AI Reputation: How to Control What ChatGPT Says About You
Two years ago, when an investor wanted to learn about you, they opened Google. Today, more and more often, they open ChatGPT and ask directly: "What do you know about [your name]?" And the AI answers. With data. With opinions. With a summary that might be completely wrong but sounds absolutely convincing.
This isn't science fiction. 65% of professionals already use artificial intelligence to research people and companies before making decisions. And unlike Google, where you can see the results and evaluate the sources, AI answers are presented as absolute truths. No visible links. No clear source. No chance to respond right away.
In this guide, I'll explain exactly how generative AI works when it talks about people and companies, where it gets its information, how you can influence what it says and what to do if it's saying something inaccurate or damaging about you.
1. The new reputation battleground
For 20 years, digital reputation was fought on a single field: Google's search results. If you owned the first page, you owned the narrative. That model still matters, but it's no longer enough.
Generative AI has created a second battleground that runs on completely different rules. And most companies and professionals don't even know this battleground exists.
Reputation on Google (traditional)
The user searches, sees 10 results, chooses which ones to click and forms their own opinion by weighing multiple sources. You can influence which results appear and in what order.
Reputation in AI (new)
The user asks and gets ONE synthesized answer. They don't see the original sources. They don't evaluate. They accept what the AI says as the truth. You can't directly control the answer, but you can influence the sources the AI consults.
The critical problem with AI and reputation
Generative AI can make things up (so-called "hallucinations"). It can mix up data from people with the same name. It can give disproportionate weight to an old negative source. And worst of all: the average user has no way of knowing whether the answer is right or not. The AI says it with total confidence, in a neutral, authoritative tone that creates instant credibility.
2. How AI tools build answers about you
To influence what AI tools say, you first need to understand how they build their answers. The process is fundamentally different from a search engine's.
2.1. The answer generation process
- Training on massive datasets
AI models like ChatGPT are trained on enormous amounts of text from the internet: web pages, articles, books, social media, forums. Anything that's been published online can become part of their "base knowledge." This training has a cutoff date: the AI knows nothing after that date. - Real-time search (models with browsing)
The latest versions of ChatGPT, Perplexity and Gemini can search the internet in real time. When you ask them about someone, they run searches, read the pages they find and synthesize the information. This is where your SEO remains critical. - Synthesis and text generation
The AI combines what it knows from training with what it finds in real time, applies a probabilistic model and generates an answer. It doesn't "understand" or "have opinions": it predicts which sequence of words is most likely given the information available. - Safety filters
Before the answer is shown, filters are applied to try to prevent harmful, defamatory or false content. But these filters aren't perfect, and damaging information can slip through if it's present in sources the AI considers "reliable."
Key point: AI doesn't verify the truth
Generative AI doesn't distinguish between true and false information. It distinguishes between frequent, consistent information and rare or contradictory information. If a false fact about you appears in many sources, the AI will treat it as true. If a true fact about you only appears in one obscure source, the AI may ignore it or downplay it.
2.2. How much weight each source carries
Not all sources influence AI answers equally. This is the hierarchy we've observed:
| Source | Weight in the answer | Why |
|---|---|---|
| Wikipedia | Very high | AI models treat Wikipedia as a primary reference source. If you have an article, it dominates the narrative. |
| News media (El País, BBC, Forbes...) | High | Major media outlets are considered authoritative sources by language models. |
| High | For professional information, LinkedIn is a highly valued primary source. | |
| Your own website | Medium-high | Depends on the domain's authority and the quality of the content. |
| Academic publications | High | Papers, books and academic publications carry significant weight in most AI models. |
| Social media | Medium | Twitter, Instagram, YouTube: they have influence, but less weight than the media. |
| Forums and smaller blogs | Low | Less weight individually, but if a fact shows up in many forums, it adds up. |
| Structured data (Schema.org) | High (growing) | AI models are learning to read and prioritize structured data. It's the future. |
3. Where AI tools get their information
Understanding the specific sources lets you focus your strategy with precision. Not every AI uses the same sources, or uses them in the same way.
