> ## Documentation Index
> Fetch the complete documentation index at: https://docs.flashduty.com/llms.txt
> Use this file to discover all available pages before exploring further.

# AI-Assisted Post-Mortem Generation

> Let AI generate a post-mortem report draft based on your incident data, dramatically reducing the manual effort of writing from scratch.

## Let AI Write the First Draft

The hardest part of writing a post-mortem is often getting started — staring at a blank page, unsure where to begin.

Flashduty's AI-assisted generation can help you get past that. It analyzes all the context around the linked incident and, following the template's section structure, produces a complete first draft. All you need to do is adjust the wording, fill in details, and add your team's conclusions.

<Tip>**Plan requirement**: This feature requires an On-call Pro or higher subscription. [Learn more](https://flashcat.cloud/flashduty/price/)</Tip>

## How to Use

<Steps>
  <Step title="Create a Post-Mortem Report">
    Create a post-mortem report. AI will begin generating content and display it in the editor in real time.
  </Step>

  <Step title="Refine the Draft">
    Once generation is complete, you can edit, trim, or add to the content directly in the editor. Other team members can join and edit simultaneously.
  </Step>
</Steps>

## What Data Does AI Use

AI doesn't write in a vacuum — it analyzes the following data to generate the report:

| Data Source        | Description                                                               |
| :----------------- | :------------------------------------------------------------------------ |
| Template structure | Organizes content according to the sections in your chosen template       |
| Incident details   | Labels, custom fields, severity, responders, and more                     |
| Response timeline  | Actions taken and status changes during incident handling                 |
| War room chat logs | Discussion content from Feishu or Slack war rooms related to the incident |

<Note>
  To provide the AI with complete incident lifecycle context for generating post-mortem reports, your IM war room integration requires additional permissions to read war room chat logs.

  1. Feishu integration requires the `im:message.group_msg` scope.
  2. Slack integration requires the `channels:history` scope.

  Please update your app permissions or reauthorize. Refer to the integration documentation for details. Due to platform API limitations, WeCom and Dingtalk do not currently support war room chat log extraction.
</Note>

## Usage Limits

To ensure service quality, AI generation has the following limits:

| Constraint           | Details                                     |
| :------------------- | :------------------------------------------ |
| Generation frequency | Up to 5 times per minute per account        |
| Concurrency          | Only 1 generation task per report at a time |

If you see a rate limit message, wait a moment and try again. If a message indicates someone else is generating, wait for them to finish.

## Tips for Better AI Output

* **Pick the right template**: AI strictly follows the template's section structure. A well-organized template leads to a well-organized report.
* **Enrich your incident context**: Labels, custom fields, and the response timeline are the primary inputs for AI analysis. The more complete they are, the deeper the generated report will be.
* **Leverage your IM integration**: If your team discussed root causes and remediation in a Feishu or Slack war room, those conversations will be factored into the AI's analysis for a more accurate report.
* **Treat AI as a starting point, not the finish line**: The AI draft is a great foundation, but the final report should include your team's in-depth discussion, genuine root cause consensus, and actionable improvement plans.

## Time Display

Time information in the post-mortem report automatically adapts to your timezone settings, ensuring team members in different regions see accurate local times.