Sveriges mest populära poddar
Notion in Practice

Notion Custom Agents Deep Dive: Triggers, Permissions, and Real Use Cases

53 min24 februari 2026

Custom Agents are the next step for Notion AI: not just answering questions, but running real workflows in the background.

In this hosts-only deep dive, Tim Jeffries and Jerwin Parker break down what Custom Agents are, how they differ from the “regular” Notion AI experience, and how to think about implementing them safely in a team.

Tim and Jerwin walks through the Custom Agents interface, including triggers, activity logs, permissions, and integrations, then shares a set of practical agents Smooth Ops has been using to reduce admin overhead and increase project context.

In this episode, we cover
  • Custom Agents vs Notion AI: what changes when AI can run autonomously
  • Personalized instructions and playbooks: the foundation for consistent outputs
  • Triggers: schedule-based runs, property-based triggers, and Slack/Calendar/Mail possibilities
  • Permissions and governance: what the agent can access, and how sharing agents can extend access
  • Real examples (Tim): meeting follow-up, Slack context capture, email-to-CRM logging, lead intake automation, and project status updates
  • Real example (Jerwin @ TrustOnCloud): a meeting triage custom agent for a marketing leader to capture actions, track accountability, and keep momentum after internal calls
  • Ambient context: why Tim rebuilds a daily context page overnight so AI stays “stateful” across conversations

Three Actionable Takeaways
  1. Takeaway #1: Start with personalized instructions + playbooks before you build agents.

    Build a reliable “baseline assistant” first, then turn repeatable playbooks into agents once you trust the workflow.

  2. Takeaway #2: Treat agents like employees: define the job, grant minimal access, then audit regularly.

    Good outcomes come from clear instructions, tight permissions, and reviewing activity logs.

  3. Takeaway #3: Use agents to capture context automatically (meetings, Slack, email) so work stays “stateful.”

    The more structured context lives in Notion, the less the AI has to guess, and the better the outputs become.

Who this is for

Operators, consultants, founders, and Notion power users who want to automate repeatable work, reduce cognitive load, and keep project context up to date without relying on memory.

Subscribe for more practical episodes on building real systems in Notion, with workflows you can copy and adapt.

Notion in Practice med Tim Jeffries and Jerwin Parker finns tillgänglig på flera plattformar. Informationen på denna sida kommer från offentliga podd-flöden.