Which AI does Atlassian use? A complete overview for 2024

It seems like every tool we use has suddenly sprouted an AI feature, and Atlassian is no exception. Its entire product suite, from the backlogs of Jira to the pages of Confluence, is getting an "Atlassian Intelligence" upgrade. These features promise to do everything from summarizing endless comment threads to answering questions and tackling boring, repetitive tasks.

But with all the buzz, it’s easy to feel a little lost. What’s actually powering these new tricks? Is it the same AI as ChatGPT? Is it something they built themselves?

This guide cuts through the noise. We’ll get straight to the point and explain exactly which AI models Atlassian uses. We’ll also walk through the features you’ll actually see in your day-to-day tools and talk about some very real limitations you should know about. Most importantly, we’ll cover what to do when your team’s brainpower isn’t just stored in Atlassian products, because let’s be honest, whose is?

Which AI does Atlassian use?

Here’s the short answer: Atlassian Intelligence isn’t a single, all-powerful AI. Think of it more like a team of specialists. It’s a hybrid system that cleverly blends Atlassian’s own in-house tech with some of the biggest names in the large language model (LLM) world. This mix-and-match approach lets them use the right tool for the right job within their ecosystem.

Here’s a look at the players on the team:

• Big-Name LLMs: Atlassian has partnered up with the heavy hitters of AI, including OpenAI (the folks behind ChatGPT), Anthropic, and Google. They tap into various models from the GPT, Claude, and Gemini families to handle the creative, language-heavy lifting, like helping you write a Confluence page or summarizing a Jira ticket. And for those worried about privacy, Atlassian is clear that your company’s data isn’t used to train these external models.

• Atlassian’s Homegrown Models: This is where things get interesting. Atlassian is sitting on over two decades of data about how teams work together, and they’ve used it to train their own specialized AI models. This is the secret sauce behind their "Teamwork Graph," a unique data layer that doesn’t just understand words, but also the relationships between your projects, goals, documents, and teammates. It knows which Jira epic is connected to which Confluence spec, and who on your team is the go-to expert for a particular topic. This context helps the AI give you much more relevant answers.

• The New Teammate, Rovo: Atlassian’s latest product, Rovo, takes all this a step further. It’s designed to be a central AI assistant that can search, answer questions, and even take action across all your Atlassian tools. It also has some ability to connect with outside apps like Google Drive and Slack, acting as a bridge to the outside world.

A breakdown of Atlassian Intelligence features by product

Atlassian Intelligence isn’t some new app you have to download or install. Instead, its features are woven directly into the Atlassian products your team already relies on. The list of features is always growing, but here are some of the most practical ways you can use it right now.

In Jira and Jira Service Management

For anyone living in Jira and Jira Service Management, the AI features are all about cutting down on manual work and speeding up workflows.

• Talk to Jira in plain English: Forget trying to remember the syntax for Jira Query Language (JQL). You can now just type what you’re looking for in simple language. For example, asking, "show me all tickets assigned to my team that are due this week" actually works, saving you from a JQL headache.

• Get the TL;DR on tickets: We’ve all opened a ticket with a description that’s a novel and a comment thread that’s even longer. AI summaries can give you the key points in seconds. This is a lifesaver for support agents or managers who need to get up to speed on an incident without reading every single update.

• A virtual agent for JSM: The virtual agent in Jira Service Management can automatically answer common customer questions by pulling information from your Confluence knowledge base. The goal is to handle the simple, repetitive questions so your human agents can focus their brainpower on the tricky stuff.

This video demonstrates how you can use Atlassian Intelligence to break down epics and create stories in Jira automatically.

But there’s a catch. The virtual agent is only as smart as the knowledge it can access. If a customer asks a question and the answer is buried in a Google Doc or a Slack thread, the bot is stumped. This leads to frustrated customers and an escalation back to your human team, which kind of defeats the purpose.

In Confluence

Over in Confluence, the AI is focused on helping you create, find, and digest information more effectively.

• A writing assistant: Stuck with a blank page? You can ask the AI to draft a project plan, brainstorm ideas for a blog post, or even just change the tone of something you’ve already written from formal to casual. It can also handle translations.

• Summaries for pages and comments: Just like in Jira, you can get an instant summary of a long, dense document or a comment section that’s gone off the rails.

• Q&A search that actually answers: Instead of just pointing you to a list of pages with your keywords, the new Q&A search tries to give you a direct answer to your question, sourced from the content across your Confluence space.

These features are genuinely useful for the information you have inside Confluence. But they can’t see anything else. If you ask a question and the answer lives in Google Docs or Notion, Confluence’s AI will come up empty-handed. This reinforces the knowledge silos we’re all trying to break down.

The biggest limitation: A walled garden for your knowledge

Atlassian Intelligence is impressive, but it has one major blind spot: it operates on the assumption that your team’s entire world exists within the Atlassian ecosystem. In the real world, that’s almost never true.

most teams find their knowledge is 'scattered beyond just Confluence.'

