1. Choose priorities from a crowded list

Use this when everything feels equally urgent. Include deadlines and who is waiting on each task; a title alone rarely explains its importance.

YOUR STARTING POINT

Here are my open tasks: [tasks, deadlines, and dependencies]. My main outcome this week is [outcome]. I have [available time] today. Suggest one main priority and up to two supporting tasks. Explain what should wait and why. Ask me about missing information that would change the order.

Check whether the answer distinguishes a real deadline from a preferred date. If it gives every task a high priority, ask it to make an actual choice. You are looking for a decision you can assess, not more labels.

2. Check whether today’s list fits

A capacity check is useful even if you prefer to arrange the calendar yourself. Include breaks and preparation in the available-time calculation.

YOUR STARTING POINT

I work from [start] to [stop]. Fixed commitments are [times], with [preparation or travel]. Tasks and estimated durations are [list]. Calculate the remaining time and compare it with the workload. Leave [buffer] unallocated. Flag any overlap, missing estimate, or dependency. Do not shorten task estimates just to make everything fit.

Verify the arithmetic. A plan that contains five hours of work in four free hours still needs a tradeoff, however confidently the assistant presents it.

3. Find a smaller starting point

A task like “sort out the launch” can hide several decisions. Ask for a starting action with a visible result.

YOUR STARTING POINT

The project is [outcome], and I am stuck at [current state]. I have [minutes] available. Suggest three possible first actions, each with a concrete output I can reach in that time. Identify what each action would clarify. Do not add new project scope.

Pick the action that removes uncertainty or unlocks another step. You do not need to do all three. Add the chosen action to the project so the suggestion becomes part of your working plan.

4. Replan an interrupted afternoon

Use the current time as the starting point. Otherwise the assistant may offer a schedule containing blocks that have already passed.

YOUR STARTING POINT

It is now [time]. I stop at [time]. I have completed [work], and my remaining fixed commitments are [times]. Unfinished tasks are [tasks and remaining estimates]. Help me choose what to finish, what to move, and what to communicate. Preserve a shutdown buffer. Show a plan beginning now.

Read “remaining estimates” literally. A task that originally needed an hour may have only fifteen minutes left. Tell the assistant what is unfinished instead of asking it to repeat the original allocation.

5. Prepare for a meeting without inventing context

Provide the agenda or notes you are comfortable sharing with your AI provider. Keep factual preparation separate from guesses about another person’s intentions.

YOUR STARTING POINT

My meeting is about [topic]. The confirmed agenda and relevant notes are [context]. I have [minutes] to prepare. Identify the decisions we need, questions I should bring, and materials to have ready. Mark anything that is an assumption. Use only the context I supplied.

Check the proposed questions against the meeting’s purpose. A short status call should not turn into an unsolicited strategy workshop.

6. Close the day with a clear restart

End with a small, usable handoff to yourself.

YOUR STARTING POINT

Today I completed [work]. The unfinished items are [list], and the current blockers are [blockers]. Help me write a short shutdown note: what moved forward, what must be captured, and the first useful action for my next working day. Do not automatically schedule the entire backlog for tomorrow.

In Calendo, keep the restart note with the relevant task. Connected AI chat requires setup in Settings; these prompts are templates, not promises that every AI connection will take every requested action. For a complete schedule example, read how to plan your day with AI.

These guides use illustrative examples. AI helped draft the content; product descriptions are grounded in Calendo’s current features. No example is a promise of a particular result.

How we write these guides