FAQ

What’s the difference between FactCheck and Sentiment?

FactCheck evaluates the factual accuracy of what's said about your brand, while Sentiment measures the tone and perception of how your brand is described. The two features are complementary. Use them together for the best results.

How often does the FactCheck run?

FactCheck runs daily across all your FactCheck-tagged prompts. When set up for the first time, it takes 7 days for FactCheck results to appear in Profound.

What is the maximum number of FactCheck prompts I can configure?

The limit is 200 prompts. If you need more, reach out to your Customer Success representative.

What types of content can I use as a source of truth?

Support documentation and help center pages work well because they're factual, structured, and ideally regularly maintained. Internal product specs, policy documents, and Q&A documents also work well. For teams getting started, a simple Q&A document listing your brand's most critical facts, such as pricing, cancellation policies, or product specifications, is the fastest way to get up and running. Adding your website is also good practice, since public-facing websites are often updated more regularly than internal documentation. Avoid uploading broad or conflicting content, as this reduces the accuracy of claim verification.

What happens if there is nothing in the Knowledge Base?

FactCheck requires a Knowledge Base to function. Without one, there's no source of truth to verify claims against, and the feature won't produce results. Even a simple Q&A document covering your brand's most important facts is enough to get started.

Can I track how often AI responses align with our official brand facts?

Yes. FactCheck surfaces a representation accuracy metric: the percentage of times AI engines state correct information about your brand across your FactCheck-tagged prompts. Track this over time to measure progress as you update citation sources, expand your Knowledge Base, or add new prompts.

Does FactCheck work only for brand-specific facts, or can it verify other types of claims?

FactCheck is designed for brand-specific, fact-based claims: product specifications, pricing, policies, safety information, features, and service terms. It isn't designed to verify general legal or medical information that isn't specific to your brand, and it doesn't verify visual content such as logos or images in AI-generated media.

Why can some claims not be verified?

FactCheck verifies only claims it can confidently check against your Knowledge Base. If the relevant information isn't in the Knowledge Base, the claim is left unverified. Expanding and updating your Knowledge Base directly increases verification coverage.

Can FactCheck help us get ahead of misinformation before a product launch?

Yes. FactCheck works best when used proactively. Before a launch, upload updated product documentation to your Knowledge Base and tag prompts that cover the facts most likely to be misrepresented. Once FactCheck identifies inaccurate citations, your content team can contact those sources or submit corrections before the misinformation spreads. The citation-level breakdown shows which third-party pages to prioritize.

Can FactCheck help recover brand reputation after a negative event?

FactCheck gives you visibility into which AI responses state incorrect or outdated information about your brand and which sources are responsible. After a negative event, use this data to identify the specific pages driving harmful claims, take action to update or correct those sources, and track whether representation accuracy improves over time. Used alongside Sentiment, it gives you a complete picture of both what AI says and whether it's factually correct.