3.1. Training sources (historical data)
Models are trained on massive datasets that include:
Common Crawl
A massive crawl of billions of web pages. It's the foundation of training for most models. Anything published on the internet and not blocked by robots.txt may be in here.
Wikipedia and Wikidata
The primary source of "facts" for AI. Wikipedia data carries disproportionate weight in answers. If Wikipedia says something about you, the AI will almost certainly repeat it.
Books and publications
Books3, OpenWebText and other datasets of books and academic publications. If you've published a book or academic papers, the AI may know your work.
Reddit, StackOverflow, forums
Platforms with user-generated content. Reddit in particular carries enormous weight because many models were trained on its data. If there are threads about you on Reddit, the AI knows about them.
3.2. Real-time sources (web search)
Modern versions of AI tools search the internet when they need up-to-date information. This is critical because it means your current SEO strategy directly affects what AI says about you.
What AI finds in real time depends on your SEO
When ChatGPT or Perplexity look up information about you, they use mechanisms similar to Google's. They read the top-ranking results, synthesize them and generate their answer. If the top Google results about you are negative, the AI will work them into its answer. That's why a reverse SEO strategy doesn't just protect your reputation on Google: it protects your reputation across every AI tool at the same time.
4. ChatGPT vs Perplexity vs Gemini vs Claude
Each AI has its own quirks. Knowing them lets you tailor your strategy.
| AI | Web search | Main sources | Transparency | Reputational risk |
|---|---|---|---|---|
| ChatGPT (OpenAI) | Yes (with browsing) | Training + Bing | Low (doesn't always cite sources) | High (200M+ users) |
| Perplexity | Always | Real-time web search | High (cites sources) | Medium-high (growing fast) |
| Gemini (Google) | Yes (integrated with Google) | Google Search + training | Medium (partially cites) | High (built into Google) |
| Claude (Anthropic) | Yes (with tools) | Training + search | Medium-high | Medium (professional use) |
| Copilot (Microsoft) | Yes (Bing) | Bing + training | High (cites sources) | High (built into Windows/Office) |
Perplexity: the most dangerous for your reputation
Perplexity is especially relevant because it always searches the internet in real time and presents its answers with citations and links. That means if negative content is ranking well on Google, Perplexity will find it and cite it directly. Unlike ChatGPT, which can "dilute" negative information with its training knowledge, Perplexity depends 100% on what it finds on the web right now.
5. Audit: what AI tools say about you right now
Before you do anything, you need to know exactly what AI tools are saying about you. This audit is the first step in any AI reputation strategy.
5.1. How to run your AI reputation audit
- Ask ChatGPT
Open ChatGPT and ask: "What do you know about [your full name]?" Then ask: "What do you think of [your company]?" Then: "Would you recommend hiring [your name] for [your service]?" Save the answers. - Ask Perplexity
The same questions in Perplexity. Pay special attention to the sources it cites. If it cites negative content, those are the URLs you need to push down. - Ask Gemini
Gemini is integrated with Google Search, so its answers fairly directly reflect what's on Google. Compare them with your search results. - Ask Claude
Claude tends to be more cautious and to acknowledge when it doesn't have enough information. If Claude says something negative about you, the source is probably very prominent. - Document and compare
Build a table with each AI's answers. Look for patterns: if they all say the same thing, the source is strong. If they differ, there's room to influence them.
5.2. Audit template
| Question | ChatGPT | Perplexity | Gemini | Claude |
|---|---|---|---|---|
| "What do you know about [name]?" | Answer... | Answer + sources | Answer... | Answer... |
| "Is [company] reliable?" | Answer... | Answer + sources | Answer... | Answer... |
| "Any controversy about [name]?" | Answer... | Answer + sources | Answer... | Answer... |
| "Would you recommend [name] for [service]?" | Answer... | Answer + sources | Answer... | Answer... |
Heads up: the answers change
Generative AI doesn't always give the same answer to the same question. Ask several times, on different days and in new conversations. One day it may be neutral, and the next it may mention something negative. Document the worst answer you get: that's the one a prospective client could receive.
