ChatGPT reached 900 million weekly active users in February 2026, while Google's Gemini app passed 750 million monthly users by the end of 2025. When someone asks these tools about a person or company, they return a one-paragraph summary instead of a page of links. That summary is fast becoming the first impression, and it cannot be pushed to page two. The source material frames this as a reputation task: AI tools repeat what they find online, so fixing those sources changes the answer.
Where AI tools take answers about people
ChatGPT, Perplexity, Gemini, Claude and Google AI Overviews do not offer a profile that an owner can log in and edit. Every answer traces back to material on the open web, with live search results and training data as the two origins. The platforms differ in which sites they trust, and those preferences shift without warning. A Semrush study of more than 230,000 prompts in September 2025 found ChatGPT citing Reddit in about 10% of answers, down from about 60% just weeks earlier, while Forbes, Medium and PR Newswire gained ground.
The practical point stressed in the material is that there is no separate fix for each chatbot. Pages that rank well, get cited often and state clear facts are what AI systems repeat across platforms. Co-founder and CEO of Reputation Resolutions Anthony Will is quoted as saying there is no back door into ChatGPT and every answer traces back to something on the open web. Reputation Resolutions, based in Englewood, Colorado and co-founded in 2013 by Will, reports more than 5,000 clients in over 40 countries and recognition by Forbes and Inc. Magazine.
An AI answer carries more weight than a traditional search page because a single bad link can become the whole story rather than one of ten results. The guide behind the source draws on client work correcting AI statements and recommends starting with a baseline before changing anything. That means running customer-style queries such as "Who is [name]?", "Is [company] legit?", "[Company] reviews", "[Name] lawsuit" and "Best [service] in [city]" across ChatGPT, Perplexity, Gemini, Google AI Overviews and Claude. Each system pulls from different sources, so results rarely match, and the guide advises fresh conversations with date and web-search status recorded.
What this means for companies managing reputation
For businesses that depend on search for leads, hiring or partnerships, the audit routine becomes part of brand control. The described method saves screenshots of exact statements, records every cited link, and opens each source because a cited page can still support a wrong answer. Perplexity numbers its sources, while ChatGPT and Gemini show links when they search the web. Claims are then sorted as accurate, outdated, false or about someone else, with unsourced statements flagged and false allegations and mistaken identity handled first.
Correction works through sources, not through editing the model directly. If a site allows OpenAI's crawler, ChatGPT adds utm_source=chatgpt. com to referral links, which lets analytics show which pages receive ChatGPT traffic. The source cautions to expect weeks to months for improvement rather than days. Answers built on live search usually improve first, while answers drawn from training data catch up as models are retrained, and the same questions are rerun every month to track whether a specific claim still appears.
The marker to watch is whether monthly retests show the disputed claim disappearing across ChatGPT, Perplexity, Gemini and AI Overviews after source pages are corrected or removed. A second signal is referral data with the ChatGPT marker pointing to accurate pages rather than harmful ones. If citation patterns shift again, as in the Reddit to Forbes and PR Newswire move recorded by Semrush, the audit will show it first. That repeated measurement will show whether source work has turned into a stable answer.
