How to Fix Negative AI Sentiment About Your Brand

This check flags AI answers that mention your brand next to negative terms such as 'scam', 'complaints', 'hidden fees', 'lawsuit' or 'poor support'. By the end of this guide you will have logged every negative answer, traced it to the pages the AI is drawing from, fixed or answered those sources, and published a crawlable factual page so newer answers have accurate material to cite.

By Updated ⏱ 3-6 hours of work, plus 2-8 weeks for AI answers to refresh🎯 Intermediate● Medium impactGEO / AI
Illustration: AI sentiment about your brand fix guide
GEO / AI · FIX GUIDE AI sentiment about your brand IMPACTEFFORT TIME3-6 hours of work, plus 2-8 weeks for AI answers to refresh LEVELIntermediate

Why it matters

ChatGPT, Perplexity, Gemini, Copilot and Google AI Overviews summarise what the web says about you, often in one or two sentences. ' or 'Example Co vs competitor'. Buyers see it before they reach your site, so it lowers click-through and conversion even when your rankings hold steady.

You cannot edit the model, but you can change the sources it retrieves and the evidence it weighs.

✕BeforePerplexity answers 'Is Example Coreliable?' with 'Some users reporthidden fees and slow refunds', citing✓AfterThe same prompt returns 'Example Copublishes flat pricing and processesrefunds within 5 business days, per

What you need

Access to ChatGPT (with search), Perplexity, Gemini and Google Search with AI Overviews
Owner or manager access to your review profiles (Google Business Profile, Trustpilot, G2 or similar)
CMS access to publish and edit pages on example.com
A spreadsheet to log prompts, answers, negative terms and cited sources

Step-by-step fix

1

Log every AI answer that frames you negatively

Run the prompts buyers actually use in each AI tool, signed out or in a fresh session, and paste the full answer into a spreadsheet. Record the tool, date, prompt, the exact negative phrase and every cited URL. Include brand-only prompts and comparison prompts, because negative framing often appears only when you are compared with a competitor.

text
Is Example Co legit?
Example Co reviews
Example Co complaints
Example Co pricing hidden fees
Example Co vs [competitor]
Best alternatives to Example Co
What do customers say about Example Co support?
Has Example Co had any security issues?
Pro tip: Run each prompt twice. Answers vary between runs, and a phrase that shows up in both runs is the one to fix first.
2

Trace each negative claim to its source

For each negative phrase, open the citations: the numbered sources in Perplexity, the 'Sources' button in ChatGPT search, the link cards in Google AI Overviews and the 'Double-check response' option in Gemini. If no source is shown, search the exact phrase in quotes on Google and on Reddit. Tag each source as one of:

  • Review (Google, Trustpilot, G2, Capterra, app stores)
  • Forum (Reddit, Quora, community boards)
  • Article (news, blog, listicle, comparison site)
  • Your own site (old pricing pages, archived blog posts, status pages)
Pro tip: Sort the sheet by how many answers cite each URL. A small number of pages usually account for most of the negative framing.
3

Fix what is true, then reply to the reviews

If the complaint describes a real problem, such as slow refunds or a confusing cancellation flow, fix the product or policy first. Replies alone will not change the overall pattern. Then reply publicly to every cited negative review: state what changed and when, and give a named contact route.

In Google Business Profile, go to your profile in Google Search or Maps, choose Reviews, then Reply. In Trustpilot Business, go to Service reviews > Reply to reviews. Report reviews that break platform policy (fake, conflict of interest, off-topic) using each platform's flag option.

text
Hi Sam, thanks for flagging this. In March 2025 we moved refunds to an automated flow, and they now complete within 5 business days. If yours is still open, email support@example.com with your order number and I'll handle it personally. - Alex, Support Lead, Example Co
Pro tip: After resolving a case, invite the reviewer to update their review. Do not offer incentives in exchange for changes, because most platforms forbid it.
4

Answer forum threads openly as the brand

On Reddit, Quora or niche forums, reply from a clearly labelled company account or a named employee who discloses their role. Correct facts with links to evidence such as your changelog, policy page or status history, and acknowledge anything that was true at the time. If a thread is locked or archived, start a new, factual 'what changed' post where the subreddit's rules permit it, and message the moderators first.

Pro tip: Never post from fake customer accounts. Moderators and AI systems both treat astroturfing as a negative signal, and exposure creates a new negative thread.
5

Get outdated articles corrected or updated

For third-party articles with stale facts, email the editor or author with a short correction: the outdated sentence, the current fact, and a link proving it. Ask for an update or an editor's note, not a takedown. On your own site, update old posts that still describe retired pricing or past incidents without context.

Redirect retired product or pricing pages with a 301 to their current equivalent, rather than leaving them live.

nginx
# Nginx: redirect a retired pricing page
location = /pricing-2022 {
    return 301 https://example.com/pricing;
}
Pro tip: Give the editor a ready-to-paste replacement sentence. Corrections that take the editor under a minute are far more likely to get published.
6

Publish a factual source-of-truth page

com/trust or a detailed FAQ. Cover pricing transparency, refund timelines, support hours, security certifications and a dated list of past incidents and how each was resolved. Write plain, specific sentences an AI can quote.

