Hallucination flags
AI engines sometimes state things about your brand that simply aren't true, a feature you don't offer, a price that's wrong, a claim you never made. A hallucination flag is how AppearIn AI surfaces those so you can correct the record.
Why engines hallucinate
Large language models generate plausible-sounding text from patterns, not from a verified database of facts about your company. When an engine is unsure, it doesn't say "I don't know", it fills the gap with something that sounds right. For a brand, that can mean an assistant confidently telling a prospective customer you integrate with a tool you don't, or quoting a price from an old page.
This is most likely when your public information is thin, outdated, or contradictory across sources, exactly the conditions the engine has to guess around.
How a flag is raised
AppearIn AI compares what each engine says about your brand against your brand profile, the structured facts you confirm about your product, pricing, features, and positioning. When an answer asserts something that contradicts the profile, or invents a specific claim that has no basis in your verified facts, it's raised as a hallucination flag with the exact wording and the engine that produced it.
Keep your brand profile accurate
Flags are only as good as the facts they're checked against. An out-of-date brand profile produces false flags (and misses real ones). Treat the profile as the source of truth and keep it current, see create a project.What to do with a flag
- Confirm it's wrong. Open the saved answer and read the claim in context. Sometimes the engine is technically right and the profile is stale.
- Find the source. Check the cited sources for that answer. A hallucination often traces back to one outdated or incorrect page, sometimes your own.
- Fix the underlying information. Update the page, clarify the fact on your site, or correct the third-party source. Engines re-read the web, so accurate, consistent public information is the durable fix.
- Track recurrence.Because prompts run on a schedule, you can watch whether a flag keeps reappearing after you've made a change.
An honest limitation
No tool can guarantee an engine will never hallucinate, the behavior is inherent to how these models work, and answers vary run to run. What AppearIn AI gives you is early warning and evidence: which engine said what, when, and what it cited, so you can respond quickly and measure whether your fix worked.