Type your name into ChatGPT. Type your company. Type the CEO you advise.
Type "is X company reputable" — "tell me about the CEO of X" — "best PR firms in AI" — "what happened with X brand controversy."
Whatever comes back — that is your reputation now. Not your website. Not your press page. Not even Google. The answer itself.
The five paragraphs the engine returns when somebody — a reporter, an investor, a board member, a junior analyst at a competitor — asks it a question about you. That is what they read first. And in 2026, more than half the time, that is the only thing they read.
The reputation surface moved.
For two decades, reputation management lived on Google. Own the first page, push down the negatives, place the right earned media, monitor social, hold the line on Wikipedia. Every senior communications operator on the planet ran a version of this playbook.
The playbook still works on Google. Google is no longer where the question gets asked.
Reporters increasingly use Perplexity the way they once used LexisNexis — as a first-pass research tool that decides whether someone is worth deeper reporting. Analysts use ChatGPT to get context before opening a browser tab. Investors use AI engines to frame questions before a meeting.
The answer becomes the briefing memo.
Reputation inside an AI engine breaks into five measurable dimensions.
- Accuracy — Are the factual claims about the subject right? Wrong dates, wrong companies, conflated identities are common.
- Sentiment — Is the overall framing positive, neutral, or negative? Hedged language is its own signal.
- Completeness — What is the engine omitting? What does it not know about you that it should?
- Consistency — Does Claude tell the same story as ChatGPT? When the engines diverge, the divergence is the reputation issue.
- Control — Who is sourcing the answer? Are the sources ones you have any path to influence?
Each dimension is measurable. Each dimension is movable. None of them is what the legacy reputation playbook was built for.
Three patterns across executive reputation audits.
The Wikipedia anchor. Subjects with a clean, current Wikipedia page get coherent answers from every engine. Subjects without one get fragmented answers stitched from a years-old press hit, a stale conference bio, a deal announcement from the wrong company. Wikipedia presence dominates more reputation surface than anything else for senior business figures.
Trade press is undervalued. Industry trades — the ones every senior communicator deprioritized over the past decade because they did not move consumer awareness — turn out to carry significant weight in the AI citation graph. A profile in a strong trade publication is now more retrievable than a piece in a general-audience outlet that did not survive the past three years of media consolidation.
The silence problem. The most consistent reputation issue inside AI engines is not a negative — it is an absence. Executives whose visible record is thin get answers stitched together from peripheral mentions, leading to inaccuracy that the legacy crisis playbook does not address. There is no crisis to respond to. There is only a vacuum the engine is filling on its own.
Why this is harder than Google.
It is invisible. There is no rankings page to audit. No "position #3" to improve. The answer is generated in real time — changing by prompt, by model, and by source set.
It moves faster. Narratives that once took months to settle can now shape in days.
It is fragmented. ChatGPT, Claude, Gemini, Perplexity — each may tell a different version of the story.
And none of them ask permission.
What the discipline looks like now.
Five operating moves, each measured against the five dimensions above:
- Source-layer work — get the right Wikipedia presence, the right primary-source content, the right trade-press footprint, the right LinkedIn surface.
- Co-mention engineering — earn placement next to authority sources the engine already trusts.
- Schema and crawl discipline — every authored piece needs to be retrievable. Sites that look beautiful and crawl badly are reputation liabilities.
- Pre-position before crisis — the engines remember what was published before the news cycle started. Building retrieval surface during a crisis is too late.
- Quarterly measurement — run the five dimensions every quarter across all five engines. Adjust the program against the read, not against the gut.
The first question to ask.
Open ChatGPT. Type a question about yourself. Your company. Your CEO. Your client.
Then read the answer as if you have never seen the name before.
Because for the person asking it — a reporter, investor, recruit, customer, or competitor — that may be the first version of you they encounter.
And increasingly, it may be the only one.
