Why is AI saying negative things about my brand?
Engines repeat what the pages they retrieve say, so negative lines usually trace to reviews, forum threads, news or old posts. Find the source of each line, separate what is wrong from what is fair criticism, correct the facts and respond honestly. No engine documents how to request a correction to an answer.
Free to read, no signup. Nova traces each negative line to its source.
What the documentation covers
No engine documents how to ask for a corrected answer. Anthropic documents how to report policy-violating content and how to keep your own content out of Claude's search results.
Reporting is for policy violations
Anthropic's help article covers reporting content that breaks its usage policies or local law, and keeping your own content out of Claude's web search. It does not describe a way to request corrections to incorrect information.
Reporting, blocking and removing content from ClaudeAccuracy and honesty come first
Google's guidance asks content to be accurate and warns against content made mainly to attract search traffic. Honest, factual responses to criticism are the safest route.
Creating helpful, reliable, people-first contentFive steps to do it yourself
Do it by hand with these steps, or hand the whole thing to Nova.
Ask the engines about your reputation
Ask 'Is [brand] any good?', 'Problems with [brand]' and 'Is [brand] worth it?' Save answers and cited pages.
Nova: asks reputation questions across engines and saves the sources.
Trace each negative line
Open the cited pages and find the sentence behind each negative line.
Nova: traces each negative line to its source.
Separate wrong from fair
Mark each claim as factually wrong, outdated, fair criticism or unclear. Only the first two are for correction.
Nova: classifies each claim for your review.
Respond to the sources
Correct facts with proof, update what you own and reply honestly to fair criticism, disclosing who you are.
Nova: drafts responses and corrections for approval.
Recheck
Re-ask the questions in a few weeks.
Nova: re-checks and reports the change.
Short on time? Nova, your AI brand visibility expert, runs this Maneuver on your brand and tracks the result.
Take it to NovaEvery negative line and what to do
| Engine | Says | Type | Likely source | Move |
|---|---|---|---|---|
| ChatGPT | [NEGATIVE CLAIM] | Outdated | Old review | Update and respond |
| Perplexity | [NEGATIVE CLAIM] | Fair criticism | Forum thread | Reply honestly |
| Gemini | [NEGATIVE CLAIM] | Wrong | News article | Request correction |
Nova then re-asks the questions at 30 days, so you see how the picture changes.
Take this Maneuver into your own AI agent
Get sentiment-watch.md, a ready-to-use skill file with a reputation question set, a claim classifier and a response template. Drop it into Claude, Cursor or any agent that reads SKILL.md files.
Skill file for Claude, Cursor and other agents
- A reputation question set
- A wrong-versus-fair classifier
- A response template
Questions, answered.
Relevant reads
All articlesLet Nova trace what AI says about you
Nova opens with this Maneuver loaded. Nothing ships without your approval.