Whether the agent may SEE what the person is looking at (a whiteboard, a shared screen, a rendered component).
ask
on
off
Whether the agent may SEE what the person is looking at (a whiteboard, a shared screen, a rendered component).
ask(default) — no answer is pre-filled: the server's policy decides what is possible, and the "agent can see" control lets the person choose.on— the person's choice is pre-set to "may see" (up to what the server's policy allows).off— pre-set to "may not see pixels": the agent still learns what is on screen as text, never as an image.