Online reputation management: what ChatGPT says about your brand

Online reputation management is the work of monitoring and shaping what people find when they look up your business: search results, reviews, social posts, press coverage, and now the answers AI assistants give about you. In Pew Research Center's February 2026 survey, roughly half of U.S. adults reported using AI chatbots, and 42% said they turn to them to look up information. AI reputation management adds one job to the old routine: reading what ChatGPT, Gemini, Perplexity and Google's AI features say about your brand, and in what tone.

Hypervisible, the software we make at NEXUS GEO, was built for that new reading: it records what five AI engines say about a brand and in which tone, on the same questions every week, so a shift in reputation shows up as a number you can follow.

What online reputation management means in 2026

Online reputation management (ORM) comes down to keeping watch on what's said about your business and replying where a reply is possible. When something is wrong, you get it fixed where it was published. What has changed is where people look. Reviews and search results still count, and AI answers now sit on top of both.

Plenty of business owners look it up: there are about 4,400 U.S. searches a month for “online reputation management” (Google Ads estimate, October 2026). The term also carries some baggage. Some firms sell it as a way to bury bad search results or to flood a profile with fake five-star reviews, and the second practice is now banned by a federal rule.

In August 2024, the Federal Trade Commission announced a final rule against fake reviews and testimonials. It bans writing, buying or selling fake reviews, AI-generated ones included, paying for reviews that express a particular sentiment, posting insider reviews from employees or managers without disclosure, and suppressing negative reviews through unfounded legal threats or intimidation. The FTC can seek civil penalties against knowing violators. The legitimate version of the job is slower and duller: monitoring, replies, corrections, and reviews from people who actually bought from you.

Reputation and awareness are easy to mix up. Awareness is whether people know your name, and reputation is what they think of it. A regional HVAC company can be unknown outside three counties and still have an excellent reputation inside them. Our guide on how to measure brand awareness, AI answers included covers the first half of that pair.

In 2026, a full ORM routine watches five places:

search results for your name, including Google's own panels and AI Overviews;
reviews on Google, Yelp, industry sites and marketplaces, plus employer reviews on sites such as Glassdoor;
social media and forums, Reddit included;
press and comparison articles, the “best X in Y” lists that buyers and AI engines both read;
AI answers: ChatGPT, Gemini and Perplexity, plus Google's AI Overviews and AI Mode.

AI answers as the new first impression

When someone asks ChatGPT whether a contractor is reliable, they get a short verdict instead of a page of links to sort through. That paragraph becomes the first impression, and nothing forces the reader to look any further.

The audience is large. Pew's February 2026 survey of 5,119 U.S. adults found that 44% say they use ChatGPT, against 18% in 2023 (Pew notes that it changed the question's wording in 2026). On Google, the written answer comes before the links too: AI Overviews appear above the results on many searches, and AI Mode, a conversational version of Search, has been open to everyone in the U.S. since May 2025. Our Google AI Mode guide covers how it chooses the pages it links.

What an AI writes about you comes from what it has read. A single answer can blend pages from your website, directory listings, reviews, news articles, Reddit threads and whatever the model learned during training, which may be out of date. How ChatGPT chooses its sources goes into the mechanics. The portrait that comes out can be accurate, flattering, unfair, or simply stale.

Picture a three-attorney estate planning firm in Denver. Its founding partner retired in 2023, and the firm's website says so. But a 2019 profile in a local business journal still calls him the managing partner, and a 2020 Yelp review complains about slow callbacks. Ask an AI about the firm and you might read that it is “led by” the retired partner and “sometimes slow to respond.” No review dashboard will ever show the firm that sentence.

Reviews feed into these answers. OpenAI's help page on shopping in ChatGPT says it may weigh reviews when it picks products to show, and that it can display review summaries written by the model from reviews on public websites, which OpenAI does not verify. Google's documentation for site owners says a Business Profile can help a company's products and services appear in its AI responses, and in its other search results too.

Being named and being linked are two different things. An AI can describe your company in detail without linking to your site, or cite your site without naming you in the text. Brand mentions in AI answers covers how to find both and what each one is worth.

How to check what AI engines say about you

Ask the AI engines your customers use the questions those customers ask, twice: once with your brand name and once without. Then write down the facts, the tone and the sources of each answer.

Start with the classic checks, because AI answers draw on the same material. Search your business name on Google, then add “reviews,” “complaints” and “scam.” Look at your Business Profile and at the knowledge panel, if Google shows one for you. Read the newest reviews on every site where you have a listing. Google Alerts is free: set an alert on your name and choose how often new pages reach your inbox (as they appear, once a day at most, or once a week at most).

