Team workload view with availability

The team dashboard in the Mail2Flow prototype shows who holds which tasks and which ones are overdue. Task 6 asked the accountants to find it, read the workload and act. Four found it on their own; three could not, because the sidebar showed icons without labels, and once the moderator opened it all three called it clear and valuable. That is a plain usability rule, icons need visible labels, and it decides whether an AI feature is used at all.

Two AI-specific lessons came with it. One senior reassigned an overdue task to a colleague who had no tasks, without noticing that he was marked absent; availability has to sit in the picker, not in a badge people can miss. Two participants asked how reliable the dashboard’s numbers were, which is the question every AI dashboard should answer on the screen: what the numbers are counted from and how fresh they are. Participants also asked for start dates and deadlines per task, drag-and-drop reassignment, a notice when an assignee has seen a task, and a way to show tasks that belong to several people.

Assigning work to people touches the AI Act’s employment area (Annex III, point 4). A view that shows load and availability, with a person who moves the task, keeps the decision with the person. Letting the system allocate tasks on the basis of how individuals behave, or evaluate their performance, is listed as high-risk in Annex III, point 4(b). Design considerations, not legal advice.