6. How to influence what AI tools say about you
You can't directly edit ChatGPT's or Perplexity's answers. But you can influence the sources they use to build those answers. Here's the complete strategy.
6.1. The 5 levers of influence
1. Own the authoritative sources
If Wikipedia, LinkedIn, news media and your personal website say positive, consistent things about you, AI tools will reflect that. These sources carry the most weight in generated answers.
2. Structured data (Schema.org)
Implement Person, Organization and Article markup on your website. Structured data is the most direct way to "tell" AI tools who you are, what you do and what your credentials are.
3. Consistent information
If your bio, title, career history and achievements are consistent across every online source, AI tools will be confident presenting that information. If there are inconsistencies, the AI gets confused and may make things up.
4. Volume of positive content
The more positive, reliable sources exist about you, the less relative weight a single negative source carries. It's the same logic as reverse SEO, applied to the world of AI.
5. Search engine SEO
AI tools with real-time search (Perplexity, ChatGPT with browsing, Gemini) use search results. If you own Google, you own what these AI tools find and cite.
6.2. Wikipedia: the source that matters most
If you're notable enough to have a Wikipedia article, it will probably be the main source AI tools use to talk about you. If you don't have one, AI tools will turn to other, less controllable sources.
If you already have a Wikipedia article
Review it and make sure the information is accurate and up to date. Don't edit your own article (it violates Wikipedia's rules and may be reverted). Find an experienced editor who can update the information with verifiable sources. Every fact must be backed by a reliable external source.
If you don't have a Wikipedia article
Assess whether you meet the notability criteria (significant coverage in independent media). If you do, a Wikipedia editor can create the article. If you don't, focus on generating that media coverage first. Don't try to create your own article: it will be deleted quickly.
6.3. News media: the amplifier
Media mentions have a double effect: they rank well on Google (influencing AI tools with web search) and they become part of the models' training data.
How to earn strategic media coverage
You don't need to appear in El País or Forbes. Trade media, industry podcasts and specialized publications count too. The key is for the mentions to be substantive (not just a passing mention of your name, but content that describes your experience, achievements and perspective) and published on authoritative domains.
Do you know what ChatGPT says about you?
We run a complete audit of your reputation across every generative AI tool and present you with a specific action plan. A free, confidential assessment in 48 hours.
Request an AI reputation audit7. Schema.org and structured data: your identity for AI
Structured data is probably the most underused tool in AI reputation management. It's invisible markup you add to your website that tells Google and AI tools exactly who you are, what you do and what your credentials are.
7.1. Essential Schema types for reputation
Person Schema
Defines your name, title, organization, education, awards, social profiles and professional relationships. It's your "ID card" for machines.
Organization Schema
Defines your company: name, founder, industry, address, contact details, logo, social profiles. Link the Organization to your Person.
Article Schema
Every article you publish should have Article markup with the author, date and description. AI tools use this to attribute authorship and credibility.
Review and AggregateRating Schema
If you have verifiable positive reviews, rating markup can influence how AI tools perceive the quality of your service.
7.2. Example Person implementation
What to include in your Person Schema
The most important fields are: name (full name), jobTitle (current title), worksFor (company, linked to its Organization Schema), sameAs (an array with the URLs of all your profiles: LinkedIn, Twitter, website, etc.), alumniOf (education), award (awards or honors) and knowsAbout (areas of expertise). The more fields you fill in with accurate information, the more complete the picture AI tools will have of you.
7.3. SameAs: connecting your digital identity
The sameAs field is especially important. It tells AI tools that your LinkedIn profile, personal website, YouTube channel and Twitter account all belong to the same person. This strengthens your digital identity and reduces the chances of being confused with someone who shares your name.