The AI engines come next:

Pick the engines. A sensible core is ChatGPT plus AI Overviews and AI Mode on Google, then Gemini and Perplexity. Add Copilot or Claude if your customers use them at work.
Write branded prompts, the kind a customer types once they have heard of you: “[brand] reviews,” “is [brand] legit,” “[brand] vs [competitor].”
Write unbranded prompts, the kind a stranger types: “estate planning attorney in Denver for a blended family,” “roof replacement company in Columbus that handles insurance claims.”
For each answer, record whether it names you and in what position, every fact stated about you, the tone (positive, neutral, mixed, negative) with the sentence that shows it, and the sites cited.
Ask the same prompts again a few days later. The same prompt can come back different in a new session, and a single bad answer may just be a bad draw.

Branded prompts show what the AI knows about you. Unbranded ones show whether it thinks of you at all when a stranger describes a need, and strangers are where new customers come from.

The first pass can also be done for you. Hypervisible's free AI visibility analysis puts 10 questions to five AI engines, half about your industry without your name and half about your brand by name, then reads the 50 replies in 5 to 10 minutes. You won't have to type in a card number, and the results stay saved to your account. For a fuller method, with competitors and sources, see our guide to running an AI visibility audit.

When an AI gets your brand wrong

Save the exact answer with the prompt, the engine and the date, find the page the error comes from, fix it there, then report the answer to the AI's maker. In our view, the fix at the source does more than the report, because the AI will read that page again the next time it searches.

The AI companies say plainly that errors happen. OpenAI's privacy policy tells users not to rely on the factual accuracy of its models' output, because the most likely next word is not always the most accurate one. Google's help page on AI Overviews warns that they “can and will make mistakes.”

For a small or mid-size business, the errors are often mundane: an old address, a discontinued product, a former CEO, last year's prices, or a mix-up with a similarly named company in another state.

Reporting the answer

ChatGPT. Click the thumbs-down icon under the response and select an issue. For content that may break the law, such as a defamatory claim, OpenAI's help center points to the “Safety or Legal concern” option and to its content reporting form. It adds that reported domains and content may be reviewed by its Model Quality team, which may apply filters so that ChatGPT relies less on unreliable sources.
Google AI Overviews. Click the thumbs-down icon under the overview, choose “Report a problem” and pick a category. Your latest search and its results are sent with the report.
Google AI Mode. Same idea: the thumbs-down icon sits at the bottom of each response, then you choose a category and submit.
Google knowledge panel. If the wrong fact comes from that panel, use “Suggest edits,” then “Edit” next to the item, and include public links that show the right information. Google compares your feedback with other public information on the web and replies by email.

When the error is about a person, say the founder, OpenAI's privacy policy offers another route: a correction or removal request through privacy.openai.com. OpenAI says it will consider the request based on applicable law and the technical capabilities of its models, which is not a commitment to change anything.

Fixing the source

Find the page behind the error. The links shown by Perplexity, by an AI Overview, or by ChatGPT when it searches the web usually point to it. Then work outward from your own properties:

Update your website: about page, locations, hours, pricing, leadership, each with a visible date.
Correct your Google Business Profile and the directories that copy your details.
Ask third-party sites (a directory, a news outlet, a comparison article) to fix the error, in a short note with a link to the correct information.
Report fake or abusive reviews from your Business Profile. Google removes only reviews that violate its policies, asks owners not to report a review just because they disagree with it, and says the evaluation typically takes several days.

Tracking tone and accuracy over time

Ask the same questions on the same engines on a fixed schedule, and log each time whether the answer is positive, neutral or negative, which facts it states, and which sites it cites. A single screenshot is an anecdote, and you need a run of readings before you can speak of a trend.

The reason is variance: ask the same prompt twice and the brands and sources can change, and so can the tone. Only repetition on identical questions separates a real shift from noise.

Five things are worth logging, week after week or month after month:

Mention rate, the share of answers that name you.
Tone, answer by answer, and the adjectives that keep coming back.
Accuracy: each fact the AI states about you, marked right or wrong, so you notice if a fixed error comes back.
The sites cited, which change when a new article or a new list comes out.
How the same prompts describe your competitors, so you know whether a caveat targets you or the whole category. AI share of voice gives the formula for comparing.

Reading the trend takes some judgment. A caveat that shows up three weeks running, right after a news story, deserves a response. A one-week dip on a single engine, with no new source behind it, is usually noise. After a correction, give it several weeks before you judge the result: AI engines that search the web reread pages on their own schedule.

