Agents overview

Agents are supervised AI teammates that act on your event graph — sourcing speakers, drafting evaluations, packaging content, and more. They run at a tunable level of autonomy, with every action logged and reversible.

Training video · 3:50Reviewer agent: pre-score submissions on your evaluation planWhat the Reviewer agent does, switching it on from the Agents page, its personas (the Active Personas panel, the Persona Library, Create New), enrolling personas on an evaluation plan round from Evaluators → Personas (Add personas, Edit assignments, Run personas, Progress), where its scores land (the persona's per-submission scores, View feedback, the session's Reviews tab and the Reviewer agent results card), and how the committee reads persona scores beside human evaluators on the plan's Review tab and keeps every decision.

1. 0:02The Reviewer
2. 0:37Turning it on
3. 0:57Personas
4. 1:33Personas on a plan round
5. 2:10What it produces
6. 2:38Review and override
7. 3:17Next steps

Training video · 3:08Team Lead: your guide on every pageTeam Lead is the guide in the sidebar. Open it from Ask on any page and it reads where you are, offers the questions to ask about that page, answers from your event's own data and the Help Center, and keeps the matching articles and training videos one tab away.

1. 0:02Team Lead
2. 0:18The page guide
3. 0:49Asking a question
4. 1:19It follows you
5. 1:36Create
6. 1:52Learn
7. 2:08Watch
8. 2:24FAQ and Attention

Note

Agents build on the same data and permissions as the rest of Sessionboard. What an agent can see and do is governed by its scopes and your Event Team permissions.

What agents can do

Source speakers

Research and enrich prospects in the CRM Pipeline — logged in the card’s Activity feed under Agent.

Screen submissions

Screener scores submissions as AI personas, and summarizes earlier rounds’ comments, against your review plans. People still make every decision.

Build the agenda

The Scheduler agent drafts and organizes sessions from the agenda sidebar.

Report & analyze

Answer questions in natural language through Insights and SbQL.

Autonomy & trust

Agents operate at supervised autonomy levels — they propose and act within scope, but you stay in control:

• Activity logging — every agent action appears in the relevant record’s Activity feed, distinct from User, System, and Rule events.
• Scoped access — agents inherit token scopes; see API tokens.
• Reversible — agent changes are auditable in History and can be rolled back.
• Run history — the Coordinator keeps a record of each pass it makes over your participants, including how many it did not reach. See Check Coordinator run history.

Connect your own AI assistant

Any MCP-compatible assistant (Claude, ChatGPT, Copilot, Gemini, Cursor) can act as an agent against your event data via the Sessionboard MCP server. See Connect AI assistants to Sessionboard.

Note

The Agents tab is in every event’s sidebar. Ask is always on for every organization and every event — it is the guide to the product, and its card has no switch. Open Ask on any page to see where you are and what is set up there. The other agents are turned on from that tab; if one shows as locked, ask your account manager about AI access. Ask also has a shortcut row on the Event Dashboard.