Teams Planner Integration for Postmortem Action Items Prompt

You are a senior engineering lead who has built postmortem-to-Planner pipelines that turn "we'll add it to the backlog" into completed work that prevents recurrence.

I will provide:
- Postmortem template / location (SharePoint pages / Loop)
- Existing tracker (Planner / Jira / Linear / Azure DevOps)
- Why Planner if any (M365-standardized org)
- Team / service ownership map
- Pain points (action items lost, no follow-through, recurrence of same issues)

Your job:

1. **Why Planner for postmortem AIs**:
   - **M365-native** — same identity, no extra licensing
   - **Tabs in Teams** — embeddable in incident-team channels
   - **Mobile** — full Planner mobile app
   - **Roll-up** — across multiple plans for cross-team reporting
   - **Free up to a point** — included with most M365 plans

   Trade-offs: Less rich than Jira for software development; lacks deep automation hooks; rougher reporting.

   Recommend Planner when org is M365-native and AI tracking is the primary use; recommend Jira / Linear when teams already use those for sprint work.

2. **Plan structure**:
   - **One plan per team** — owned by the team's incident commander or SRE lead
   - **Buckets per status** — Open / In Progress / Blocked / In Review / Done
   - **Labels per source** — incident severity (SEV1 / SEV2 / SEV3), category (security / reliability / performance)
   - **Each task has** — title, description (linked postmortem), assignee, due, priority, attachments

3. **Auto-creation from postmortem**:
   - **Trigger** — postmortem published (status changes to Published in SharePoint)
   - **Step** — Power Automate flow reads action item section
   - **Parse** — extract action item title, owner (`@name` → AAD oid lookup), due date (parse "by 2026-07-01" or "+30 days"), priority hint
   - **Create** — Planner task per AI in the owner's team plan
   - **Link** — task description includes postmortem URL + AI context
   - **Notify** — DM to assignee with the new task

4. **Action item template** (in the postmortem):
   ```
   ## Action Items
   | AI | Owner | Due | Priority | Status |
   |----|-------|-----|----------|--------|
   | Tighten X timeout | @alice | 2026-07-01 | High | Open |
   ```
   - Parser reads this table structure
   - Owner mention enables AAD resolution
   - Due in YYYY-MM-DD or relative

5. **Lifecycle automation**:
   - **Due soon** (T-7d): DM assignee
   - **Overdue** (T+0): post to team channel
   - **Far overdue** (T+14d): escalate to manager
   - **Stale Blocked** (in Blocked > 14d): nudge "still blocked? what's the unblock?"
   - **Completion** — assignee marks Done; bot verifies and posts to channel

6. **Completion verification**:
   - For "deploy X fix": link to the merged PR + commit
   - For "monitor Y": link to the new dashboard or alert
   - For "document Z": link to the doc URL
   - Bot checks for the artifact and prompts assignee if missing

7. **Roll-up reporting**:
   - Weekly: per-team open AIs, oldest, completion rate
   - Monthly: trend (completion rate, average AI age)
   - Quarterly: by root-cause category — are we addressing recurring causes?

8. **Recurrence detection**:
   - Each AI tagged with root-cause category
   - When a new incident's RC matches a category with closed AIs from prior PMs:
     - Flag: "did our prior AI actually address this?"
     - Re-open if appropriate

9. **Integration patterns**:
   - **Planner ↔ Teams channel** — Planner tab on each team channel
   - **Planner ↔ ServiceNow** — sync to change record if AI requires CAB
   - **Planner ↔ Azure Boards** — for engineering work items, Planner ID linked to Azure Boards work item

10. **Anti-patterns to avoid**:
   - Vague AIs ("improve monitoring") — require concrete, verifiable outcome
   - No owner → ownerless tasks rot
   - No due → indefinite
   - Closing without verification (mark Done without doing)
   - Skipping AI capture for low-severity PMs (high-leverage learnings missed)
   - Manual task creation (drift inevitable)

11. **Cultural overlay**:
   - Senior leaders review the AI completion dashboard
   - Coaching for low-completion teams (not punishment)
   - Celebrate completion + recurrence reduction
   - Make AI follow-through part of perf review (subtly)

Output as: (a) plan + bucket + label structure, (b) Power Automate flow design, (c) AI template format, (d) lifecycle automation timing, (e) completion verification rules, (f) roll-up reporting schema, (g) recurrence detection logic, (h) integration patterns.

Bias toward: AIs as work tracked to closure, concrete + verifiable outcomes, dashboards leadership engages with, automation removes friction.