Reddit

Think about it. Your engineers probably have critical documentation in Google Docs. Your support team has a treasure trove of solutions saved as macros in Zendesk. Important project decisions are made and buried in conversations on Slack or Microsoft Teams.

This creates a "walled garden." The AI is trapped inside, unable to see the full picture. The JSM Virtual Agent can’t find a solution if it’s in an old Zendesk ticket. The Confluence Q&A search can’t pull an answer from a Notion database. Your AI is essentially working with one hand tied behind its back. This leads to incomplete answers, missed opportunities for automation, and frustrated employees who still can’t find what they need.

This infographic explains which AI does Atlassian use and how a unified platform avoids the walled garden problem by integrating all knowledge sources.:

For an AI assistant to be truly helpful, it needs to learn from all of your company’s knowledge, no matter which app it lives in. This is where you might need to look beyond the built-in tools. Platforms like eesel AI, for example, are designed specifically to solve this problem by connecting to your entire tech stack, not just the tools from a single company.

Atlassian Intelligence pricing

One of the first questions on everyone’s mind is, "How much does this cost?" The pricing model is actually pretty straightforward: Atlassian Intelligence pricing is included with all cloud-based Standard, Premium, and Enterprise plans.

There isn’t a separate line item or add-on fee for the AI features themselves. If your company is on an eligible plan, the features are automatically available. This means your access to AI is directly tied to the subscription level you have for Jira Software, Confluence, or Jira Service Management.

Just keep in mind that the most powerful AI capabilities, like the JSM Virtual Agent or advanced automation rules, are often reserved for the pricier Premium and Enterprise tiers. So while you aren’t paying for "AI" as a separate product, you do need to be on a higher-tier plan to unlock everything it can do.

A better approach: Unifying all your knowledge with eesel AI

While Atlassian Intelligence is a nice perk for teams living exclusively in Atlassian tools, its walled-garden design is a serious drawback. A truly smart system needs a bird’s-eye view of your entire organization’s knowledge. This is where a solution like eesel AI offers a much more complete and practical alternative.

Connect every knowledge source, not just Atlassian

The biggest difference with eesel AI is its ability to knock down those walls between your apps. It integrates with over 100 platforms right out of the box, including:

• Atlassian Tools: Of course, it connects to Confluence and Jira Service Management.

• Company Wikis: It pulls knowledge from Google Docs, Notion, and SharePoint.

• Help Desks: It learns from every ticket in Zendesk, Freshdesk, and Intercom.

• Chat Platforms: It finds answers buried in Slack and Microsoft Teams.

This means your AI agent can draw from every relevant document, ticket, and conversation to give you and your customers a complete, accurate answer every single time.

The eesel AI platform showing how it connects to multiple applications to build a comprehensive knowledge base, which is a key differentiator from the AI Atlassian uses.:

Get set up in minutes, not months

Some enterprise AI tools feel like they require a team of consultants and a six-month implementation plan. eesel AI is built to be the exact opposite. It’s radically self-serve. You can sign up, connect your apps with a few clicks, and have a functioning AI assistant ready to go in less time than it takes to get through your morning meetings.

Even better, you can test it before you fully launch. eesel AI’s simulation mode runs your AI against thousands of your past support tickets. It shows you exactly how it would have performed, giving you a precise forecast of its resolution rate. This lets you deploy with confidence, knowing exactly what to expect.

A screenshot of the eesel AI simulation mode, a feature that shows its potential performance, contrasting with the features of the AI Atlassian uses.:

You’re in complete control of your AI automation

eesel AI puts you in the driver’s seat. You get fine-grained control over how your AI automation works. You can set rules for which types of questions the AI should handle and what actions it’s allowed to take, whether that’s escalating a ticket to the right person in JSM or looking up live order details from Shopify. You can even customize its personality and tone to match your brand’s voice perfectly, ensuring a smooth and consistent experience for everyone.

A look at eesel AI's customization settings, where users can define rules and guardrails for their AI, a key aspect when considering which AI to use instead of just what AI does Atlassian use.:

Beyond the built-in solution

Atlassian Intelligence is a solid step forward, especially for teams that are deeply committed to the Atlassian ecosystem. By combining models from OpenAI with its own Teamwork Graph, it delivers some genuinely useful AI features directly within Jira, Confluence, and JSM.

However, its greatest strength is also its biggest weakness: it’s all about Atlassian. In today’s world, where company knowledge is spread across dozens of different applications, a siloed AI will always struggle to keep up.

If you’re looking for an AI strategy that reflects how your team actually works, you need a solution built to connect everything. A platform like eesel AI bridges the gaps between your scattered sources of information, creating a single, trustworthy source of truth that can power your support automation, make your team more productive, and deliver a far better customer experience.