# What is Meeting Assistant? Also called AI Notetaker, Meeting Bot. A meeting assistant is a system that joins or connects to a video or audio meeting to capture what is said and produce usable output afterward, such as a transcript, a summary, and action items. It typically joins as a participant through a calendar invitation, or connects through the conferencing platform's own recording interface. Two joining models exist. A bot participant dials into the meeting like anyone else, appears in the participant list, and receives the mixed or per speaker audio. A platform integration instead uses the conferencing provider's recording and transcription hooks, which avoids an extra attendee but depends on what that provider exposes. Calendar access is what tells the assistant which meetings to attend. Raw transcription is rarely the deliverable. The assistant segments the conversation by speaker, identifies decisions, commitments, questions left open, and dates mentioned, then writes those into a summary and often into a task tracker or customer record. The value comes less from the notes existing than from what happens to them after the meeting ends. Recording people carries obligations. Participants should be informed, and in several jurisdictions consent is required from everyone present, which is why most assistants announce themselves on joining. Accuracy is the other caution, since crosstalk, jargon, and unfamiliar accents produce transcription errors, and a confidently written summary can carry an error further than a raw transcript would. Adoption usually stalls on trust rather than capability. External guests may object to an unfamiliar participant in the list, and internal teams stop reading summaries that are long, generic, or occasionally wrong on names. Short output, visible timestamps linking back to the recording, and an easy correction path do more for adoption than better models. ## Key points - Joins as a participant or connects through platform recording hooks - Calendar access determines which meetings it attends - Output is summaries and action items, not only transcripts - Participants should be informed; consent rules vary by region - Summaries can carry transcription errors confidently ## In practice A consultancy has its assistant join client calls listed on the shared calendar. It announces itself, records, and afterward posts a summary to the project channel: three decisions, two open questions, and four commitments with owners and dates. The commitments become tasks in the project tool. A partner reviewing the notes spots that a client name was misheard and corrects it before the summary is circulated. ## Related terms - [Meeting Transcription](/en/glossary/meeting-transcription) - [Calendar Integration](/en/glossary/calendar-integration) - [Speech to Text](/en/glossary/speech-to-text) - [Microsoft Teams App](/en/glossary/microsoft-teams-app) - [Omnichannel](/en/glossary/omnichannel) [Back to the AI Glossary](/en/glossary)