Documentation
How Jaha takes part in a call
Jaha can wait to be asked, speak up on its own when it has a fact the room is looking for, send one person an answer privately, and read what's on the shared screen. Each of these is a setting, per meeting, per playbook or for the whole workspace.
When asked, or reactive
In on-request mode, Jaha speaks only after “Hey Jaha”. In reactive mode it listens the whole time and speaks up briefly only when it has a specific fact the room is clearly looking for, like the date in the SOW or the number in the CRM. It never guesses. A meeting takes its mode from its playbook, and the playbook from the workspace default. The composer shows which one applies.
Reactive mode is paced: a pause between contributions, a limit per meeting, and a confidence threshold below which it stays quiet. Every time it speaks without being asked, it's logged on the meeting page with the reason.
Private answers
When guests are on the call and the answer is internal, like the lowest approved price or a term that isn't public yet, Jaha tells the room it has nothing it can share and sends the answer only to the account owner, logged with the reason. The discretion setting controls this: automatic (guarded when guests are present), open, guarded or private.
The shared screen
When someone shares their screen, Jaha captures each new slide or page (never a camera or a face), reads it, and can answer questions about it (“which milestone slipped?”). Screen reading is off by default. Turn it on in Settings › Screen share. Each captured screen is kept with the meeting.
Fine-tuning
Settings › Jaha has the numbers: how long it waits after a question before answering, how long a silence has to last, the pause and per-meeting limit for reactive mode, the confidence threshold, and how often it captures the screen. Each value has a range and a default, and a meeting can override the mode without changing the workspace.