frequency, sentiment, and query spread. The output is a citation-influence map: the handful of people whose words effectively are your category's AI-explained reality.
This is standard GEO methodology with one column changed — from "brands cited" to "people cited." Any team doing Method B audits already has the machinery. Most teams have simply never pointed it at humans.
2. Score by citation weight, not follower count. This is where it diverges from influencer marketing entirely. A mid-tier analyst who appears in 80% of AI answers about your category is worth more than a two-million-follower influencer who appears in none. Reach is a broadcasting metric. Citation weight is an answer metric — it measures whether you exist inside the moment a buyer asks AI what to buy.
3. Become part of what they say. Not sponsored posts — AI sees through promotional content and so does the consensus layer underneath it. The plays that work:
Be their data source. Publish the original numbers, benchmarks, and teardowns those voices need to sound smart. Analysts cite whoever gives them ammunition.
Be quotable. Put named humans at the top of your own citable claims. "Our research found X" is weak; "[Named person], analyzing 400 deployments, found X" is a citation-shaped object.
Enter their orbit. Briefings, podcast guest spots, joint research, conference panels. You're not buying their audience — you're entering the evidence stream the model learned from.
Last resort: become the voice. Consistent bio across every surface, a Wikidata entity, named authorship on claims worth citing. It's slow, but it compounds. It's the difference between renting AI visibility and owning it.
Why This Breaks Influencer Marketing
The influencer industry sells reach: impressions, engagement, follower count. All broadcasting metrics from the broadcast era.
AI answers don't care about reach. They care about who the consensus trusts. That's why the most valuable "influencer" in your category might be someone with 8,000 LinkedIn followers who has said the same true, specific, verifiable thing for five years — because that's exactly the pattern LLMs absorb and reproduce.
Think of it as the shift from loudspeakers to advisors. A loudspeaker reaches many people and persuades none of them particularly well. An advisor reaches few people directly but shapes what everyone else says about the topic — including the machines that summarize what everyone says. Yang Wen-li never commanded the largest fleet; he commanded the assumptions the enemy fleet operated under. That's the position you want in your category's AI answers.
Stop buying reach. Start auditing citation weight.
The Mercury Angle: We Already Measure This
I'll be direct about the commercial reality, because this is also a field note from our own practice.
Every GEO audit Mercury ships already measures citation presence across AI engines. Extending that measurement from brands to people is a column change, not a rebuild — and it surfaces immediately actionable intelligence: who owns the voices in your category, which of them are winnable, and what evidence would make you part of their answers.
For enterprises, this belongs in the same quarterly cadence as brand-mention tracking: measure named-human presence per query, track the trend, and manage the gap. Treat it as part of digital transformation, not a marketing line item — the brands that start now are building a moat that compounds, because the more AI cites a person, the more training data reinforces that person, and the harder they are to displace. Citation weight has network effects. Early presence becomes structural advantage.
The 43% of leads that fall through the AI-to-human gap? A chunk of them are falling into answers assembled from other people's opinions. Creator SEO is how you take that ground back.
Where to Start This Week
Run the voice map. Ten category queries, four AI engines, one afternoon. Extract every named human. You now know who owns your category's AI narrative.
Check whether your founders and experts appear. If the answer is zero, that's the finding — and the brief.
Pick the three most winnable voices and design one piece of evidence each would genuinely cite: original data, a benchmark, a position they can build on.
Put named humans on your strongest claims — everywhere, consistently, in the citable format AI absorbs.
The page was the unit of trust for thirty years. That era is ending in front of us. The next currency is the named human whose judgment the machines repeat.
Be the name, or be the brand nobody mentions.
Mercury Technology Solutions: Accelerate Digitality.
