There’s a specific moment most people remember. You search your own name, or your company’s name, and something is wrong. An old article you’d forgotten about. A review from a client who never became a client. A competitor outranking you for your own brand. Or — increasingly — you ask ChatGPT about yourself and it repeats something inaccurate with total confidence. When people come looking for Cory Maki reputation management guidance, it’s almost always right after that moment. And the first thing worth saying is that the panic response and the correct response are rarely the same thing.
What online reputation management actually is
Online reputation management (ORM) is the practice of shaping what people find when they look you up — across search engines, social platforms, review sites, and now AI answer engines like ChatGPT, Perplexity and Google AI Overviews (the AI-generated summary that sits above traditional blue links).
That’s the textbook version. The working version is narrower and more useful: ORM is the discipline of making sure the most accurate, most useful, most representative information about you is also the easiest information to find.
Notice what that definition does. It doesn’t promise that nothing negative exists. It doesn’t promise a clean slate. It says that among everything published about you, the material that best reflects reality should be the material that surfaces first — and that if such material doesn’t exist yet, the job is to create and earn it.
Over a decade in reputation and search, working with startups, law firms and public figures, I’ve found that most reputation problems are actually absence problems. There isn’t a smear campaign. There’s a vacuum. The person never built a search presence, so search engines filled the space with whatever third parties happened to publish — a directory listing, an old bio, a single news mention, an aggregator page. Nature abhors a vacuum and so does a search results page.
What ORM is not
The category has a credibility problem, and it earned it. So let’s be precise about what falls outside the work.
- It is not buying reviews. Fake reviews are a fraud risk, a platform-ban risk, and — practically speaking — obvious. Real customers write differently than paid ones.
- It is not deleting the internet. With rare exceptions (defamation, doxxing, clear terms-of-service violations, legally protected categories), content you don’t own doesn’t come down because you’d prefer it gone.
- It is not spam suppression. Firing forty thin press releases and doorway pages at a search result used to move things temporarily. Now it mostly builds a low-quality footprint that AI systems read and summarize back to you.
- It is not a one-time project. Search results are a live system. A result you pushed down in March can resurface in September because a news cycle, an algorithm update, or an AI model refresh changed the weighting.
- It is not PR, exactly. PR earns attention. ORM makes sure that attention is discoverable, accurate, and durable in the places people actually check. They’re adjacent, not identical.
The through-line: reputation is earned, not bought. Anything that tries to shortcut the earning either doesn’t hold or actively backfires.
How it works: the mechanism
Search engines and AI systems are both trying to answer a question — who is this, and should I trust them? — using the material available. The difference is what they do with it.
A classic search engine ranks documents. Ten links, you pick one. An AI answer engine synthesizes: it reads across many sources, forms a summary, and cites a handful. That shift matters enormously for reputation, because citations beat rankings in that environment. Ranking eleventh for your name used to mean invisibility. Being one of five sources an AI model draws on to describe you means you’re in the answer itself — whether or not anyone clicks.
A concrete example. Say a founder has a Crunchbase profile, a LinkedIn, one podcast appearance, and a mention in a lawsuit filing that a court-records aggregator republished. Traditional SEO advice: build pages until the aggregator drops to page two. But an AI model summarizing that founder isn’t ranking those sources — it’s reading all of them and weighing which ones are clear, consistent, corroborated and structured enough to rely on. If the founder’s own material is thin and vague while the aggregator page is specific and well-structured, guess which one shapes the summary.
That’s why clarity and structure make content citable. Ambiguity is not neutral. An AI system that can’t confidently parse who you are, what you do and what you’re known for will lean on whatever source states something plainly — even if that source is unflattering or outdated. Much of modern ORM is really Generative Engine Optimization (GEO) applied to identity: making the true version of your story the easiest version to extract, quote and attribute.
The Cory Maki reputation management approach, in practice
This is the part I care about most, because it’s where the work actually happens. At Reputation Pros, where I run fulfillment, the sequence looks roughly like this — and it’s the same sequence I’d recommend to someone doing it themselves.
1. Audit before you act
Search your name and brand in a clean browser. Then ask ChatGPT, Perplexity and Google’s AI Overviews the same questions a prospect would: Who is [name]? Is [company] legitimate? What is [name] known for? Write down what comes back. Most people have never done this, and the gap between what they assume and what appears is usually the entire brief.
2. Build the owned layer first
A properly structured personal or company site is the single highest-leverage asset in ORM. It’s the one source you fully control, and it’s the reference point other systems check against. It should state plainly who you are, what you do, where you’re based, what you’ve published and where you’ve been covered. Even a simple first post establishing a home base is more useful than a beautiful site that says nothing specific.
3. Earn the corroboration layer
One source saying something is a claim. Several independent sources saying the same thing is a fact, as far as a search or AI system is concerned. Legitimate coverage, contributed writing, podcast appearances, conference listings, a book, industry directories — these are the corroboration that makes your owned layer credible. This is the slow part. It’s also the part that can’t be faked.
4. Make every asset citable
Clear headings. Direct answers near the top. Consistent name, title, and location everywhere. Specific, checkable facts rather than adjectives. This is the logic behind the ARC Method and AI-citation frameworks I’ve developed — structuring content so that an AI system can lift a clean, attributable statement without guessing.
5. Handle the negative result correctly
Assess it honestly first: is it accurate? If it’s factually wrong, there may be a correction or removal path — pursue it through the publisher or the platform, properly. If it’s accurate but old, context and volume of newer, better material is the answer. If it’s a review, respond publicly, calmly, once. What I’m seeing across AI search is that a measured response often gets summarized alongside the complaint, which is a far better outcome than silence.
6. Monitor as a system, not an event
Alerts on your name and brand. Quarterly re-runs of the AI prompts from step one. Systems and automation scale quality here — the goal isn’t to check obsessively, it’s to notice drift early, when it’s cheap to correct. If you’re curious how this extends to AI reputation management specifically, the monitoring layer is where most of the new work lives.
The durable principle
Platforms change. Google ships an update, a new answer engine gets traction, a review site changes its display logic. What doesn’t change is this: the systems that describe you are trying to find the most reliable account of who you are. Online reputation management is the work of making sure that account exists, is accurate, is corroborated by people other than you, and is structured clearly enough to be found and quoted.
Everything else — the scrubbing, the suppression tricks, the review farms — is an attempt to avoid that work. It doesn’t hold. Show up where the answers are formed, say true things clearly, and let the evidence accumulate. That’s not a tactic. That’s just the job.