Draft a ServiceNow Incident Description

Context

Incident descriptions written under pressure are often incomplete, technically dense, or focused on what happened rather than what the responder needs to know. A well-structured ServiceNow incident description captures the title, impact, observed symptoms with timing, current status, and next steps in plain language — written for someone who does not know the system. This prompt converts rough notes into a complete, clear draft.

Works with both Gemini (for notes already in Google Docs or Gmail) and build-cli (when you need to reference technical log context alongside the notes).

Security note: Do not include credentials, system passwords, or internal IP addresses in the notes you paste.

The Prompt

Write a ServiceNow incident description from the following notes.

The description should include:
- Short title (under 80 characters)
- Impact: who is affected and how
- Symptoms: what was observed and when
- Current status: what has been done so far
- Next steps / owner

Keep the language factual and clear. Avoid jargon. Write for someone who doesn't know the system.

Notes:
[PASTE YOUR ROUGH NOTES HERE]

Usage Instructions

Choose your tool:
- Gemini: use this directly in Google Docs (Gemini sidebar) or Gmail if the notes are already in a Doc or email thread.
- build-cli: useful if you need to reference a technical log file or codebase context alongside the notes.
Copy the prompt above.
Paste your rough notes after [PASTE YOUR ROUGH NOTES HERE]. Notes can be bullet points, Slack messages, or anything you have at hand.
Review the generated description before submitting — the AI may soften language around severity that should stay strong.

Example Output

The AI will produce a structured description such as:

Title: Daily customer_dim load producing null country_code — 12% of records affected

Impact: The customer segmentation and tax calculation reports downstream of customer_dim are producing incorrect results for approximately 12% of customer records starting from the 15 November load. Business users running these reports are seeing incorrect country groupings.

Symptoms: Observed at 08:15 on 2026-11-15: 4,200 of 35,000 records in the daily load have country_code = NULL. Historical null rate was <0.1% (< 40 records per load). First affected load: 2026-11-15 00:30 UTC.

Current status: Root cause investigation in progress. Pipeline has not been reverted. Downstream reports have been flagged with a data quality warning banner by the BI team.

Next steps / owner: R. Chen to investigate ETL changes deployed 2026-11-12 and identify the change responsible. Estimated update by 2026-11-15 14:00.