Voice for Incident Reports

Dictate incident postmortems while the event is still fresh. Capture timeline, impact, root cause, and remediation without spending hours writing.

The Problem

Incident postmortems are critical for organizational learning but are often delayed for days or weeks because writing them is exhausting, especially after the stress of the incident itself. By the time someone writes it up, timelines are fuzzy, details are lost, and the report loses much of its value.

The Solution

Ummless lets you dictate the incident report immediately after resolution, when every detail is vivid. The Technical Documentation preset structures your narrative into a proper postmortem with timeline, impact assessment, root cause analysis, and action items.

Workflow

  1. Resolve the incident

    Once the incident is mitigated and stable, take a brief break to decompress.

  2. Open Ummless

    Select "Technical Documentation" for structured incident report output.

  3. Walk through the timeline

    Start from when the incident was detected. Cover each major event, decision, and action chronologically.

  4. Explain root cause and impact

    Describe what went wrong, why it happened, and how many users or systems were affected.

  5. List action items

    Dictate the follow-up work needed to prevent recurrence. Assign owners if you can.

  6. Review and share

    Review the report for accuracy, then share it with the team and stakeholders.

Before & After

Raw Transcript

OK so the incident started at like 2:15 PM when our monitoring alerted on a spike in 500 errors. Turns out the database connection pool was exhausted because a query in the search service had a missing index and was holding connections for way too long. It affected about 30 percent of requests for roughly 45 minutes. We fixed it by adding the index and restarting the connection pool. We need to add query timeout limits and better connection pool monitoring going forward.

Refined Output

## Incident Report: Database Connection Pool Exhaustion ### Timeline - **14:15** - Monitoring alert triggered: spike in 500 errors - **14:20** - Investigation began; identified elevated database connection usage - **14:35** - Root cause identified: unindexed query in search service holding connections - **14:45** - Mitigation applied: added missing index, restarted connection pool - **15:00** - Error rates returned to baseline; incident resolved ### Impact - ~30% of requests returned 500 errors - Duration: approximately 45 minutes ### Root Cause A query in the search service lacked a database index, causing long-running queries that held connections from the shared pool. Under load, the pool was exhausted, causing failures across all services. ### Action Items - [ ] Add query timeout limits to the database client configuration - [ ] Implement connection pool utilization monitoring and alerting - [ ] Audit other services for missing indexes on frequently queried columns

Suggested Presets

Technical Documentation

Frequently Asked Questions

Should I write the postmortem during or after the incident?

After resolution and a brief break. During the incident, focus on mitigation. Use Ummless immediately after resolution while details are fresh.

Can multiple team members contribute?

Yes. Each team member can dictate their perspective and timeline. Combine the reports into a comprehensive postmortem that captures the full picture.

Ready to try it?

Download Ummless and start using voice for your workflow.

Related Content