Take a Shopify skincare brand that shipped late during the 2025 holiday rush. The complaints piled up on review sites in December. By spring, deliveries are back on time and new reviews say so, yet AI answers may keep mentioning “shipping delays” as long as the older reviews dominate the pages the engines read. Replying to those reviews, publishing a clear shipping policy page and asking recent buyers for reviews gives the engines fresher material. The next readings tell you whether the phrase is gone.

Daily tracking earns its cost during a crisis or a launch, when a change has to show up within 24 hours. Most brands can follow tone and accuracy with weekly readings on a fixed set of prompts, and report on them once a month.

Tools for brand monitoring in AI answers

Brand monitoring in AI answers takes its own kind of tool: alert tools cover the web, review platforms cover reviews, and neither reads what ChatGPT or Gemini says. Most brands end up with one of each kind, and the first two kinds can be free. We make Hypervisible: it takes the first line of this table, and the next section lays out the facts behind that choice.

Features and prices as shown on vendor websites, checked October 22, 2026; SE Visible, Similarweb, AthenaHQ and Scrunch read again October 29, 2026.
Tool	What it monitors	AI engines	Frequency	Entry price
Hypervisible	How often AIs cite you, your position, your rank against rivals, the sites cited, what AIs say about you and in which tone	AI Overviews, AI Mode, ChatGPT, Gemini, Perplexity	Every Monday	Solo at €49/mo for 25 questions, or €29/mo for the first 15 accounts
Google Alerts	New web pages that mention a name or term	None	As it happens, daily or weekly	Free
Birdeye	Reviews, listings and social, plus an “AI Search” product	Not stated	Not stated	On request, by number of locations
SE Visible (SE Ranking)	Brand visibility on tracked prompts, in seven countries	ChatGPT, Gemini, AI Overviews, AI Mode	Daily	Basic, €80/mo for 75 prompts
Similarweb AEO Intelligence	Visibility on tracked prompts, plus visits from AI platforms (AI Traffic)	ChatGPT, AI Mode, Gemini, Perplexity	Daily	$129/mo, or $99/mo billed annually, for 150 prompts
AthenaHQ	Brand visibility in AI answers, counted in credits (one per AI response)	Starter: 11, Claude included	Not stated	Starter $295/mo
Scrunch	Presence, position and cited sources, plus the AI bots that visit your site	ChatGPT, Claude, Gemini, Perplexity, AI Overviews, AI Mode, Meta	Not stated	$300/mo, or $250/mo billed annually

Google Alerts and your Google Business Profile cost nothing and cover the basics: new pages about you, and reviews on Google, which you can answer and report from the profile. Review management platforms such as Birdeye bring reviews, listings and social messages from many sites into one place, which suits multi-location businesses. Birdeye's site now lists an “AI Search” product as well, but its pages don't say which AI engines it covers, and pricing comes after a form.

AI visibility tools are the newer category, and they differ in price, in engines and in how they count usage. SE Visible and Similarweb check daily on four engines each, with Perplexity on Similarweb and not on SE Visible. AthenaHQ sells credits, one per AI response, across 11 engines on its Starter plan. Scrunch reads seven engines on every plan and adds a view of the AI bots that read your pages. For reputation work, the column that counts is tone: ask each vendor whether the plan you'd buy records how the answers describe you, on top of whether they name you, and get the reply in writing. Our comparison of AI visibility tools sets them side by side on price, engines, frequency and markets.

Hypervisible: tracking the tone of AI answers about you

Hypervisible reads what AI engines say about your brand and how they say it, the reputation question this guide keeps returning to. We make it, so here are the facts behind our recommendation.

Tone, answer by answer. Each reading notes whether you appear and how high, your standing against rivals on identical answers, the domains cited, and the adjectives and tone the engines use for you. On paid plans the questions stay fixed and come back every Monday, and the change shows in points.
Both Google AI features from the entry plan. AI Overviews and AI Mode come with ChatGPT, Gemini and Perplexity on every plan, starting with Solo at €29 monthly for the first 15 accounts and €49 after that (pricing). SE Visible Basic, at €80, reads four engines without Perplexity, and Similarweb's $129 plan lists AI Mode without AI Overviews (prices read October 29, 2026).
Built by an agency that only does GEO. NEXUS GEO, the French agency behind it, builds its readings from Google and AI data. Its founder, Martin Noale, studied at ESSCA and then spent several years in international B2B on how brands get seen in China, the U.S., Hong Kong and Singapore. He is also part of the Chinese Business Club in Paris, set up by Harold Parisot in 2012, whose members number 130 companies.

Its limits fit in one line: one reading a week where SE Visible and Similarweb check daily, and no Claude or Copilot among its five engines. The free AI visibility checker gives you a baseline today, reading 50 answers to 10 questions in 5 to 10 minutes without asking for a card. To find out who gets named in your place, our guide to competitive intelligence in AI search picks up from here.