- Your LinkedIn profile URL
- Your personal website URL
- Your Twitter/X profile URL
- Your YouTube channel URL
- Your Instagram profile URL
- Your Wikipedia page URL (if one exists)
- Your profile URL on Crunchbase, About.me or other directories
Structured data is the bridge between your web presence and AI. Without it, AI tools have to "guess" who you are. With it, you're telling them directly.
8. How it differs from traditional SEO
SEO for AI (sometimes called GEO, Generative Engine Optimization, or AEO, Answer Engine Optimization) shares principles with classic SEO, but there are fundamental differences you need to understand.
| Aspect | Traditional SEO (Google) | SEO for AI (GEO/AEO) |
|---|---|---|
| Goal | Rank in the top positions | Be the source the AI synthesizes |
| Keywords | Optimize for specific searches | Optimize for natural-language questions |
| Content | Focused on relevance and backlinks | Focused on authority, consistency and clarity |
| Format | HTML structure matters (H1, H2, meta...) | Clear statements and structured data matter |
| Result | 10 organic links on page 1 | 1 synthesized answer with no visible links |
| Speed of change | Days to weeks | Can take months (depending on the model) |
| Measurement | Rankings, CTR, traffic | The AI's answers (no standard metrics) |
8.1. What works in traditional SEO and in AI
Long-form, high-quality content
Long, well-structured articles with valuable information work for both Google and AI. Google ranks them; AI tools use them as sources.
Authoritative backlinks
Links from high-authority sites strengthen your domain on Google AND your credibility as a source for AI tools.
Presence across multiple reliable sources
Appearing in recognized media, directories and platforms has a positive impact on both ecosystems.
8.2. What's different with AI
Consistency matters more
On Google, having contradictory information on different pages isn't a big problem. For AI tools, it is: if your data is inconsistent, the AI can get confused and generate "hallucinations" about you.
Wikipedia carries disproportionate weight
On Google, Wikipedia is just one more result. For AI tools, Wikipedia is one of the main sources of "truth." A Wikipedia article about you has more impact on what AI says than 50 backlinks.
No control over the "snippet"
On Google, you can optimize the title and meta description. In AI tools, the answer is generated from scratch every time. There's no "snippet" to optimize: you influence the sources, not the presentation.
9. Monitoring mentions in AI tools
Unlike Google, where you can use automated tools to track rankings and mentions, monitoring AI tools requires a more manual approach. But it's absolutely necessary.
9.1. Monitoring protocol
| Frequency | Action | Tool |
|---|---|---|
| Weekly | Ask ChatGPT and Perplexity about your name/company | ChatGPT + Perplexity (manual) |
| Every two weeks | Ask Gemini and Claude | Gemini + Claude (manual) |
| Monthly | Full audit with every question from the template | All AI tools + tracking document |
| Quarterly | Review structured data and information consistency | Google Rich Results Test + manual review |
9.2. Key questions to monitor
Don't just ask "What do you know about me?" Use these questions, which simulate what a prospective client or investor might ask:
Informational questions
"Who is [name]?" / "What does the company [name] do?" / "What industry does [name] work in?" / "What is [name]'s professional background?"
Evaluative questions
"Is [name/company] reliable?" / "Would you recommend hiring [name]?" / "What do people say about [company]?" / "What is [name]'s reputation?"
Risk questions
"Is there any controversy about [name]?" / "Any complaints about [company]?" / "Is [company] a scam?" / "Are there any lawsuits against [name]?"
Watch out for "trick questions"
Generative AI is suggestible. If you ask "What are the complaints against [your name]?", the AI may invent complaints that don't exist just to satisfy the question. That's a problem because your competitors or detractors can use this technique to generate negative content about you. Monitor regularly with risk questions to catch this kind of hallucination.
10. Mistakes that destroy your AI reputation
Over the past two years, I've seen many professionals and companies make mistakes that have cost them dearly. These are the most dangerous ones.
