Missions
Most AI forgets the moment you close the chat. Ayven maintains persistent organisational concerns — monitoring changes, tracking strategic questions, improving processes continuously. Issues that matter don't disappear after one conversation.
Confidence-based execution
Ayven knows when she's sure and when she's not, and her behaviour changes accordingly:
| Confidence | Behaviour |
|---|---|
High | Share findings directly, create public docs. |
Medium | Propose actions, request approval first. |
Low | Draft only, keep private, ask questions. |
Human input when needed
When Ayven gets stuck or needs direction, she doesn't spin. She asks:
- Blocker detected: the mission pauses and notifies you
- Question routed: a specific request is sent to you in chat
- Response captured: your answer feeds back into the mission
- Work resumes: she continues automatically with the new context
Mission lifecycle
| State | Meaning |
|---|---|
Open | Actively working: research, documentation, actions in progress. |
Paused | Blocked waiting for input, or circuit-broken after repeated failures. |
Closed | Goal achieved — but it may reopen if conditions change. |
How to use it
Give @Ayven a concern that outlives one conversation — "keep an eye on the onboarding funnel", "track the open questions from the platform review" — and she opens a mission for it. She reports progress, asks when she's blocked, and closes it when the goal is met.