What Agent draws on
- Notes captured during the week
- Calendar history and upcoming commitments
- Project updates and task-board snapshots you add to Mem
- Goals and review preferences
Your week already left a trail. Agent turns it into a plan.
The Agent loop
Agent keeps the thread alive, then brings it back to you for judgment and action.
You verify which commitments are truly open and approve the priorities that carry into the next week.
Mem Agent can turn a week of scattered activity into a clear review and a short plan for what comes next. It reads the context available in your notes, messages, calendar, and saved project updates, then separates meaningful progress from activity and unresolved commitments from things already handled.
A useful review is not a longer task list. It should answer four questions:
Agent can group notes that refer to the same project, recognize that a promise in Monday’s meeting is still unresolved on Friday, and compare the week with goals you wrote earlier. If you save a task-board snapshot or status update in Mem, it can incorporate that too. It does not silently change the external board or claim that every inferred action item is correct.
Imagine your week includes six meetings, a handful of quick captures, a project update, and several changes to next week’s calendar. Agent might identify that a launch date moved, two customer questions point to the same risk, and a promised document never received a final decision. It can then propose a plan with three outcomes instead of repeating 25 activities.
Ask for evidence beside each recommendation. “Prioritize the pricing decision because it blocks the launch brief and came up in two meetings” is far more useful than “Work on pricing.” Evidence also makes the human checkpoint fast: you can correct a misunderstanding before it becomes next week’s plan.
The one-question weekly review is a simple starting ritual. With Agent, the same ritual can continue week after week rather than becoming an isolated summary.
A one-off chatbot can summarize a weekly dump, but it cannot reliably know why an item mattered last week, which constraint you chose, or whether a follow-up was later completed unless you supply that history again. Mem Agent works across accumulated context. It can remember the rationale, ask about an unresolved commitment, and stop following up after you mark it handled.
This is especially useful for projects that move slowly. A small decision from three weeks ago may explain today’s delay. The review can reconnect those points without requiring you to reread every note.
Before accepting the plan, verify the source material, remove stale or speculative tasks, and decide how much capacity you actually have. Make task-board, calendar, and communication changes yourself. Agent’s role is to assemble the evidence and prepare the decision. Your role is to decide what becomes a commitment.
A simple playbook
Give Agent your current goals, recurring responsibilities, and the questions your review should answer.
Agent reviews recent context, groups related developments, and identifies open loops without requiring a perfect task system.
Review the proposed priorities, correct any assumptions, and choose the commitments that should carry forward.
What this can include
Starter prompt
You do not need a perfect command. This is enough to start the thread.
“Every Friday, review my notes, calendar, and project updates from the past seven days. Summarize meaningful progress, decisions, unresolved commitments, and risks. Then propose no more than five priorities for next week, with the evidence behind each. Ask me before carrying forward anything ambiguous.”
Where you stay in control
Mem Agent can review, organize, and draft a weekly plan. It does not silently update an external task board, reschedule meetings, or report status to other people.
Your checkpoint:You verify which commitments are truly open and approve the priorities that carry into the next week.
Run a one-question weekly reviewFrequently asked questions
Clear answers about what Mem Agent can do, what it needs, and where you stay in control.
It is a synthesis of recent notes, calendar context, decisions, progress, and unresolved commitments, followed by a proposed plan for the next week.
It can surface commitments mentioned in context available to it, such as meeting notes or messages with Agent. You should verify that each inferred task is real and still relevant.
Not by default. You can save task-board updates in Mem for review, then make or approve changes in the system that owns the work.
Mem Agent can retain the reasoning behind priorities and follow unresolved threads across weeks. A one-off chatbot only sees the snapshot you paste into that session.
The goal is a short review you can actually finish. Ask for a concise synthesis, inspect the evidence, and spend your time choosing rather than collecting.
Yes. Tell Agent whether to combine them into one capacity plan or keep separate sections while still flagging conflicts between them.
Confirm that it is handled. Agent can then stop carrying that item into future reviews instead of repeatedly resurfacing it.