10.1. Ignoring the problem
The most common and most damaging mistake
Most professionals haven't even checked what AI tools say about them. They assume that if Google is clean, everything's fine. But AI models have their own training datasets and may contain information that's no longer on Google. I've seen cases where a news story removed from Google years ago still shows up in ChatGPT's answers because it's part of the training data.
10.2. Inconsistent information across the web
Different titles on every platform
If you're "CEO" on LinkedIn, "Founder" on your website and "Managing Director" in a media article, the AI doesn't know which one is right and may invent or mix up titles.
Dates that don't add up
If your LinkedIn says you founded the company in 2015 but your website says 2016, the AI picks up on the inconsistency and may question your overall credibility.
A different bio on every site
If your bio tells a different story on each platform, the AI has no clear narrative to synthesize and may produce inaccurate or confusing summaries.
No structured data
Without Schema.org, you force the AI to "guess" your information by reading unstructured text. With structured data, you hand it over on a silver platter.
10.3. Trying to "hack" AI tools
Don't try to manipulate the answers with tricks
There are techniques like "prompt injection" (embedding hidden instructions on your website for AI tools to read and follow). Don't use them. Models are constantly updated to detect and penalize these practices. If you're caught, your content may be flagged as unreliable and excluded from answers, making your situation worse. The right strategy is to create genuinely good content, not to try to trick the machines.
11. An integrated strategy: Google + AI
In 2026, digital reputation isn't managed only on Google or only in AI. You need a strategy that covers both fronts at the same time. The good news is that many actions have an impact on both ecosystems. Our service to audit what language models like ChatGPT say about you is designed precisely to manage both fronts in a coordinated way, instead of treating them as separate projects.
11.1. Integrated action plan
- Dual audit
Check what Google says AND what AI tools say. Document both. Identify discrepancies: content that appears in AI but not on Google (old training data) or content that appears on Google but not in AI (very recent content). - Consistent information
Unify your bio, title, career history and details across ALL platforms. This directly affects both ecosystems. - Structured data
Implement Schema.org Person and Organization on your website. This benefits your SEO on Google (rich snippets) and your reputation in AI (clear, structured data). - Authoritative content
Publish long-form articles on your website, on Medium and in trade media. Quality content ranks on Google AND feeds AI answers. - Media coverage
Pursue interviews, media mentions and podcast appearances. The media are high-authority sources for Google AND for AI. - Ongoing monitoring of both
Check Google weekly (automated tools) and AI tools weekly (manually). Act immediately on any negative change.
11.2. Action impact matrix
| Action | Impact on Google | Impact on AI | Priority |
|---|---|---|---|
| Optimize LinkedIn | High | High | Critical |
| Implement Schema.org | Medium | Very high | Critical |
| Build a personal website with a blog | High | High | Critical |
| Publish in the media | High | Very high | High |
| Wikipedia (if applicable) | Medium | Very high | High |
| Reverse SEO | Very high | High (AI with web search) | High |
| YouTube and video | High | Medium | Medium |
| Data consistency | Low | Very high | Critical |
The highest-ROI actions
The three actions with the highest return are: 1) Unify your information and make it consistent across every platform (huge impact on AI, zero cost). 2) Implement Schema.org structured data on your website (huge impact on AI, low cost, a few hours of work). 3) Optimize your LinkedIn to the max (high impact on both ecosystems, zero cost).
12. The future of reputation in the AI era
The digital reputation landscape is changing at unprecedented speed. These are the trends we're seeing that will define the coming years.
12.1. Trends that are already here
Conversational search as the norm
More and more people use AI instead of Google to look up information about people and companies. By 2027, an estimated 50% of informational searches will be done through conversational AI.
AI built into everything
Microsoft Copilot in Office, Gemini in Gmail, Claude in business tools. AI isn't a separate app: it's built into the tools people use every day. Your AI reputation follows you everywhere.
