Glossary

AI Hallucination

A hallucination is an AI assistant stating something false with full confidence — a business that doesn't exist, a service you don't offer, an address you left years ago. Grounded assistants hallucinate less because they read live sources, which makes the accuracy of those sources a business concern.

For local businesses, common hallucinations include incorrect hours, service areas, and phone numbers, or a competitor's specialty attributed to the wrong company. Thin or contradictory evidence makes these errors more likely.

A language model generates a plausible continuation from the question and available context. Clear, consistent retrieved facts give it better evidence for an accurate answer. Conflicting phone numbers or missing Saturday hours can lead it to select an incorrect value or infer an unsupported one.

That is why the errors cluster where documentation is thinnest. Assistants rarely invent a business name from nothing; they routinely invent hours, pricing ranges, and service areas, because those are the facts businesses most often leave unstated or state only in an image, a PDF, or a booking widget a crawler cannot read.

Reduce the risk by using one name, address, and phone number across current website pages and profiles. State hours, service radius, offered services, and important exclusions in crawlable text so retrieval has a supported fact to use.

When an error appears, trace the cited sources before changing the website. Correct an outdated directory or profile when it supplied the wrong fact, then confirm that the canonical page states the current value. Retest with the same question and settings, but expect some variation because retrieval and generation are non-deterministic.

Check what assistants currently claim about the business because many incorrect answers will not produce customer feedback. The free report returns sampled answers and the sources attached to them.

Related terms
  • Grounding (Web Search) Grounding is an AI assistant's use of live retrieved sources — usually web search — to answer, instead of relying on training memory alone.
  • Retrieval-Augmented Generation (RAG) Retrieval-augmented generation is the technique behind grounded AI answers: before responding, the system retrieves relevant documents — search results, web pages, database records — and generates its answer from them.
  • NAP Consistency NAP consistency means keeping a business's name, address, and phone number accurate across its website, Google Business Profile, directories, and social profiles.

See where your business stands.

The free AI visibility report shows what ChatGPT, Gemini, and Perplexity say when buyers ask about your market — with the sources behind every answer.

Get the free report All terms

No charge and no call required — the report arrives by email.