Agenda Agent
Draft agenda from standing items, matters arising and submitted papers
They draft, check, chase and cite. People decide. Nothing becomes an official record until someone adopts it.
Every run is on behalf of a named user and can never see or do more than they can.
Claims without a source are rejected. Unclear transcripts get [CHECK], not a guess.
Official records, external messages and policy changes go to the Agent Inbox.
Inputs, tool calls, sources, model, prompt version, cost and approver — for every run.
24 agents
Draft agenda from standing items, matters arising and submitted papers
Completeness and format feedback to the author
Compiled, bookmarked, watermarked board pack
One-page brief per paper, linked to source pages
Pointed, cited questions and gap analysis per member
Suggested actions and decisions in a side panel
House-style minutes, every sentence linked to the transcript
Reminders, collected updates and escalations
Formal minute extracts to each faculty
Answers cited to minute numbers and pages
Annual governance report
Policy set with clauses cited to pack requirements
Vendor research, proposed risk tier, DPIA pre-fill
Missing or unclear AI-use declarations flagged
Assessment formats that hold up at the chosen AI level
Five-minute lessons, quizzes and scenarios
Answers cited to the exact policy clause
Readiness score, gap list and tasks
Draft regulator report citing live evidence
Severity, routing and DPO alert
Decisions extracted into the register, flagged "Imported — verify"
Answers about lessons, cited to the lesson and policy clause
Help limited to the assessment's AI-use level, logged in the process timeline
VC briefing: meetings, overdue actions, compliance risks
Measured, not assumed
The Minutes Agent is scored against gold-standard minutes for decision recall, action recall and unsupported claims before every release.