This guide gives you a reusable starting pattern. It is designed to help you see the work more clearly; it is not a substitute for judgment, source checking, or responsibility for the result.
Prepare the inputs before you ask for output.
The model only sees what you give it. Spend a few minutes naming the reader, desired result, and uncertain information. This makes a first draft easier to assess and reduces the need for decorative rewriting later.
Give the task a useful brief.
Help me choose a realistic weekly priority plan from this project list. Outcome for the week: [state it] Available time and people: [constraints] Tasks with known deadlines or dependencies: [paste list] Risks or decisions that need an owner: [paste list] Group tasks into: must move this week, useful if capacity remains, blocked or dependent, and not for this week. For each proposed priority, explain the dependency or reason using only supplied information. Do not estimate effort or promise dates that are not in the notes. End with questions a human owner must answer.
Replace every bracketed field with your real context. Read the output before reuse.
Four steps that keep the result usable.
- 1
Define one outcome that makes the week meaningful.
- 2
List hard deadlines and dependencies separately from preferences.
- 3
Let the model group work without pretending it can choose trade-offs for you.
- 4
Confirm the final priorities with the person who owns the trade-off.
Priorities need a limit
A crowded list becomes useful only when someone decides what will not be attempted. Asking the model to label work as blocked, optional, or out of scope makes the limit visible instead of turning every task into a vague priority.
Dependencies change the sequence
A task may sound urgent but still depend on a decision, source, or person. Put those dependencies in the input so the first draft can surface why an apparently simple sequence may not be credible.
Keep trade-offs with the owner
AI can help compare options and state constraints, but it cannot decide which relationship, risk, or opportunity matters most this week. Use the output to prepare that conversation, not to avoid it.
A worked example: a support engineer's week
A list with a vendor incident, a long-open ticket, a documentation task, and two review requests is crowded. Grouping by “must move”, “if capacity remains”, and “blocked” shows that the incident and the review requests depend on other people. The weekly plan then names one outcome — keep the incident moving — instead of claiming all four are priorities.
Treat “blocked” as information, not failure
A blocked task with a named dependency is more useful than a vague “pending”. When the list marks what is waiting on whom, the weekly plan can surface the real question: do we wait, escalate, or drop it? That is a human decision, not something a model should decide for you.
Run a human check.
- Does the plan name one outcome rather than a long list of activities?
- Are blocked tasks and missing dependencies visible?
- Did the draft avoid creating new dates or effort estimates?
- Has the person who owns the trade-off confirmed the final focus?
AI is strongest here when it makes missing information, structure, and options easier to see. The moment an output becomes a claim, commitment, or decision, bring a person back into the loop.