Atlassian CEO: AI won't replace SaaS tools—it will make them smarter

Atlassian's platform of work-management tools—including Jira, Trello, and the newer Rovo—sits at the center of one of tech's biggest debates: whether AI will destroy the SaaS category entirely. In a podcast interview, CEO Mike Cannon-Brookes pushed back against that "SaaSpocalypse" narrative, arguing instead that AI will amplify rather than replace companies' need for visibility and control.

What Atlassian really does#

The company describes itself as a platform that helps teams across an organization collaborate and manage work—connecting developers, operations, sales, finance, and service teams. Cannon-Brookes emphasized that the core function is enabling organizations to understand what's happening across their business at multiple levels of abstraction: what an individual contributor needs to do, what a team is executing, and what the organization is chasing strategically.

Critically, the actual work doesn't happen inside Jira or Trello. Instead, these tools serve as what Cannon-Brookes called a "human reference to work"—a layer of legibility that lets people see, coordinate, and make decisions at scale.

Why AI adoption increases, not decreases, SaaS usage#

Cannon-Brookes argued that AI will actually drive more reliance on platforms like Atlassian's. As businesses deploy AI to handle routine parts of workflows—say, AI systems that approve 80% of straightforward sales discounts—the need to track, measure, and optimize those processes only intensifies. Someone still needs to understand how many exceptions the company is granting, whether more staff is required, or if processes should change.

The fundamental dynamic mirrors what happened to bank branches after mobile banking arrived. Most predicted brick-and-mortar branches would vanish, but instead, "bank branches have adapted," Cannon-Brookes noted. They've shifted from transactional work to higher-level customer service and complex problem-solving, and their numbers have grown.

The human layer that AI can't replace#

Cannon-Brookes was direct about what large language models can't do: run an entire business automatically. Most organizations are "global conglomerations of rules, compliance, laws, staff in different places, human inconsistency—which is the same as saying human creativity." Even with five to ten years of AI improvement, knowledge work will remain abundant because competitive bars will rise across industries, requiring more developers and skilled workers, not fewer.