Apr 1, 12:00 – 1:00 AM (UTC)
Natural language To Jira Query Language (JQL) is a feature designed to take the guesswork out of writing (JQL) queries. Over the past year, Atlassian has been investing in improving this feature - upgrading the underlying machine learning models, adding custom field support and enhancing search relevance. Join Machine Learning engineer, Matt Burton to learn more about these improvements.
59 RSVPs
The NL2JQL (Natural Language to JQL) feature has received significant enhancements aimed at improving the accuracy and efficiency of converting natural language queries into Jira Query Language (JQL).
Key Improvements:
1. Custom Field Search: Users can now perform complex queries involving custom fields and values, meeting the demand for detailed search capabilities.
2. In Context Search: Natural Language to JQL now understands how users navigate Jira and uses this context to infer the correct JQL syntax to use.
3. Improved models: Atlassian has been at the forefront of language model adoption, benefiting from the gradual improvement in language models over time.
Technical Foundation:
These improvements are built on AI models that accurately process natural language, utilizing prompt caching and model-specific features to optimize performance.
Impact on Teams:
The enhancements streamline issue tracking and management, allowing teams to focus on resolving issues rather than constructing complex JQL queries, ultimately boosting productivity and creating a more intuitive experience within Jira.
Atlassian
ML Engineer – Central AI
