The parrot problem: How AI answers misrepresent your brand

Beyond Visibility: Mastering Brand Representation in the Age of AI Search Answers

James Cadwallader, Co-founder and CEO of Profound, addressed AI search at Web Summit Vancouver 2026. He highlighted a new paradigm: users query AI engines like ChatGPT, Perplexity, and Gemini, receiving direct responses. Marketers face a critical challenge: ensuring brand accuracy.

Mr. Cadwallader termed this the “parrot problem.” Southwest Airlines was misrepresented by AI citing outdated seating policies. Nike was compared to a competitor without solicitation. These instances show AI models “opining” subjectively.

The core issue is no longer just showing up, but how a brand is represented. Profound’s study analyzed 50,000 AI responses across industries and prompt types. They queried major AI engines, revealing insights.

ChatGPT responses averaged 3,000 characters (eleven tweets), while Claude was more concise. Crucially, 50% of AI responses consisted of unsolicited augmentation, including rationales, comparison tables, or follow-up actions.

For marketers, the new imperative is to manage both accuracy and sentiment of AI-generated content. Accuracy pertains to verifiable facts, while sentiment refers to subjective opinions. This dual control is vital.

To address this, Profound developed a “Fact check” feature, inspired by Meta’s Fact Score paper. This system ingests millions of LLM responses, scanning verifiable claims against “ground truth” data. It monitors and corrects inaccuracies.

Controlling the narrative requires strategic content creation. Mr. Cadwallader emphasized original, non-commodity marketing content that AI cannot easily derive from its pre-training data. This means creating content based on unique experiences or proprietary insights.

Content must also be fresh and specific. Profound’s research showed 50% of top-cited AI content is less than 13 weeks old, highlighting the need for up-to-date information. Marketers should prioritize granular, atomic details over broad platitudes.

Run Repeat, a publisher reviewing sneakers by cutting them in half, exemplifies this. Their highly specific, dated content is valued by AI, making them a “gold standard” for AI search optimization. Profound empowers marketing teams to deploy agents that monitor and shape brand presence.