Why did the AI answer incorrectly?

Like any AI system, Meetbook's chat and Second Brain can occasionally get something wrong — misattributing a quote, missing context, or stating something with more confidence than the transcript actually supports.

Common causes

• Thin or inaccurate source material. If the underlying transcript has errors (see How transcription works), an otherwise-correct AI answer can still be wrong because it's working from a flawed transcript.
• Ambiguous questions. A vague question ("what did we decide") across a meeting with several small decisions can get summarized in a way that drops nuance.
• Cross-meeting confusion in Second Brain. With many similar meetings (e.g. recurring calls with the same client), the AI can occasionally conflate details from two different sessions.

Reporting a bad answer

There's no dedicated "report this answer" button in chat today — for a report-level issue, use the thumbs up/down under a meeting's Overview (see How your report is generated). For a chat answer that was clearly wrong in a way worth flagging, contact support with the meeting and the question you asked — see How to contact support.

Reducing the risk

Treat AI answers as a fast starting point, not a verbatim source — for anything high-stakes (a legal commitment, a specific number someone quoted), check the underlying transcript directly rather than relying solely on the summarized answer.