Atlassian's AI playbook: Why bolting on AI beat rebuilding from scratch

Atlassian's AI playbook: Why bolting on AI beat rebuilding from scratch

Oct 10, 2026
SaaStr AI SprinklerAS Gtm_strategy

The Gist

  • Atlassian bolted AI onto 20+ legacy apps, now used by 5M+ users
  • Almost skipped chat interface but shipped it as a discovery tool
  • Chat backend became universal interface like DOS command line
  • Whiteboard prompts evolved into full features and workflow tools
Key Quotes

The standard advice is to never bolt AI onto an existing product and to reimagine everything from scratch. Atlassian did the thing you're told not to do.

Sherif Mansour

Juniors are up to 38% more likely to use AI, nearly twice as likely to experiment, and use more tools... Seniors adopted more slowly but are much better at spotting slop and at getting better output from the AI.

Sherif Mansour
Key Insights
  • Bolt-on AI integration can be more effective than rebuilding from scratch, as demonstrated by Atlassian's success with Rovo chat and Jira automation.
  • Chat interfaces serve as a universal starting point for AI, revealing which use cases warrant dedicated features, similar to how DOS command lines evolved into specialized apps.
  • Agents (AI tools) require the same four elements as humans: tools, context, goals/accountability, and visibility into teammates' work.
  • AI builder roles should not replace PMs or designers, as their decision-making and customer context are critical for steering engineering teams.
  • Juniors adopt AI faster and experiment more, while seniors excel at quality control and refining AI output, making a combined approach optimal.
  • Atlassian's AI Builder Week program pairs juniors and seniors to share tools/techniques and quality control practices, accelerating org-wide AI adoption.
Actionable Takeaways
  • Implement a chat interface in your product to uncover high-value AI use cases, even if chat isn't the end-state UX.
  • Bolt AI onto existing workflows (vs. rebuilding) to accelerate learning and leverage existing user behavior data.
  • Structure teams to maintain PM/designer roles for decision-making, even as engineers become more productive with AI.
  • Pair junior and senior talent deliberately, with juniors driving experimentation and seniors ensuring quality/output refinement.
Data Points
  • 5 million users (Regularly use Atlassian apps for AI capabilities)
  • 38% more likely (Juniors are more likely to use AI compared to seniors)
  • 1 PM/designer per 10-40 engineers (Ratio shifts with AI adoption, increasing demand for PM/designer input)

RevBots.ai View:

The AI Sprinkler approach works when you treat chat as a discovery layer that reveals which use cases deserve dedicated interfaces.

Full Story: SaaStr →