<aside> ⚡

I lead a data science team at Notion. My days are a mix of 1x1s, cross-functional syncs, Slack threads, email, docs, and dashboards, all competing for attention before 9am. I used to start every morning in full triage mode: skimming calendar, scanning Slack, checking email, opening up docs from yesterday. It took 30–45 minutes, and I still missed things.

So I built an agent to do it for me. I call it Jessica’s Chief of Staff 😅 a nerdy name for a very practical job: help me start the day calm, prepared, and on top of what matters.

</aside>

✅ What my Chief of Staff actually does (day to day)


🛠️ How it works under the hood

The agent is a Notion custom agent. Here’s what I have connected (nothing fancy, just the basics 💅):

Tool What it does
Notion Reads and writes pages, databases, docs, comments. Full workspace awareness.
Calendar Pulls today's meetings, attendees, and agendas. Optionally uses shared calendars (where you already have access) to help with scheduling context and meeting prep.
Mail Scans inbox for unread/unanswered emails. Filters for importance + urgency signals.
Slack Reads public channels and specific channels the org lives in. Surfaces threads with new replies or unresolved questions.

The agent has 4 triggers:

  1. Daily recurrence (weekdays 8am PT): generates the morning brief
  2. Weekly recurrence (Mondays 9am PT): updates the Living Context
  3. Weekly recurrence (Fridays 8am PT): Weekly org wrap
  4. @mention: I can talk to it anytime in Notion for ad-hoc questions

image.png

💽 Every brief lands in a dedicated "Daily Briefs" database with properties like date, status, topline, risks, and key links. The weekly context goes into a single Living Context page that accumulates over time. This is the agent's long-term memory.


📝 The instructions are the product

<aside> 💡

If there’s one thing I want you to take away, it’s this: the instructions page is where all the value lives.

</aside>

The agent runtime, the connectors, the triggers, those are infrastructure. The instructions are the strategy. They tell the agent: