A weekly planner template should protect attention and make commitments visible. For human-agent work, it must also separate AI proposals from approved tasks, reserve capacity for review and exceptions, and connect Friday’s evidence to next week’s decisions.

Copy this weekly planner template
Week of [date]
Weekly contract
Field | Value |
|---|---|
Week ID | WK-2026-30 |
Week owner | [Person] |
Planning window | Monday–Friday |
Available capacity | [Hours] |
Planned focus capacity | [Hours] |
Review and coordination | [Hours] |
Buffer | [Hours] |
Maximum active outcomes | 3 |
Agent policy | Read / Propose / Approved execution |
Friday review | [Date and owner] |
The short answer
A useful weekly planner has eight parts:
three or fewer measurable weekly outcomes;
an honest capacity budget;
one intake queue for human and agent proposals;
an approval gate before work becomes committed;
daily focus blocks with owners and evidence;
a visible exception queue;
a Friday outcome review;
a decision record for what carries forward.
The week is not successful because every box is checked. It is successful when the intended outcomes move, important exceptions are handled, and the next decision uses current evidence.
Weekly outcomes
Outcome | Why this week | Evidence of success | Owner | Confidence |
|---|---|---|---|---|
Increase activation by 8% in pilot cohort | Launch decision Friday | Dashboard and cohort query | Mina | 70% |
Reduce reset-ticket backlog by 15% | Support SLA risk | Queue snapshot | Alex | 85% |
Ship approved self-service reset flow | Required for pilot | Release and acceptance evidence | Priya | 75% |
Write outcomes as changed states, not activities.
Weak: “Work on onboarding.”
Stronger: “Pilot users complete onboarding without assisted setup, with activation rising from 42% to 50% and no permission-error increase.”
Capacity budget

Capacity type | Hours | Rule |
|---|---|---|
Committed human work | 31 | Named owner accepted it |
Approved agent-assisted work | 14 | Exact proposal reviewed |
Review and coordination | 10 | Meetings, approvals, decisions |
Buffer | 9 | Interruptions and uncertainty |
Total | 64 | Must not exceed available capacity |
Start with real availability after holidays, support duty, meetings, and existing commitments. Do not build an 80-hour plan for a 40-hour week.
A simple policy:
focus capacity =
available hours
- fixed meetings
- operational duty
- review time
- risk bufferLimit planned focus work to the remaining capacity. Buffer is not unused time; it is protection against uncertainty.
Work intake
Use one Tasks database related to Weeks, Projects, and Outcomes.
Property | Type | Required |
|---|---|---|
Task ID | ID | Yes |
Task | Title | Yes |
Status | Status | Yes |
Week | Relation | Required when committed |
Outcome | Relation | Yes |
Project | Relation | No |
Source | Select | Human / Agent / Meeting / Incident / Recurring |
Proposal state | Select | Proposed / Approved / Rejected / Deferred |
Owner | Person | Required when approved |
Estimate | Number | Yes |
Priority | Select | Yes |
Due or focus date | Date | Yes |
Acceptance evidence | Text or URL | Yes |
Dependencies | Relation | No |
Risk | Select | Yes |
Agent policy | Select | Human only / Agent assist / Approved execution |
Last verified | Date | No |
The Week relation should be empty until work is accepted. A proposal appearing in the inbox is not a commitment.
Human and agent lanes
Human-committed work
Tasks accepted by a named person. The owner understands the expected outcome, time cost, dependencies, and evidence of completion.
AI-proposed work
Suggestions generated from goals, projects, meetings, support data, or knowledge gaps. They remain Proposed until reviewed.
Approved agent-assisted work
A human approves the task and the allowed agent behavior: research, draft, analyze, update a field, create a bounded artifact, or execute a specific tool.
Deferred or rejected work
Preserve the reason. This prevents the same low-value suggestion from returning every day without new evidence.
Approval gate
Before moving work into the week, validate:
linked outcome or explicit operational necessity;
owner;
estimate;
acceptance evidence;
source;
priority;
dependency state;
permission and risk class;
agent authority;
available capacity;
material external commitment.
The reviewer can approve, edit and approve, reject, or defer. The system should record reviewer identity and time.
Daily plan
Monday — commit and remove ambiguity
confirm outcomes and capacity;
approve or defer intake;
validate dependencies;
reserve focus blocks;
identify decisions needed before Wednesday.
Tuesday — protect execution
work on the highest-risk outcome;
review early evidence;
stop low-value work before it consumes the week;
route missing access or ownership to exceptions.
Wednesday — midpoint decision
compare outcome evidence with the plan;
review capacity used and remaining;
re-estimate uncertain work;
decide what to stop, change, or escalate.
Thursday — close critical loops
finish acceptance evidence;
reconcile downstream state;
resolve launch or handoff decisions;
prepare Friday review.
Friday — verify and learn
measure outcomes against targets;
inspect guardrails and exceptions;
record completed work with evidence;
review agent proposal quality and impact;
decide what carries, stops, or becomes next-week intake.
Daily focus record
Field | Example |
|---|---|
Date | Monday |
Focus block | 09:00–11:00 |
Task | Customer interviews |
Outcome | Pilot activation |
Owner | Mina |
Agent role | Prepare evidence brief |
Interruption policy | Incident only |
Completion evidence | Interview notes and synthesis |
Result | Completed / Changed / Blocked |
Do not use the daily view as another place to copy tasks. It should be a filtered view of the canonical Tasks database.
Agent planning contract
Agents may
retrieve authorized goals, projects, tasks, meetings, and outcome evidence;
identify overdue, blocked, unowned, or duplicate work;
propose weekly outcomes and tasks;
estimate using declared historical data;
draft a capacity scenario;
prepare daily briefs;
summarize evidence for the Friday review;
reconcile approved changes.
Agents must not
commit a person’s time;
invent an owner, estimate, deadline, or dependency state;
move a proposal into the committed week;
approve their own work;
change external promises;
mark Done without evidence;
silently carry incomplete work;
hide an exception or failed write.
Planning packet
An AI proposal should be reviewable as data:
{
"proposal_id": "WKP-204",
"week_id": "WK-2026-30",
"task_id": "TASK-4821",
"outcome_id": "OUT-114",
"proposed_owner": "Mina",
"estimate_hours": 2,
"reason": "Three pilot interviews are needed before Friday's launch gate",
"evidence_ids": ["PRD-104", "DEC-071"],
"dependencies": [],
"risk": "low",
"agent_action": "prepare_interview_brief",
"requires_approval": true
}Validate IDs, estimate, owner availability, dependency state, permissions, and capacity before approval.
Capacity controls
Plan to less than 100%
A week without buffer assumes no ambiguity, interruption, review delay, or failed automation. The assumption is usually false.
Separate effort from elapsed time
A two-hour task blocked on legal review may consume little effort and still threaten the outcome.
Reserve review capacity
Agent-assisted work can reduce drafting time while increasing review volume. Budget approval and verification explicitly.
Count operational duty
Support, incident response, hiring, and customer commitments consume real capacity even when they are not roadmap tasks.
Re-plan visibly
If priorities change, record the removed item, replacement, decision owner, and reason. Do not pretend the original plan still exists.
Exception queue
Create a view for:
approved task without owner;
task without estimate;
task without acceptance evidence;
capacity over plan;
blocked dependency;
permission missing;
agent proposal awaiting review;
agent write rejected;
destination state mismatch;
external commitment at risk;
overdue decision;
incomplete work proposed for carryover;
outcome without current evidence.
Each exception needs severity, owner, next action, due time, and verification.
Carryover policy
Do not automatically copy incomplete work into the next week.
For each item:
verify the current state;
record why it did not finish;
decide whether the outcome still matters;
update the estimate and dependencies;
compare against new capacity;
approve, change, defer, or stop;
preserve the original week and decision.
Carryover without review hides planning errors and makes every week start over capacity.

Friday review template
Outcome scorecard
Outcome | Target | Actual | Evidence | Decision |
|---|---|---|---|---|
[Outcome] | [Target] | [Actual] | [Link] | Continue / Change / Stop |
Work completed
List only work with observable evidence.
Work changed or stopped
Record the decision, owner, and reason.
Exceptions
unresolved:
newly created:
resolved with proof:
Capacity
planned:
used:
review:
buffer consumed:
unplanned work:
Agent contribution
proposals reviewed:
proposals accepted:
hours saved or shifted:
errors or exceptions:
useful evidence:
policy changes needed:
Next-week decisions
carry forward:
return to intake:
stop:
new evidence required:
decision owner:
Weekly planner views
This week
Approved tasks related to the current Week.
