Atlassian Intelligence create post-incident reviews

Alright, the fire is out. The system is back online. Now what?

Once the immediate crisis is over, the real work begins. The post-incident review (PIR) is where you figure out what actually happened, why it happened, and how to stop it from happening again. It's a critical step, but let's be honest, it can be a drag to put together. Atlassian is trying to ease some of that pain by building AI into Jira Service Management to help generate these reports.

This guide will give you a straight-up look at how to use Atlassian Intelligence to create post-incident reviews. We'll cover how it works, what it's good at, where it stumbles, and what it’ll cost you. We'll also touch on what to do when you need a bit more power and control than the built-in tools can offer.

The basics of Atlassian Intelligence for post-incident reviews

Before we jump into the AI side of things, let's make sure we're on the same page about the key pieces.

What is a PIR?

A post-incident review (or postmortem, if you prefer) is a meeting and a document. It’s the process your team uses to break down an incident after it’s been fixed. The point isn't to play the blame game; it's to get a clear timeline of events, understand the impact, and pinpoint the root cause. As Atlassian themselves put it, the whole idea is to learn from the incident so you can make your services more robust.

What is Atlassian Intelligence?

Atlassian Intelligence is the AI that Atlassian has sprinkled across its cloud products, like Jira Service Management, Confluence, and Bitbucket. It’s a mix of their own AI models and some tech from OpenAI. It uses your organization’s data (what they call a "Teamwork Graph") to understand your projects and teams, which helps it provide slightly more tailored assistance. One of its specific jobs in Jira Service Management is helping to draft those PIRs.

How to use Atlassian Intelligence to create post-incident reviews

Atlassian has kept the process pretty simple, which is great for teams already living in Jira Service Management (JSM). The feature’s main job is to spit out a summary of the incident to give you a head start.

Here’s how it works, in a nutshell:

1. Find your incident: In your JSM project, go to "Incidents" and click on the one you need to review.

2. Start the PIR: Look for the Add PIR button on the ticket. This will open up the creation screen.

3. Let the AI do its thing: You'll see a Suggest description button in the description field. Click it, and Atlassian Intelligence will generate a summary.

4. Clean it up: The AI-generated text will appear in the description box. Now it's on you to read it, edit it, and fill in any gaps before you finalize the PIR.

The AI pulls its information from the incident ticket and any associated Slack channel you might have created.

While that’s a decent starting point, real incidents are messy. The full story is often scattered everywhere, from design docs in Google Docs to internal guides in Notion. For a PIR to be genuinely useful, your AI needs to see everything. This is a big blind spot for Atlassian's tool, but it's exactly what platforms like eesel AI are built for. It connects to over 100 different sources to give you the complete picture, not just the Jira slice.

The good, the bad, and the pricey

Atlassian Intelligence has its moments, but it's important to be realistic about its limitations before you commit.

The good stuff: Convenience and integration

The biggest win here is that it's built right into JSM. If your team is all-in on Jira, it feels seamless.

• Quick summaries: It saves your team from staring at a blank page by drafting an initial summary based on the ticket and Slack chatter.

• Everything in one place: The PIR gets saved as a work item in JSM and can be easily pushed to Confluence, which helps keep all your incident docs together.

• Simple to use: There's no complicated setup. Anyone on the team can click the "Suggest description" button and get a result.

Where it gets tricky: A lack of control

For all its convenience, the feature has some serious drawbacks that can get in the way of a proper review.

• It only sees what's in Atlassian: The AI is stuck inside its own little world. It can’t pull in crucial context from your internal wikis, Microsoft Teams chats, or monitoring tools. A PIR is only as good as its inputs, and this siloed approach can lead to some pretty big gaps.

• You can't tell it what to do: The AI summary is a black box. You can’t tweak the prompt, adjust the tone, or give it a template to follow. If your company has a specific way of doing PIRs, this one-size-fits-all approach probably won't cut it.

• No way to test it: You can't run the AI on old incidents to see how well it works. You're essentially testing it on a live incident, which isn't ideal when you need accurate documentation.

These are the kinds of limitations that separate a basic feature from a dedicated platform. A tool like eesel AI is designed to handle this complexity. You get a workflow engine that you can actually control. You can connect all your knowledge sources, edit the AI’s persona with a prompt editor, and even run it on thousands of past tickets in a simulation mode to make sure it's ready for prime time.

Pricing: What's the damage?

Cost is always a factor, and getting access to AI features like the ability to create post-incident reviews isn't free. You'll need to be on one of Jira Service Management's higher-tier plans.

If you're on a Free or Standard plan, you're out of luck.

Here’s a quick look at the public pricing for Jira Service Management Cloud (per agent, per month) to give you an idea:

Heads up: Pricing can change. Always check the official Atlassian pricing page for the latest numbers.

What this means is you might have to pay for a big platform upgrade just to get one AI feature, along with a bunch of other things you may not need. This is a pretty different model from solutions like eesel AI, which offers straightforward pricing based on the AI features you actually use. There are no per-resolution fees, so you don't get hit with a surprise bill at the end of the month.

Thinking beyond the summary

Atlassian Intelligence gives you a decent first draft of a PIR, which is a start. But a mature incident management process does more than just write reports. The real goal is to turn what you've learned into concrete actions and automate as much of the grunt work as possible.

Think about the Atlassian workflow: an incident is resolved, so someone has to remember to go in, click 'Add PIR', then click 'Suggest description'. Then they have to read through it, figure out the next steps, and manually create and assign all the follow-up tasks. It helps with one small piece, but the process is still very manual.

Now, imagine a more automated workflow. An incident is resolved, and a trigger automatically kicks off the PIR process. An AI reads through everything, not just Jira and Slack, but Confluence, Google Docs, and any other knowledge base you’ve connected. It then generates a PIR based on a custom template you created. And here's the best part: the AI automatically creates, tags, and assigns the follow-up tasks in Jira. Nothing gets forgotten.

This is the difference between a tool that assists with a manual step and a platform that automates the entire workflow. eesel AI is built for the latter. It connects to your systems, generates PIRs based on your rules, and uses AI Actions to create and assign those crucial follow-up tasks, making sure lessons learned actually lead to improvements. You can even simulate the whole thing on past incidents to get it just right before you turn it on.

The verdict: Is it enough for you?

For teams that are deep in the Atlassian ecosystem, using Atlassian Intelligence to create post-incident reviews is a nice little add-on. It's convenient, it simplifies that first draft, and it keeps everything inside JSM.

But if your team is serious about operational maturity, you’ll probably outgrow it fast. Its inability to access outside data, the lack of customization, and the manual steps required after the PIR is generated are all major ceilings. When you need to connect all your team's knowledge, control how your AI behaves, and automate the follow-up that prevents future incidents, you’ll need to look beyond the built-in features.

See how eesel AI can plug into your JSM setup and give your team the power and flexibility it needs. You can get it running in minutes and start building a more resilient incident management process today.