Find Requirements Problems Before Development Starts

AI-assisted quality checks for User Stories and Acceptance Criteria - bringing a systematic QA perspective into refinement.

Instead of discovering requirements problems during testing, I identify and resolve them before development starts.

Seconds
for the AI quality analysis
AI-assisted
requirements review
Jira
integrated workflow
Daily
production use

The Problem

Poorly specified User Stories can create unnecessary discussions, clarification loops, ambiguity, late testability problems, rework, and eventually defects. The problem is often not discovered until development or testing has already started.

The goal is to move the quality check upstream. Before development, not after it.

The Approach

I developed an AI-assisted tool that evaluates User Stories against IREB-based requirements quality criteria. The tool analyses aspects such as atomicity, consistency, completeness, unambiguity, understandability, testability, correctness, feasibility, and traceability.

It provides a quality score and criterion-by-criterion assessment. It also identifies problems, generates questions for the Product Owner, and proposes improved User Stories and Acceptance Criteria. The AI accelerates the analysis. The QA methodology and quality criteria provide the structure.

Workflow

Jira User Story
AI Quality Check
Score & Analysis
Questions for PO
Improved Story + Acceptance Criteria
Optional update back to Jira

The tool can retrieve tickets through the Jira API or analyse a pasted User Story description.

Before / After

Before After
Requirements reviewed mainly through discussionSystematic quality check before refinement
Ambiguity discovered during developmentAmbiguity identified early
Testability problems discovered laterTestability checked upfront
Missing information creates clarification loopsAI generates concrete questions
Manual QA review of every StoryAI-assisted quality assessment
Story improved through multiple iterationsAI proposes improved Story and AC

The Output

Quality Score
An overall assessment of the Story.
Criterion Analysis
A structured evaluation of individual quality characteristics.
Problems
Specific explanations of identified weaknesses.
Questions
Concrete questions for the Product Owner where information is missing or unclear.
Improved Story + Acceptance Criteria
A proposed improved formulation with more complete and testable Acceptance Criteria.

The Product Owner and QA remain responsible for the final decision.

The Result

The quality check takes seconds and is used as part of my daily production workflow. Instead of discovering requirements problems during testing, I can identify and address them before development starts. The result is a repeatable requirements-quality gate that brings QA thinking earlier into the software delivery process.

Business Impact

Earlier detection of requirements problems
Better testability before implementation
Fewer unnecessary clarification loops
More structured refinement
Better Acceptance Criteria
Earlier QA involvement
Reduced risk of requirements-related rework
Repeatable requirements-quality checks

Where This Fits

Requirements - Find quality problems before development starts
Test Design - Turn approved Acceptance Criteria into structured test cases
Development - Build quality into the development process
Deployment - Verify the deployed software automatically

Move quality earlier. Automate repetitive work. Keep humans responsible for quality decisions.

Instead of discovering requirements problems during testing, I identify and resolve them before development starts.