Trustpilot on April 11 launched an AI Visibility Suite, a set of tools designed to help brands track and optimize their presence in AI search results from ChatGPT, Claude, and Perplexity. The product responds to a shift in consumer behavior that is already measurable: 58% of consumers now use generative AI for product recommendations, according to Capgemini research.

The suite ships three components. The In-App Review Collector gathers verified feedback at the point of experience. The Invitation Optimizer times review requests for higher engagement. The AI Visibility Metrics dashboard tracks how frequently a brand appears when AI systems surface recommendations.

The Citation Spike

The product launch is backed by concrete data. Trustpilot’s structured review data drove a 246% surge in ChatGPT citations between June and August 2025. By January 2026, Trustpilot had become the 5th most-cited page by ChatGPT, according to PromptWatch data.

The platform hosts more than 361 million active reviews as of June 2025 and removed 7.8 million fake reviews in 2025, with 91% detected automatically, according to the company’s announcement.

Answer Engine Optimization Takes Shape

Trustpilot’s move makes “answer engine optimization” tangible. Traditional SEO optimizes for Google’s ranking algorithm. Answer engine optimization targets a fundamentally different system: AI models that synthesize information from multiple sources to generate recommendations rather than listing links.

The difference matters for brands. In traditional search, a company can rank without being explicitly recommended. In AI search, the model decides whether to mention the brand at all. Brands that AI models don’t surface during the recommendation step lose deals they never see, as CMSWire also reported.

For Trustpilot, the suite represents a pivot from reviews platform to infrastructure for AI-driven commerce. For every company that sells products online, it raises a practical question: what does your brand look like to an AI agent making a purchase recommendation, and can you measure it?