Positive brand sentiment may be valuable for customer trust, but it is not a dependable shortcut to visibility in AI-generated answers. A dataset spanning 9 million answers across nine AI platforms found that brands with more negative sentiment were, if anything, slightly more likely to be cited on five major platforms. The finding challenges a simple assumption behind many brand-visibility efforts: that an AI system will preferentially cite the companies discussed most favorably.
The analysis, reported in Search Engine Land's coverage of the 9 million-answer study, draws on Writesonic data covering more than 400 enterprise brands. It included ChatGPT, Google AI Overviews, Google AI Mode, Gemini, Perplexity, Copilot, Grok, Claude and Meta AI. This is an observational result, not proof that negative sentiment causes AI citations. Still, it is a useful warning for businesses treating reputation management and Generative Engine Optimization, or GEO, as the same discipline.
What the data changes about AI visibility
Across ChatGPT, AI Mode, Gemini, Copilot and Grok, the study found that brands with more negative sentiment were slightly more likely to appear as citations. It also found that positive sentiment did not reliably increase citation likelihood. The more plausible interpretation is not that businesses should seek negative coverage. Instead, AI answer systems appear to respond to the breadth and relevance of available discussion, rather than using favorable sentiment as a simple selection rule.
That distinction matters because a citation is not necessarily an endorsement. An AI answer may cite a review, comparison, complaint, news report or help page because it helps answer the user's question. A brand can therefore be visible in a generated answer while receiving little or no positive framing in that answer.
| Observed pattern | Finding from the research | What it means for measurement |
|---|---|---|
| Sentiment and citations | More negative sentiment was slightly associated with more citations on ChatGPT, AI Mode, Gemini, Copilot and Grok. | Do not use favorable sentiment as a proxy for AI visibility. |
| Source types | 64% of AI citations came from ordinary pages rather than specialized lists. | Useful, comprehensive pages can matter beyond rankings in list-style content. |
| Commercial queries | 82% of bottom-of-funnel commercial-query citations came from third-party sources. | Independent reviews and discussions can shape buyer-facing AI answers. |
| Brand recognition | Up to 75% of answers that cited a page did not name the brand in the answer text. | Citation counts alone may overstate recognizable brand influence. |
The study also points to a measurement problem often overlooked in AI visibility work: ghost citations. In this pattern, an answer cites a page but does not explicitly mention the relevant brand. The underlying content may have informed the answer, yet the brand receives little direct recognition from the reader. This makes it important to distinguish between being a source used by an AI system and being a company named in the final answer.
For businesses, the practical conclusion is not to abandon sentiment monitoring. Negative reviews, complaints and unfavorable reporting can affect conversion, trust and customer retention even if they do not reduce AI citations. Rather, sentiment and AI visibility should be measured as related but separate outcomes.
A practical approach to reputation and AI content
The data comes from Writesonic's enterprise client base, and results can vary by platform, industry, language and prompt type. Businesses should validate the pattern against the questions their own customers ask. A local service company, software provider and ecommerce retailer may face very different citation sources and answer formats.
A useful working process includes:
- Monitor both citations and answer language. Track whether an AI answer cites your pages, names your brand, recommends it, or presents it in a negative context.
- Separate ORM from GEO goals. Reputation work should address legitimate customer concerns and improve trust. GEO work should improve the availability, clarity and coverage of useful information that answers relevant questions.
- Audit third-party sources for commercial topics. Since third-party sources represented 82% of citations for bottom-of-funnel commercial queries in this research, reviews, comparisons and industry coverage can be important inputs to what AI systems surface.
- Build durable owned content. The study notes that owned pages can account for a small share of citations, about 3% in some contexts. That is a reason to improve pages that explain products, services, use cases and limitations clearly, not a reason to stop investing in them.
- Use automation with editorial controls. AI can help teams identify repeated customer questions, organize source material and flag content gaps. Human review remains necessary to ensure published information is accurate, specific and aligned with the business's real offering.
The strongest content strategy is not a campaign to manufacture praise. It is a sustained effort to make reliable information available wherever customers and AI systems may need it. Comprehensive owned pages can support long-term authority, while credible third-party coverage and real customer feedback shape the broader discussion surrounding a brand.
AI visibility is difficult to manage when citation data, brand mentions and customer sentiment are treated as one metric. GEO Search Leads can help businesses turn AI-search visibility checks into a clearer view of where their brand is surfaced, omitted or discussed by third-party sources. That creates a practical basis for prioritizing content improvements and reputation work without guessing which signals matter. Start an AI visibility scan.
Frequently Asked Questions
Does negative brand sentiment improve AI citations?
No. The research found a slight association between more negative sentiment and citations on five platforms, but it does not establish that negative sentiment causes a brand to be cited or that the pattern applies universally.
What are ghost citations in AI answers?
Ghost citations occur when an AI answer cites a page but does not name the associated brand in the answer text. The source may influence the response without creating direct brand recognition.
Why should businesses track third-party sources for AI visibility?
The study found that 82% of citations for bottom-of-funnel commercial queries came from third-party sources. Reviews, comparisons and other independent pages may therefore influence what buyers see in AI answers.
Should a business stop publishing owned content if third-party sources receive more citations?
No. Owned content remains important for explaining a business's products, services and expertise. The research suggests it should be part of a broader strategy that also monitors third-party discussion and AI answer outcomes.
Conclusion
The Writesonic findings do not make negative sentiment a visibility tactic. They show that AI citation behavior is more complex than a positivity score. Businesses should protect their reputation while separately measuring citations, explicit brand mentions, third-party coverage and the quality of their own content. That approach offers a more realistic foundation for improving visibility in AI-generated answers.
Top comments (0)