Outcomes
The maximum three outcome records and their live evidence.
My focus
Current owner’s committed tasks grouped by day.
Agent proposals
Proposed items not yet authorized.
Approval queue
Proposals with reviewer, risk, capacity impact, and needed-by time.
Blocked and at risk
Dependencies, external commitments, and exceptions.
Capacity
Hours grouped by person, source, and work type.
Friday review
Outcomes, evidence, completed tasks, exceptions, and carryover decisions.
Future intake
Valid work not committed to this week.
Repeating weekly templates in Notion
Notion database templates can repeat weekly. Use a recurring Week template to create the standard review structure and default properties.
Be careful with relations: Notion warns that a relation filled inside a database template will point every created page to the same existing record. Leave week-specific relations empty or populate them after creation.
Notion also supports nested databases in repeating weekly, monthly, or yearly templates, with documented nesting limits. Prefer filtered linked views of canonical databases over copying entire task lists into each week.
Setup steps
Create Weeks, Tasks, Outcomes, and Decisions databases.
Relate Tasks and Outcomes to Weeks.
Add Proposal state and Agent policy to Tasks.
Create the weekly page template.
Add filtered linked views for outcomes, committed tasks, proposals, exceptions, and review.
Configure the Week template to repeat weekly.
Define the capacity and approval policies.
Test that week-specific relations do not reuse old records.
Test agent proposals remain outside the committed view.
Test Friday carryover requires a decision.
Common planner mistakes
The week is a flat checklist
Connect tasks to outcomes, capacity, owners, and evidence.
AI suggestions look committed
Use a separate Proposed state and approval queue.
Capacity ignores reviews
Budget time to inspect agent output and approve consequential work.
Every task has Friday as its due date
Use focus dates, dependencies, and actual external commitments.
Incomplete work rolls forward automatically
Require a carryover decision.
The planner duplicates source tasks
Use relations and filtered views of canonical records.
Done means a checkbox
Require observable acceptance evidence.
No buffer
Plan uncertainty explicitly.
Templates prefill the wrong relation
Leave week-specific relations empty inside repeating templates.
Friday review is a retrospective essay
Use evidence and decisions: outcomes, capacity, exceptions, continue/change/stop.
When Notion is enough
Notion can support a weekly planning system with Tasks databases, relations, filtered views, database templates, and repeating weekly templates. Its official guidance frames database templates as a way to standardize recurring processes and prefill structures and properties.
Add a dedicated scheduling or capacity system when you need complex resource optimization, shift coverage, time tracking, or transactional workforce planning. Keep the weekly outcomes, evidence, decisions, and exceptions connected.
Frequently asked questions
What should a weekly planner template include?
Include weekly outcomes, available capacity, committed work, proposals, owners, estimates, daily focus, dependencies, acceptance evidence, exceptions, Friday review, and carryover decisions.
How many weekly priorities should I have?
Use three or fewer outcome-level priorities. Tasks can be more numerous, but they should support those outcomes or explicit operational obligations.
Should AI plan my week?
AI can retrieve context, propose work, detect gaps, and prepare scenarios. People should authorize commitments, owners, deadlines, external promises, and major changes.
How much buffer should a weekly plan have?
Use historical interruption and uncertainty data. If none exists, start with a visible buffer and adjust through weekly evidence rather than planning to 100%.
Should unfinished tasks roll over automatically?
No. Revalidate importance, estimate, dependencies, and capacity, then explicitly carry, change, defer, or stop.
Can Notion create a new weekly planner automatically?
Yes. Repeating database templates can create pages weekly. Test relations and filters so each new page points to the correct week.
What is the difference between a weekly outcome and a task?
An outcome is a measurable changed state. A task is a bounded action that contributes to it.
How should agent work be measured?
Track accepted proposal rate, review time, errors, exceptions, verified output, and whether human attention shifted toward higher-value decisions.
Last verified: July 21, 2026.
Use the reusable template
Duplicate the weekly planner below and separate human commitments, AI proposals, and approved agent-assisted work. Capacity and approval must remain visible.
Sources
Product facts were verified on July 21, 2026.
Where Dokki fits
A Dokki weekly planner can combine outcomes, capacity, daily focus, delegated agent work, approvals, exceptions, and Friday evidence in one table. The planner becomes a reviewable operating system rather than a list that resets without learning.
