← Notion
Complex day at Notion

Prove the agent is safe before rollout

You’re the ai / ml engineer. Your team is in the room. Printed Aug 6, 2026.

A useful summary can become a serious problem when its source boundary is unclear.

Teams must weigh a committed launch against the cost of discovering a rare failure in live operational knowledge.

Who you’d be doing this for

“I need to know what the summary saw, what it changed, and who was allowed to see it.”

Arunee Intharasak · Director of Compliance Operations

Owns controlled procedures that will be summarized into an enterprise operational database.

What is at stake

The launch date is fixed, but the generated tests missed rare paths through permissions that reviewers only found later. You have to decide what evidence is enough to sign off on a rollout you cannot reverse.

Why it isn’t already fixed

Every obvious fix costs something else. That’s the part you’d have to decide.

  • launch commitment versus verification certainty
  • summary usefulness versus permission preservation
  • automated coverage versus human evidence
  • vendor assurances versus product-specific control

Why Notion

At Notion, this often matters because shared workspaces combine knowledge, permissions, and operational databases.

Written with these in mind

AI safety engineerML evaluation specialistResponsible RAG systems builder

Not your kind of problem? 14 more at Notion, or browse every organization.

This is the setup. The work is inside.

Running it puts you in the room: the full situation and its constraints, stakeholders who push back in their own words, and the decisions that are yours to make. What you produce becomes a Day One Plan — work you can show someone instead of describing.