Board first, guided brief on request

Mail2Flow’s prototype offered two views of the same morning. The task board is a kanban with three columns, needs action, needs review and needs follow-up, plus filters; the accountant picks what to open. The guided brief is the AI’s answer to the same inbox: the most urgent task already open with a summary, two proposed actions and a chat assistant, and the other tasks listed below it. Everyone saw both, in counterbalanced order.

The board won on every measure the test could take. In the six-second look people recalled its columns and its count; the brief was too much text to parse. While exploring, all seven counted their four tasks on the board; on the brief only one of the six recorded answers was right, and four people thought they had two, because the open task read as an e-mail and the rows below read as subtasks. Five of seven preferred the board, and their reasons came in the same words: overview, my own choice, less effort.

Two people preferred the brief, both juniors, because the urgent task was already open and ready to act on. That is the part worth keeping. The report’s path, and this pattern, is a combination: the board as the landing view, the brief as a focus mode that opens from a task. What a bigger test should separate is the ordering from the rest, because the brief in this test also carried the two side-by-side proposals that confused people on their own (see One proposed action, one named button).