Epic Validator
Description
From idea to a review-ready epic
Epic Validator is a dialogue assistant for product managers and business or systems analysts. It accepts an idea, free-form description, meeting notes or supporting materials, identifies what is confirmed and what is missing, asks a small number of focused questions and prepares a structured Jira epic.
Quality checks for one epic
The service checks the epic against a defined internal methodology: value-oriented title, epic type and Change/Run classification, purpose and dependencies, team and owner, completion criteria, quarter-end criteria, design flag, effort estimate, RICE inputs and calculation, 12-month effect and driver, due date and labels. It explains gaps, flags contradictions and proposes ready-to-review wording instead of returning an opaque score.
Batch audit
Epic Validator accepts Excel, CSV and JSON exports with different column names and layouts. It reads the structure first, maps available columns to the methodology, shows ambiguous matches, checks only the criteria supported by the file and distinguishes a missing column from an empty field. It returns coverage, a summary across the set, recurring issues and detailed recommendations for problematic epics.
Why it matters
The service turns implicit expectations about a well-formed epic into a repeatable, explainable quality workflow that supports both epic creation and readiness checks before planning.
My Role
Designed and independently delivered the AI-assisted service using AI development tools. Defined the validation criteria from the team's requirements standard, the review workflow, expected outputs and acceptance logic. Used AI-assisted tools to translate these decisions into a working service, test the results and deploy it alongside the BA Tools product family. My ownership covered product definition, validation and delivery; code generation and framework-level implementation were AI-assisted.