AI agents that make decisions
AI agents that research vendors, evaluate candidates and prepare reports are already here. An AI agent preparing a due diligence report on you will automatically consult multiple AI tools and sources. Your digital reputation is being evaluated by machines.
Emerging regulation
The EU is developing legislation on the right to an explanation and the right to rectification in AI. In the future, you'll be able to demand that an AI correct false information about you. But that legislation isn't mature yet.
12.2. What's coming in the next 2-3 years
12.3. How to prepare today
- Don't wait until it's urgent — AI training data accumulates. The sooner you start building a solid digital presence, the easier it will be to maintain
- Invest in structured data — Schema.org is going to become increasingly important as AI evolves
- Prioritize consistency — Unify your information across every platform. It's free and has an enormous impact
- Generate media coverage — Media mentions are the longest-lasting, highest-impact sources on both Google and AI
- Monitor regularly — AI answers change with model updates. What's fine today may change tomorrow
- Work with professionals who understand both ecosystems — Most SEO and reputation agencies only work with Google. You need someone who understands the whole ecosystem
AI reputation isn't a problem for the future. It's a problem right now. Every day that goes by without a strategy, AI tools are forming opinions about you based on what they find, not on what you want them to say.
13. Frequently asked questions
Not directly. ChatGPT doesn't have a mechanism like Google's right to be forgotten. OpenAI has a form for reporting factually incorrect information, but the process is slow and results aren't guaranteed. The most effective strategy is to influence the sources ChatGPT uses to generate its answers: if web sources say positive, consistent things about you, ChatGPT will reflect that.
Yes. Generative AI suffers from "hallucinations": it generates information that sounds plausible but is completely false. It can attribute titles you've never held, companies you've never worked for or even controversies that don't exist. This is one of the biggest risks to your AI reputation and the main reason to regularly monitor what these tools say about you.
Partially. AI tools with real-time web search (Perplexity, ChatGPT with browsing, Gemini) will reflect changes on Google relatively quickly. But AI that relies on training data may keep repeating old information for months until the model is updated. That's why cleaning up Google is necessary but not sufficient: you need a dedicated AI strategy running in parallel.
Yes, and increasingly so. Modern AI models are designed to read and prioritize structured data because it's more reliable than free text. A well-implemented Person Schema gives the AI accurate, structured information that reduces the likelihood of hallucinations and improves the accuracy of answers about you. It's currently one of the highest-return investments in digital reputation.
It depends on your audience. By volume, ChatGPT (200M+ weekly users) is the most relevant. In terms of specific risk, Perplexity is more dangerous because it always searches in real time and presents sources with citations, giving them instant credibility. Gemini is especially important because it's built into the Google ecosystem (Gmail, Docs, Search). The right answer is to monitor all of them.
You can handle the basic actions (audit, information consistency, structured data) yourself at minimal cost. A comprehensive professional service covering both Google and AI ranges from €2,000 to €6,000 per month, depending on the complexity of the case. Since the same actions affect both ecosystems, the incremental cost of adding AI management to an existing Google reputation strategy is relatively low.
Legislation is still being developed. The EU AI Act sets transparency obligations for high-risk AI systems. The GDPR already allows you to request the rectification of inaccurate personal data, and some legal scholars argue this applies to AI. In practice, OpenAI, Google and Anthropic have forms for reporting incorrect information, although the process is slow. Specific EU legislation on the right to rectification in AI is expected within the next 2-3 years.
Operating in your city
We manage the Google and AI reputation of clients in Bilbao, Bogotá, Buenos Aires and Cancún, along with dozens of other cities in Spain and Latin America.
Your AI reputation is being shaped right now. Are you in control?
Generative AI is already talking about you. The question is whether what it says is what you want it to say. Request a complete audit of your reputation on Google and across every relevant AI tool. Free, confidential and delivered in 48 hours.
I want to control my AI reputationIf you have questions or want to discuss your specific case, message me directly on LinkedIn. This is a new field and the rules are changing fast: the sooner you act, the better positioned you'll be.
Last updated: April 2026