AI Reputation Management: What Does ChatGPT Say About You?
Ask ChatGPT, Gemini or Perplexity “Is [your name or company] trustworthy?” and you will get a confident, two-paragraph answer. Your future clients, investors and hires are already asking. That answer is now part of your reputation, and most people in New York have never read it. This is where AI reputation management comes in: the practice of checking and improving how AI tools describe, compare and recommend you. This guide shows you how to audit what AI says about you, why it says it, and what you can realistically do when the answer is wrong, outdated or unfair. It ends with a 30-day action plan you can start this week. What Is AI Reputation Management? Classical online reputation management deals with the things that can be seen at the top of Google results page: news, reviews, complaint websites, social media profiles. Reputation management includes removal, suppression and creation of positive content. AI reputation management involves additional platforms, AI answer engines like ChatGPT, Google AI Overviews, Gemini, Copilot and Perplexity. Those systems not only give links but sum up the found information into one answer usually in terms of “pros and cons.” What does your AI search reputation consist of? It is the answer that AI gives out. When the AI gives you an answer based on the information found from an outdated complaint or limited sources, you have a problem. There is a lot of intersection between those two types of reputation management. However, there are some differences that are shown in the table below. Traditional online reputation management AI reputation management Where it shows up Google results, review sites, social profiles ChatGPT, Gemini, Perplexity, Copilot, Google AI Overviews What the reader sees A list of links One written summary, sometimes with cited sources What you can change Your own pages; removal requests to sites and platforms The same sources, not the AI’s answer itself How you check it Search your name and brand Ask the questions an AI user would ask, and log the answers How often to check Monthly, or after a news event Monthly, because answers change between runs Why It Matters Now Individuals have increasingly turned to AI applications to do research about a business or professional prior to reaching out. According to the 2025 Attest survey quoted by GoFish Digital, over half of consumers (52%) identified ChatGPT as their go-to AI search application. The analysts have also forecasted a drop in traditional search volumes due to the increased role of AI answers in the process. For a law firm, wealth manager, clinic, or startup located in New York, this alters the first impression in two ways, making AI reputation management a must: You might never get mentioned:If there is insufficient reliable information available on you, an alternative company will be recommended. You might be negatively characterized: It takes only one viral review, a closed case in court or an old article for you to appear in the AI answer. The true AI reputation management consists in knowing both ahead of the prospect. How AI Tools Decide What to Say About You Although no outside authority dictates how it happens and although it will differ from one AI application to another, there are certain features that recur: Third party sources: Newspapers, reviews, directories, and industry publications count for more than your own website. Community conversations: Reddit threads have grown to be very visible on Google and serve as a frequent element in AI-generated responses. For example, according to one study, which quotes Search Engine Journal, Reddit’s share of Google top-three results grew from 8.56% to 10.24% following the May 2026 core update. Consistency: With consistency of your name, services provided, and other details on various sites, the margin of error for the AI becomes lesser. Recency: More recent is better, and therefore, new information can be favored over outdated information. Structure: Properly organized and clearly understandable pages leave fewer chances for errors. The lesson here is simple: the AI does not invent your reputation from scratch. The AI reproduces what your sources say about you. This is why the task of AI reputation management is largely about correcting and improving these sources. What a Weak and a Strong AI Answer Look Like The example below is an illustration only. “Hartwell Legal” is a made-up New York employment-law firm. Weak answer: “The firm Hartwell Legal is a small one in New York. Reviews are not all positive and some customers have pointed out poor communication. No independent information is found.” Better answer: “Hartwell Legal is an employment law firm based in New York. It is included in two directories and is mentioned in the local business press. Most client reviews are positive about responsiveness. Some older ones have pointed out delays.” The firm did not change the AI. It changed the sources: it answered old reviews professionally, completed its directory profiles, and published clear pages about its practice and team. That is the pattern behind most AI reputation management wins. Step 1: Audit What AI Says About You Open ChatGPT, Gemini, Perplexity and Google (find the AI Overviews tab) in a new private window. Start by assessing your ChatGPT Brand Reputation, then your Google AI Overviews Reputation, because each one uses a different source of information. You will use these prompts “Who is [name]?” “Is [company] legit or a scam?” “What are the complaints about [company]?” “Best [your service] in New York” (to see whether you are recommended) “[Company] vs [competitor]” Copy each answer into a spreadsheet and note three things: Accuracy: Is anything false, outdated or mixed up with another person or company? Sentiment: Is the tone positive, neutral or negative? Sources: Which pages does the tool cite? Perplexity and Google usually show links; for ChatGPT you can ask, “What sources support that?” This ensures consistency in the comparison of results from month to month. Below is an initial template with a


