Explore two real-world approaches to building AI agents for Atlassian’s stock plan support—one trained on curated data, another that learns dynamically from user interactions. See how these models perform in production across two environments: within Jira Service Desk to intercept employee questions before tickets are submitted, the other as a global assistant via Rovo with broad access to plan data. Through a side-by-side comparison, you’ll see how differences in design, behavior, and tone impact the user experience and outcomes. You’ll come away understanding how each approach delivers accuracy, scales effectively, and earns user trust—and what that means for reducing support volume and improving response quality.