What a $42B Software Co. Really Spends on AI Tools

What It Covers

Mike Cannon-Brookes, CEO of Atlassian ($42B valuation), explains how his company approaches AI implementation across developer and business teams. He reveals metrics on 3.5+ million monthly AI users, discusses measuring developer productivity versus developer joy, and shares why Atlassian uses 75+ AI models simultaneously while maintaining focus on sustainable long-term growth over rapid revenue maximization.

Key Questions Answered

• •Business Process Automation Philosophy: Atlassian views AI as automating specific boxes within workflow flowcharts rather than eliminating entire processes. Organizations assign work to AI agents (from platforms like Salesforce, Google Agent Space, or GitHub Copilot) at specific workflow points, then route completed work back to humans or other agents for approval, authorization, or next steps. This approach preserves essential business collaboration while accelerating specific tasks.
• •Enterprise AI Architecture: Atlassian built a teamwork graph containing 100+ billion objects and connections, growing 50%+ quarter-over-quarter. This graph tracks links between Jira issues, Confluence pages, Google Docs, Figma files, GitHub pull requests, and Salesforce records across hundreds of SaaS applications. The system maintains permissions and enables semantic search, creating what functions as the largest enterprise search engine by providing AI tools organizational context.
• •Developer Productivity Measurement: Atlassian tracks developer joy rather than raw productivity metrics, believing creative professionals produce best work when satisfied. The company acquired DX to combine quantitative metrics (pull request cycle times, build speeds) with qualitative surveys identifying pain points. This reveals where perceived AI efficiency gains don't match actual output improvements, helping allocate thousands of dollars per developer in AI tool spending across platforms.
• •AI Tool ROI Analysis: Atlassian deploys four major coding tools (Rovo Dev, Claude Code, Cursor, GitHub Copilot) to 10,000+ R&D staff, tracking which tools deliver better ROI in specific contexts. Rovo Dev excels at finding bugs and security issues by leveraging the teamwork graph for prior solutions across repositories. Code generation speed improves significantly, but developers must rebuild mental models when reviewing AI-generated code rather than writing it themselves.
• •Code Maintenance Applications: AI coding agents excel at large-scale code maintenance tasks across existing codebases. When Atlassian needed to update 500+ repositories for an API change, developers wrote examples in JavaScript and Java, then agents found similar patterns and generated pull requests. This human-AI collaboration loop prevents compounding errors while handling tedious refactoring work that would consume senior developer time.
• •Sustainable Growth Strategy: Atlassian deliberately avoids maximizing quarterly revenue growth, instead allocating resources toward infrastructure and capabilities that enable 20-30% growth rates five years forward. The company grew cloud revenue 26% and revenue performance obligations 40% in recent quarters at $56B+ valuation by balancing short-term harvesting with long-term seeding, ensuring architectural investments survive technology disruptions like AI transformation.