Add Organization markup with sameAs links to your official profiles so systems can match your brand to the right entities.

html
<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "Organization",
  "name": "Example Co",
  "url": "https://example.com",
  "logo": "https://example.com/logo.png",
  "sameAs": [
    "https://www.linkedin.com/company/example-co",
    "https://www.trustpilot.com/review/example.com",
    "https://www.g2.com/products/example-co"
  ],
  "contactPoint": {
    "@type": "ContactPoint",
    "contactType": "customer support",
    "email": "support@example.com"
  }
}
</script>
Pro tip: Put a visible 'Last updated' date on the page. Recency helps both readers and retrieval systems trust the page over older third-party claims.
7

Let AI crawlers read your corrections

txt to confirm that the search and retrieval bots used by AI answer engines are not blocked. Blocking them leaves third-party pages as the only sources these engines can retrieve about you. Training crawlers such as GPTBot and Google-Extended are a separate choice and do not control live search citations.

Also confirm the new pages are indexable: no noindex tag, a self-referencing canonical, and inclusion in your XML sitemap.

text
User-agent: OAI-SearchBot
Allow: /

User-agent: ChatGPT-User
Allow: /

User-agent: PerplexityBot
Allow: /

User-agent: Googlebot
Allow: /

User-agent: Bingbot
Allow: /

Sitemap: https://example.com/sitemap.xml
Pro tip: Copilot and ChatGPT search draw heavily on Bing's index, so submit the new page in Bing Webmaster Tools under URL Submission as well as in Google Search Console under URL Inspection.
8

Re-run the prompts and re-audit

After 2-4 weeks, re-run the exact prompts from step 1 in the same tools and add the results to your sheet. Compare the negative phrases and cited URLs with the originals. Then run the audit again.

The check passes when no AI answer that mentions you includes negative framing. Repeat monthly, because new reviews and threads can bring the problem back.

Pro tip: Track the share of answers that cite your own pages. A rising share is an early sign of improvement, even before every negative phrase disappears.

Platform-specific notes

Google Business Profile

Reply from Google Search or Maps by opening your profile and choosing Reviews. To report a review, use the three-dot menu and select 'Report review', or use the Reviews Management Tool to track the status of your removal requests. Google removes reviews only for policy violations, not for being negative.

Trustpilot

Reply under Service reviews in Trustpilot Business. Use 'Report review' only for guideline breaches. Trustpilot shows companies that reply to negative reviews, which AI summaries often repeat as a positive signal.

Reddit

Each subreddit has its own self-promotion rules. Read the sidebar and message the moderators before posting as a brand. Brand accounts should state their affiliation in the profile and in the comment.

WordPress

Edit outdated posts in place rather than duplicating them. Use a plugin such as Redirection, or Yoast/Rank Math's redirect manager, for 301s on retired pages. Check that your SEO plugin has not set the new trust page to noindex.

Cloudflare

Bot Fight Mode, the AI Crawl Control settings, and the managed robots.txt option can block AI search crawlers even when your own robots.txt allows them. Under AI Crawl Control, check that search bots such as OAI-SearchBot and PerplexityBot are allowed.

Check that it worked

✅ How to verify

  • Re-run your logged prompts in ChatGPT, Perplexity, Gemini and Google AI Overviews and confirm the original negative phrases no longer appear.
  • Check that the citations now include your trust or FAQ page, or updated third-party articles, instead of the old threads and reviews.
  • Use URL Inspection in Google Search Console and the URL Inspection tool in Bing Webmaster Tools to confirm the new pages are indexed.
  • Re-run the audit and confirm it reports no negative framing in AI answers that mention you.

⚠️ Common mistakes

  • Publishing positive content without fixing the underlying problem, so new complaints keep feeding the same narrative.
  • Posting fake reviews or sockpuppet forum replies, which breaks platform rules and creates new negative coverage when discovered.
  • Blocking AI search crawlers in robots.txt or at the CDN, leaving third-party pages as the only sources AI engines can retrieve about you.
  • Testing once while signed in. Personalised or memory-influenced answers hide what a new buyer actually sees.

Is this still an issue on your site? Run a free audit — ZORO checks this and 294 other SEO, AI-visibility, accessibility and security signals, with fixes for your platform.

Check my site free

FAQ

Can I ask ChatGPT or Google to remove negative information about my brand?

Not for opinions or accurate reports. You can submit feedback on specific answers, use the thumbs-down option, and use legal or privacy removal routes for defamatory or personal data. The dependable fix is changing the sources the AI retrieves.

How long does it take for AI answers to change?

Answer engines that use live search, such as Perplexity, ChatGPT search and AI Overviews, can reflect updated sources within days to a few weeks after recrawling. Answers drawn mainly from a model's training data can take months, until the next model update.

Why does AI mention a complaint that was fixed years ago?

The old thread or article is probably still well-linked and easy to retrieve, and nothing newer contradicts it. A dated 'what changed' page, plus replies or corrections on the original source, gives the AI a more recent fact to cite.

Do review replies actually affect AI answers?

Yes. AI summaries of review sites often note whether a company responds to and resolves complaints. Specific replies that state a fix and a date are often quoted directly, which softens or replaces the negative framing.