Back to guides

Give the task a clear brief.

Set the items, conditions, and result before scheduling repeat work. These are example requests to adapt and test with your model and target apps.

Organize a set of information

Provide the list, source apps, and fields. For example, organize places from a specified collection by area and activity, leaving unavailable details blank. Start with a few items, check them, then expand the list.

Compare the same products

Specify the model, specification, quantity, and platforms. Record visible prices, offer conditions, and timestamps separately. Membership, coupons, and delivery can change the total; do not treat prices with missing conditions as directly equivalent.

Schedule a follow-up

Test the lookup first, then choose a supported schedule. State which workspace file to read and write, what history to keep, and how to compare it. The preview has no built-in product watchlist, price charts, or price-drop alerts. Prepare and check the workflow yourself.

Schedule a check-in

For each app, specify the instructions, phone, and time. Run it manually first. Records distinguish checked in, already done, unconfirmed, failed, and needs attention. Login or verification may require you to intervene and run the check again.

Review a batch

Ask for item-by-item completed, failed, and uncertain results. Check a small batch before expanding it. The current queue runs serially; one task can operate its bound phones in sequence.

Keep a useful procedure

Save a tested method as a skill. Supply the current items and criteria with later tasks. A skill does not grant extra device or channel permissions.