AI meeting memory agent for prep, follow-up, and durable context

The job of a meeting memory agent

The useful agent is not just a transcript summarizer. It remembers who was involved, what changed, what was promised, which project it affects, and what needs follow-up before the next meeting.

OpenHuman-style workflows are a strong fit because they combine memory, notes, voice, connectors, and model routing into one assistant loop.

Useful scenarios

Meeting memory works best when the same people or projects appear again.

• Customer success calls with repeated commitments and risks.
• Founder or operator meetings that produce follow-up drafts.
• Engineering planning where decisions and owners matter.
• Research calls where source notes should become durable wiki entries.

Operation steps

Make every meeting artifact reviewable before it affects memory.

• Choose the calendar and note sources for the first meeting type.
• Generate a prep brief from prior context.
• Capture outcomes after the meeting and mark uncertain facts.
• Create follow-up draft language but require human approval before sending.
• Update project and people notes with concise durable facts.

Common risks

Meeting agents can create confidence faster than accuracy. Review remains essential.

• A transcript can include private or third-party restricted information.
• A follow-up draft can sound confident about unconfirmed decisions.
• People memory can become intrusive if preferences and context are not scoped.
• Voice and meeting features should not be enabled without consent rules.

Product connection

OpenHuman Online previews a meeting prep brief and Memory Tree before checkout, then directs high-intent users to Team annual hosted payment.

Quick answers