The same AI that was supposed to kill SaaS just posted its rescue numbers.

The Summary

The Signal

For months, the story went like this: AI agents would let companies build their own tools, making SaaS products redundant. Why pay Atlassian for project management when your AI can spin up a custom tracker in an afternoon? The market believed it. SaaS stocks took a beating in 2026 on exactly this thesis.

Thursday's earnings calls flipped the script. The companies supposedly most vulnerable to AI replacement just posted growth numbers powered by AI adoption. Atlassian's 32% revenue growth wasn't despite AI. It was because of it. Same story at Twilio and Five9, where AI features drove customer expansion instead of customer exodus.

"The very same technology that has pummelled their share prices this year over fears that companies could use AI to build the software they sell."

Here's what the market missed: most companies don't want to build software. They want working software, fast, that integrates with everything else they're already running. Building custom tools with AI sounds great until you're three months in, debugging edge cases, and your internal team is now maintaining code instead of shipping features. SaaS companies that embedded AI into their platforms gave customers the productivity gains without the maintenance burden.

The revenue split tells you who figured this out:

  • Atlassian: 32% growth (collaboration, project tracking)
  • Twilio: 20% growth (cloud communications APIs)
  • Five9: 9% growth (contact center software)

The pattern isn't random. Atlassian and Twilio sell developer-adjacent tools where AI augmentation makes existing workflows faster. Five9's slower growth suggests contact centers are still experimenting with whether to augment human agents or replace them entirely. The companies selling "AI plus your existing stack" are winning. The ones selling "AI instead of people" are still figuring it out.

Kate Leaman at AvaTrade is right to pump the brakes. Three earnings beats don't end the SaaSpocalypse. But they do clarify the dividing line: SaaS companies that treat AI as a feature die. SaaS companies that treat AI as the new interface for their underlying value prop survive. Atlassian's moat isn't Jira's UI. It's the accumulated project data, integrations, and workflow logic that took years to build. AI just made accessing that moat easier.

The Implication

If you're holding SaaS stocks, these earnings don't mean you're safe. They mean you need to ask better questions. Is the company building AI features, or is it building AI-native products? Is it defending existing customers with better tools, or is it expanding into new markets AI just opened up? Revenue growth this quarter matters less than whether the product strategy assumes agents exist.

For companies buying software, the calculus just shifted. The "build vs. buy" debate isn't dead, but it's more expensive than the AI optimists claimed. Watch which SaaS vendors are shipping agent-friendly APIs and which are still selling seats. The next wave of growth won't come from human users clicking buttons. It'll come from agents calling endpoints.

Sources

Business Insider Tech