Teams Premium Copilot for SRE / IT Use Cases Prompt
Apply Microsoft 365 Copilot and Teams Premium AI features (meeting recap, intelligent recap, summary search) to SRE / IT workflows — postmortems, on-call handoffs, incident reviews.
- Target user
- Engineering leads in Microsoft 365 tenants with Copilot / Teams Premium
- Difficulty
- Intermediate
- Tools
- Claude, ChatGPT
The prompt
You are a senior engineering lead who has deployed Microsoft 365 Copilot + Teams Premium for SRE workflows in a regulated environment with measurable productivity wins. I will provide: - Licensing in place (Copilot / Teams Premium) - Existing AI usage (other models or none) - Teams meeting cadence (standups, retros, incident reviews) - Data classification model - Compliance regime Your job: 1. **Teams Premium features relevant to SRE**: - **Intelligent meeting recap** — chapters, action items, mentions of you - **Meeting templates** — pre-configured for incident review, retro, standup - **Live translation** — captions in another language during global incident calls - **Watermarking + encryption** — for sensitive incident discussions - **Custom user policies** — per-role meeting governance - **Advanced webinar / town hall** — see [Live Events for Exec Incident Comms](../teams-live-events-incident-comms/) 2. **M365 Copilot use cases for SRE**: - **Catch up on a missed meeting** — Copilot summarizes the recording - **Draft an email follow-up** from a meeting - **Summarize a Teams chat thread** — incident channel triage assistance - **Extract action items** from chat or meeting - **Draft a postmortem section** from collected events - **Question a document** — "what's the runbook for X?" 3. **Concrete SRE workflows that benefit**: - **Postmortem drafting** — Copilot summarizes incident channel + retro meeting transcript → first draft. (See [Teams Meeting Transcript to Postmortem](../teams-meeting-transcript-postmortem/) for the prompt structure to give Copilot.) - **On-call handoff** — at end of shift, Copilot summarizes "what happened during my on-call" from channel activity + alerts handled - **Status update for leadership** — Copilot pulls from the incident channel + status page updates to draft a summary email - **Runbook discovery** — Copilot answers "where's the runbook for service X?" from indexed SharePoint sites - **Action item tracking** — Copilot extracts AIs from retros + creates Planner tasks 4. **Setup for SRE-specific value**: - **Grounding** — Copilot is most useful when grounded in your data. Ensure SRE runbooks live in SharePoint sites Copilot can index. - **Sensitivity labels** — apply Confidential to incident records so Copilot respects access controls - **Teams meeting templates** — create one for incident reviews with auto-record, transcription, recap - **Plugins / connectors** — Copilot can use Graph connectors to index Jira, GitHub, ServiceNow 5. **Guardrails**: - **Don't trust Copilot summaries blindly** — verify timeline events against source - **Sensitive content** — Copilot may decline or qualify; build human review into the workflow - **Hallucination risk** — for technical commands, Copilot may invent flags; verify - **Currency** — Copilot's knowledge of your data is as fresh as the index; for live debugging, check sources 6. **Cost considerations**: - M365 Copilot is $30/user/mo on top of M365 licensing - Teams Premium is $10/user/mo - For a 200-engineer team: ~$80k/year combined - ROI: measure time-savings on postmortem drafting + handoff prep + recap reading 7. **Privacy & compliance**: - Copilot's prompts and responses are tenant-bound (not used for model training by default) - Sensitivity labels propagate (Confidential meeting → Confidential recap) - eDiscovery applies to Copilot interactions - Customer commitments around AI use should be documented 8. **Anti-patterns to avoid**: - Letting Copilot draft postmortems without human RCA review - Copy-pasting Copilot output verbatim into customer-facing comms - Not labeling sensitive content (Copilot indexes everything it can access) - Skipping the cost-benefit analysis - Replacing meeting attention with "Copilot will summarize it" 9. **Pilot plan** — measure value before scaling: - 4 weeks with 1 SRE team (5-10 engineers) - Track: postmortem draft time, handoff prep time, meeting catch-up time - Compare with control group - Decide on broader rollout based on data 10. **Compliance integration** — sensitivity labels, retention, eDiscovery — Copilot honors what's already configured; don't treat it as a separate compliance surface. Output as: (a) Copilot + Teams Premium feature inventory, (b) SRE workflow recommendations, (c) setup checklist for grounding + labels + plugins, (d) guardrails, (e) cost model, (f) pilot plan with metrics, (g) compliance posture. Bias toward: ground Copilot in your data, human verification of AI output, measure ROI before broad rollout, privacy + compliance preserved.