{"type":"video","version":"1.0","html":"<iframe src=\"https://www.loom.com/embed/296aaeed7af640c6850a96d9c3906e95\" frameborder=\"0\" width=\"1280\" height=\"960\" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>","height":960,"width":1280,"provider_name":"Loom","provider_url":"https://www.loom.com","thumbnail_height":960,"thumbnail_width":1280,"thumbnail_url":"https://cdn.loom.com/sessions/thumbnails/296aaeed7af640c6850a96d9c3906e95-03d804b1e955517d.gif","duration":2725.24,"title":"Natural Language to JQL - New improvements including Custom Field search","description":"The meeting focused on enhancing Jira's Issue Search through JQL and AI, addressing user challenges, and introducing Atlassian's AI tool, Rover. Key improvements include Natural Language to JQL, error reduction, and user feedback integration. Future plans involve expanding AI capabilities to AQL, with upcoming demos at a team conference.\n\n### Introduction and Agenda Overview 2:22\n\n- Matt Burton introduced the session as part of the \"How We Built This\" series, focusing on making Issue Search effective.\n- The session aimed to cover Jira Query Language (JQL), Natural Language to JQL, maximizing Issue Search, and Atlassian's AI tool, Rover.\n- The presentation was structured to include a Q&A session at the end, with an expected duration of 30 to 45 minutes.\n\n### Jira Query Language (JQL) Overview 4:58\n\n- JQL is essential for searching issues in Jira, interfacing with multiple databases.\n- The issue search UI has undergone design changes, including renaming \"issue\" to \"work items.\"\n- JQL consists of fields, operators, and values, forming clauses and queries.\n- Despite its complexity, JQL is crucial for detailed searches, though 93% of users avoid it due to its complexity.\n- Basic search is limited, prompting the development of Natural Language to JQL.\n\n### Natural Language to JQL 9:33\n\n- Natural Language to JQL converts natural language queries into JQL, available only on cloud with AI enabled.\n- The feature has improved from generating valid JQL 65% of the time in 2023 to nearly 99% recently.\n- Despite improvements, user retention for AI-powered search is lower compared to basic and advanced searches.\n- The feature is being refined to enhance user experience and retention.\n\n### Limitations and Feedback on JQL 20:50\n\n- JQL's main limitation is its inability to easily explore issue hierarchies, such as searching for child issues.\n- Users cannot compare fields between issues, a common request.\n- Third-party plugins offer some solutions, but Atlassian does not provide these features out-of-the-box.\n- User feedback is actively monitored to guide improvements.\n\n### Tips for Effective Issue Search 24:12\n\n- Users are encouraged to use JQL or Natural Language to JQL for complex searches.\n- A new JQL debugger feature is in beta, expected to be available next month.\n- Managing searches within a project scope can enhance search relevance.\n- The text field in JQL allows for comprehensive text searches across text-based fields.\n- Community resources and documentation are recommended for learning and troubleshooting JQL.\n\n### Introduction to Rover 29:24\n\n- Rover is Atlassian's AI offering, integrating language models to enhance data interaction.\n- It allows for multi-turn interactions and can chain product features for deeper data exploration.\n- Rover can perform actions like summarizing issues and assigning tasks, enhancing workflow efficiency.\n- The tool supports integration with various third-party data sources, expanding its utility.\n\n### Future Developments and Q&A 39:44\n\n- Arielle Unterberger inquired about plans for AQL development similar to JQL.\n- Matt Burton confirmed internal testing for NL to AQL, with plans to integrate it into Rover.\n- Upcoming demos are expected at a team conference in the near future.\n- Participants were encouraged to provide feedback and reach out with questions via email or feedback tools."}