| TL;DR: Wall Street spent a year betting AI would disrupt SaaS by making enterprise software easier to replicate. Atlassian’s Q4 results suggest the opposite may be true. Companies with years of embedded workflows, organizational data, and customer context may become harder to displace as AI improves, not easier. Atlassian’s 28% revenue growth, accelerating cloud numbers, and $250M bet on itself are early signals. The strongest SaaS platforms may be absorbing AI into their moat, not competing against it. |
For the past year, a single question has been hanging over every SaaS company: what happens when AI makes your product obsolete?
The fear was specific. If AI tools can write code, automate workflows, and shrink enterprise headcounts, the per-seat SaaS model starts looking shaky. Fewer employees means fewer licenses. And if AI can help companies clone enterprise software, why pay for the real thing?
Well, Atlassian’s latest results offer a pretty interesting answer.
The company reported $1.77 billion in Q4 revenue, up 28% year-over-year, against a Wall Street forecast of $1.66 billion. It swung from a $23.9 million net loss in the same quarter last year to a GAAP profit of $139.1 million. On a non-GAAP basis, profit came in at $473.1 million, up 83%.
That’s not exactly what you’d expect from a company whose business is supposedly being made obsolete.
The Fear Made Sense. The Results Tell a Different Story.
The “AI kills SaaS” argument was missing something from the start. It assumed enterprises would respond to AI by building their own tools or shrinking headcounts fast enough to hurt vendors.
But Atlassian’s results suggest that assumption may be a little too simple.
More than 350,000 organizations have years’ worth of work, knowledge, and workflows living within Atlassian’s ecosystem. That’s not something companies can simply rip out and replace overnight. For deeply embedded customers, that history and workflow context makes switching platforms expensive and complex.
This is where things get interesting. The company is betting that AI can make that existing context more valuable, not less.
Why Rovo Works When Generic AI Does Not
In 2024, Atlassian launched Rovo, an AI platform built directly into Jira and Confluence. Rovo combines enterprise search, chat, AI agents, and Rovo Studio for building agents and automations. The company also offers Rovo Dev separately for software development workflows.
The core idea behind Rovo is straightforward. AI is only as useful as the context it can work with. And Atlassian has extensive organizational context across Jira, Confluence, Loom, Bitbucket, and connected third-party tools. It calls this the Teamwork Graph, which it says contains more than 200 billion objects and connections across customer graphs.
The company has numbers to back up its argument. Its internal testing found agents using Teamwork Graph context produced 44% more accurate answers while consuming 48% fewer tokens than agents without that connected context. Cisco tested this in practice. The company built AI agents using Rovo to automate reporting workflows, and some processes became 40 times faster.
The customer numbers are interesting, too. Rovo adopters are growing their ARR more than twice as fast as non-adopters. Customers that adopt Rovo are also completing 20% more Jira work items and creating or editing 25% more Confluence pages than non-adopters.
That does not mean Rovo alone caused the higher ARR growth. Still, it suggests that deeper AI adoption is happening alongside greater engagement and expansion across Atlassian’s platform.
Getting Enterprise Software Running is Only Half the Problem.
Most enterprises that consider building their own tools underestimate what comes after the build.
Because building the thing is only the beginning.
Running enterprise software at scale requires infrastructure, security compliance, uptime guarantees, and ongoing technical support. All of that costs money and people. For many enterprises, those costs can make buying an established platform more economically attractive, particularly when the software is already deeply embedded across the organization.
Buying a platform like Atlassian also means the vendor carries a significant share of the operational burden, support, uptime, and security updates that an internal team would otherwise own.
So, ultimately, the replication argument was never just about whether you could build the software. It was also about whether you could operate, secure, maintain, integrate, and support it at enterprise scale.
What the Pipeline Actually Looks Like
The Rovo story sits inside a broader acceleration. Cloud revenue grew 31% to $1.2 billion in Q4. Remaining performance obligations reached $4.8 billion, up 44% year-over-year, reflecting a substantial increase in contracted future revenue.
The company has also said customers are making larger and longer commitments to the platform. It reported a record quarter for $1 million-plus, $3 million-plus, and $5 million-plus annual contract value deals.
And that matters. The AI-era SaaS debate is not only about how many seats a company can sell. It is also about whether its software becomes more deeply embedded in how organizations operate.
What Spending $250M on Your Own Stock Actually Signals
CEO and co-founder Mike Cannon-Brookes announced that he plans to personally invest up to $250 million in Atlassian shares on the open market. That is different from the company announcing a $250 million corporate buyback.
The move came alongside Atlassian’s strong Q4 results and can reasonably be read as a significant expression of Cannon-Brookes’ confidence in the company’s long-term prospects.
Of course, the investment does not erase the challenges ahead. Atlassian is still forecasting slower overall revenue growth in FY2027.
The SaaS Companies that Win Will Have to Adapt to AI, Not Compete Against It
The question was never really whether AI would change SaaS. It was what that change would mean for companies whose products suddenly became easier to replicate.
Atlassian’s latest results offer one possible answer. The value of enterprise software may not lie only in the software itself, but in everything built around it: the data, workflows, integrations, and institutional knowledge that accumulate over years.
After all, an AI tool can help you build something new. It can’t instantly recreate years of context inside an organization.
Atlassian still has to prove that this advantage can translate into sustained growth. But for now, its results suggest that AI may be less of a SaaS killer than a test of which software companies have built something worth keeping.
And, ultimately, that may be the real shift happening in SaaS: AI isn’t necessarily replacing the platforms businesses already use. It may be making the strongest ones harder to replace.