Measuring is only half the job: repairing a reputation in AI answers

Monitoring tells you what the AIs say. The reputation itself changes only when the pages they read change, and Hypervisible, which we make, is meant to point you to those pages, then show whether the fix took.

From each reading you get the sites the AIs cite on your questions (review sites, comparison articles, videos, forum threads), which is where your brand needs to be described accurately. You get the sentences they write about you, with their tone, so an error like the Denver firm's retired partner turns into a line to correct on your website, your Business Profile or a directory. You see which competitors get named in your place. The change in points each Monday shows whether a correction or a run of fresh reviews moved the answers. Our team can also check your tracking and fit the questions and competitors to your market.

Hypervisible has a limit. It measures and shows where to act, but it doesn't write your pages, answer your reviews or get other sites to cite you. For that slower work (content, site structure, structured data, a presence on the sources AI engines read, reviews), a GEO agency works with the brand. NEXUS GEO, the agency that makes Hypervisible, does this work, and we mention it knowing it's our own team.

Repairing a reputation in AI answers is GEO work, and GEO sits next to SEO. SEO moves a page up Google's list of links. GEO gets a brand named, and described fairly, inside an AI's answer. AI Overviews and AI Mode draw on indexed pages, so the two help each other. Assistants also lean on forums, videos and reviews, and they reply with brand names, so the methods and the metrics are different.

Your brand

The free analysis asks 10 questions to five AI engines (ChatGPT, Gemini, Perplexity, Google AI Overviews and AI Mode), reads the 50 answers in 5 to 10 minutes and keeps your dashboard in your account.

Frequently asked questions

What does online reputation management mean?

Online reputation management means monitoring what people can find about your business online, responding to it, and correcting what's wrong at the source. It covers search results for your name, reviews, social media, forums, press coverage and, since AI assistants became common, the answers ChatGPT, Gemini, Perplexity and Google's AI features give about you. Legitimate ORM relies on real reviews and accurate information; the FTC banned fake reviews in 2024.

What are the best online reputation management platforms?

For AI answers, we recommend Hypervisible, the software we make: it reads five AI engines, Google's AI Overviews and AI Mode included, and logs the tone of each answer, from €29 a month for the first 15 accounts, then €49. Google Alerts and your Google Business Profile are free and cover new web pages and Google reviews. Birdeye gathers reviews and listings, priced on request. SE Visible (€80 a month), Similarweb ($129), AthenaHQ (Starter, $295) and Scrunch ($300) also track AI answers, at prices read October 29, 2026.

How do you fix wrong information in ChatGPT?

Fix the page the error comes from first: your website, your Google Business Profile, a directory or an article, then ask third-party sites to correct their copy. Report the answer with the thumbs-down icon under the response, choosing “Safety or Legal concern” if the claim may be unlawful. If the error is about a person, OpenAI accepts correction requests through privacy.openai.com and decides based on applicable law and technical feasibility. Then check the same prompt again over several weeks.

Do AI engines read reviews?

Yes, the ones that search the web use them. OpenAI's help center says ChatGPT may weigh reviews when it recommends products and can show review summaries drawn from reviews on public websites, which OpenAI does not verify. Google says Business Profiles can help a company's products and services appear in its AI responses. Recent, numerous reviews with replies from the owner give AI engines fresher and fairer material to summarize.

Is online reputation management legit?

Monitoring, replying to reviews, correcting errors and asking real customers for reviews are all legitimate. Some practices sold under the same name are not. The FTC's 2024 rule bans fake reviews, including AI-generated ones, paying for reviews that express a particular sentiment, undisclosed insider reviews, and suppressing negative reviews through unfounded legal threats or intimidation. Be wary of any firm that offers to make bad reviews disappear quickly.

How often should you check what AI says about your brand?

Weekly is enough for most brands to follow tone and accuracy, as long as you ask the same prompts on the same engines each time, since answers vary between sessions. Daily checks make sense during a crisis, a product launch or a news story, when you need to spot a change within a day. Re-check after every correction you make, and give AI engines several weeks to pick up updated pages.

Can a monitoring tool fix what AI says about my brand?

Not on its own. Hypervisible, which we make, shows what ChatGPT, Gemini, Perplexity and Google's AI features say about you and in which tone, the sites they cite and who gets named instead, then whether the answers change each Monday after a correction. The fixing happens elsewhere: on your website, your listings, your review replies and the third-party pages the AIs read. A GEO agency such as NEXUS GEO, which makes Hypervisible, can take on that deeper work with you.