From Planning Docs to Jira Task Uploads, Tasky

Tasky is an AI-based workflow automation system built by Yeogieotte Company that reads Confluence planning documents and automatically generates tickets up to Jira. Its core goal is to reduce the repetitive work of reading planning documents and breaking them into tasks, allowing developers to focus more on implementation and problem-solving.

The adoption value is also clear. It centers on three axes: improved developer productivity, reduced project preparation time, and standardization of task quality. It aims to shrink the document-to-ticket conversion process—which used to take tens of minutes—down to a matter of minutes, and to reduce the task breakdown methods and omission/duplication issues that used to vary from person to person.

The operating flow is organized in stages.

• Document reading: Fetches documents via the Confluence API and strips HTML tags to extract the title, body, and Space information.
• PRD generation: Further organizes unstructured, human-written documents so that AI, given an assigned role, produces a PRD in JSON format.
• Task extraction: Based on the standardized PRD, creates feature-specific development tasks and calculates dependencies, priorities, and estimated time required.
• Frontend task selection: Comprehensively analyzes task titles, descriptions, and details to separately select frontend work only, such as UI development, interactions, and API integration.
• Context enrichment: References the PRD together with metadata.json to reflect the project's component structure, API integration patterns, and UI patterns.
• Detailed task breakdown: Further splits tasks into execution-level subtasks so developers can start immediately, adding context, estimated work time, and importance to the descriptions.
• Jira upload: Registers the final results to the selected Epic via the Jira REST API, preserving the hierarchical structure and priorities.

Since adoption, teams can now quickly grasp the full context just by looking at the tickets, and duplicate/omitted tasks have been greatly reduced. Going forward, planned improvements include enhanced duplicate task detection, automatic updates of start/due dates, and Figma context integration. Ultimately, Tasky is not a tool where AI decides everything, but an automation system that safely converts documents into tickets within rules and context set by humans.