An open-source agentic pipeline that turns raw meeting transcripts into validated Jira epics, user stories, acceptance criteria, and technical task trees via Atlassian REST API.
Engineering planning meetings generate long, messy transcripts. Engineering managers and product owners spent 1 to 2 hours every sprint manually reading transcriptions to write structured Jira stories, acceptance criteria, and estimation sub-tasks.
Built a python-based agentic pipeline utilizing Custom AI Agents for multi-stage section extraction. The system enforces strict Pydantic validation to guarantee output schema integrity before automatically invoking the Atlassian REST API to generate linked Epics, User Stories, and Sub-tasks in Jira.
Reduced sprint refinement and ticket creation time from 60 minutes to 5 minutes per sprint. Auditable end-to-end with zero missing acceptance criteria.
Ask Bilal’s AI Clone