Atlassian taps OpenAI as it expands intelligent features across its products

As large language models become increasingly accessible, they are helping people find new ways to create, communicate and collaborate. That's prompting major enterprise tech companies like Atlassian to look at adding new levels of AI capability to their products. Like many others, it has decided to tap OpenAI's groundbreaking AI research to complement its own areas of expertise.

Its new suite of AI capabilities announced today, Atlassian Intelligence, uses OpenAI technology as well as Atlassian's own large language models. Erika Trautman, Head of Product for Atlassian’s Work Management offerings, says this builds on earlier work at the vendor:

We've been working on solving problems around teamwork — how teams plan, track and communicate – for 20 years. That gives us unique insights into work graphs, specifically around advancing projects or advancing service work. Those are two different types of universally applicable work. And that can help make our Atlassian Intelligence solutions significantly more effective for teams.

Atlassian has been leveraging machine learning to enhance its products for years, with point features like personalized search. The new Atlassian Intelligence is designed to be more comprehensive. Built on the Atlassian Platform, the technology underlying Atlassian's complete portfolio, Atlassian Intelligence will enhance all of the company's products, including Jira Software, Confluence and Trello. The new capabilities are designed to speed up work, help users tap into institutional knowledge and help users answer questions.

It does this by creating a ‘teamwork graph’ that illustrates the types of work accomplished within an organization, as well as team relationships. The teamwork graph gets additional context and data from the third-party apps teams use.

All of the Atlassian Intelligence features are opt-in, so administrators have control over how their data is used. Atlassian's standard privacy policy applies, meaning Atlassian doesn't have visibility into user-generated data.

Making Jira workflows more accessible

The most powerful use case for Atlassian Intelligence can be found in Jira Software, Atlassian's flagship product. Jira holds wide appeal because it allows teams to create essentially any business workflow — no matter how complicated — using rules and conditions. For instance, a procurement team could set up a workflow for contract reviews, which require notifying personnel from the legal team and getting their approval on certain conditions. Sherif Mansour, Head of Product for Atlassian Intelligence, says:

Many organizations can use this to track processes at scale, large process at scale, whether they be software teams, procurement teams or legal teams, or whatever. That's the power of Jira. Because Jira is a highly configurable system, people can add custom fields and workflows and integrate it with downstream systems. Asking Jira where a piece of work is at, or what's happening with this thing, can be quite sophisticated.

However, Jira is admittedly pretty complicated to use. Atlassian has a SQL-like language called JQL that enables effectively infinite configurations. Mansour says:

This has always been a learning curve for customers. But with natural language and artificial intelligence, we can reduce that barrier and make that power available to everyone. So you don't need a someone who knows how to write query language to do this. And so you can just ask it a natural language question.

A virtual agent for Jira Service Management

For Jira Service Management, a dose of AI has an obvious application — a virtual service agent. Workers looking for answers to their questions will be able to get help from the new virtual agent from within Microsoft Teams or Slack. It will deliver answers based on internal knowledge base articles, and if needed, it can ask follow-up questions.

The new tool can write up summaries of pending requests to help support teams get up to speed, and it can surface information to help teams resolve issues.

Atlassian Intelligence can do more than just find information — it can help users deliver that information with the right ‘tone’. While adjusting content for tone will be useful across products, there's a clear use case within Jira Service Management, as Mansour explains:

We know sometimes agents who are on the receiving end of these service requests receive dozens of them a day. So they struggle with empathy sometimes or feeling some of the pain that the customer might feel with whatever problem they might have. AI can help here as well. We know good responses to customers can impact customer satisfaction majorly.

Speeding up work in Confluence

On the Confluence platform, speeding up work can mean summarizing transcripts and pulling out action items and key takeaways from a meeting. It can also mean using AI to create a first draft of a new document based on notes, or changing the tone of an existing document. It could even use existing content to draft up tweets or other kinds of social media posts.

In terms of institutional knowledge, Atlassian Intelligence will offer a feature across Confluence and other products that gives users the ability hover over a term and learn more about it. For instance, if you're looking at a Confluence page that references a special project, you can hover over the project name to get some more context.

Similarly, Atlassian Intelligence can tap into institutional knowledge to answer natural-language questions. Instead of searching through your company intranet to find lengthy documents about company policies, you could simply ask a question such as, “How much can I spend on my home office setup?”

The waitlist for Atlassian Intelligence is now open. Read our separate story about the full set of new Confluence features announced at today's Team 23 conference. 

My take

The hype around AI is going strong, and every technology company is scrambling to ensure their brand is part of the conversation. For a company like Atlassian, however, the value of a LLM-driven feature set is clear. The developers, knowledge workers and support teams that are the heart of Atlassian's user base are the prime customers for the services that AI can support. Meanwhile, the ability to implement Atlassian Intelligence across products, thanks to the underlying Atlassian Platform, should encourage users to embrace the new tools.