Australian collaborationware slinger Atlassian has licensed OpenAI’s tech and sprinkled generative AI functionality on its flagship products.
The centerpiece of the enterprise software giant’s machine-learning push, announced today at its Team ’23 conference, is its intention to have AI work as a mercifully unnamed “virtual teammate” that takes on chores such as mining existing data and applying generative artificial intelligence to produce drafts of Confluence documents, or respond to Jira tickets.
Offering improved natural language queries instead of requiring users to employ Jira Query Language is another goal.
Tiffany To, the company’s head of enterprise platform and data, told The Register that in a previous role she used AI to scan for flaws in software, and while that worked, results were better when humans also considered the code.
That experience informed Atlassian’s strategy of AI as a teammate, not an independent operator.
To added Atlassian is yet to decide how AI will impact its pricing, or which licensing tiers will offer different AI features and functions to its different pricing tiers.
Once users have had a chance to understand which AI features add value, Atlassian will solidify such matters. One thing is for sure, these AI features are coming to Atlassian’s cloud-based products, not its on-prem editions. Or not yet, anyway.
Also at Team ’23, Atlassian debuted an analytics product based on its 2021 acquisition of Chartio.
The product is essentially a data lake that sucks in info from Jira Software, Jira Work Management, Jira Product Discovery, Jira Service Management (including incident and asset data), and Confluence, and can connect to Snowflake, Amazon Redshift, Google BigQuery, and other databases.
As is always the case with analytics products, Atlassian promises searing insights will be yours once you’re up and running. The company suggests software development team leads will gain the ability to “measure work output and team performance at the individual team, cross-team, and organization levels with detailed scorecards for metrics like cycle time and time spent in review.”
IT operations teams will be able to “proactively spot and fix bottlenecks with out-of-the-box dashboards for request management, incidents, changes, and service performance.”
Confluence has also been given a new content type – whiteboards – in line with Atlassian’s desire to have every conceivable form of info managed by its tools. Another Confluence enhancement is a Chrome plugin that can suck in a web page so it can be annotated or shared as part of a collaborative process. And Confluence’s licensing has been tickled so that guest users are now gratis. The tool has gained a data residency function too.
To told The Register that Atlassian reckons its AI efforts will improve over time because while it has an enormous amount of data about business workflows, software development practices, and ITSM, it has only run OpenAI for a few months.
She added that as an OpenAI licensee, Atlassian runs its own infrastructure meaning customer data won’t be ingested by models used for public-facing services.
“We don’t share any data,” she said. “This is all per instance or per customer level.” ®
