User Story Refinement for AI Features

This skill helps product managers refine user stories for AI-powered features, ensuring clarity, testability, and alignment with AI capabilities and limitations.

תחום: ניהול מוצר

מתי להשתמש

Use this skill when drafting or reviewing user stories for features that incorporate machine learning, natural language processing, or other AI technologies.

תגיות: product-management, ai, user-stories, feature-development, requirements

SKILL.md

--- name: User Story Refinement for AI Features description: This skill helps product managers refine user stories for AI-powered features, ensuring clarity, testability, and alignment with AI capabilities and limitations. --- ## Overview Developing AI-powered features requires a nuanced approach to user story writing. Traditional user stories might overlook the inherent uncertainties, data dependencies, and evaluation metrics crucial for AI components. This skill provides a structured method to refine user stories specifically for AI features, focusing on defining expected AI behavior, data considerations, potential failure modes, and clear acceptance criteria. ## When to use Invoke this skill when you are drafting new user stories for AI capabilities, or when reviewing existing user stories to ensure they adequately address the unique aspects of AI implementation. It is particularly useful during sprint planning, backlog refinement, or feature definition phases. ## How it works 1. **Input User Story:** Provide the initial user story you wish to refine. This could be a high-level story or one already partially defined. 2. **Identify AI Component:** The skill will analyze the user story to identify the core AI functionality or component involved. 3. **Prompt for AI-Specific Details:** You will be prompted to consider and add details related to: * **Expected AI Output/Behavior:** What should the AI achieve or produce? (e.g., "The system suggests relevant articles based on user browsing history."). * **Data Requirements:** What data is needed for the AI model? (e.g., "User browsing history, article metadata, user engagement metrics."). * **Performance Metrics:** How will the AI's success be measured? (e.g., "90% accuracy in article relevance, 5% increase in click-through rate."). * **Edge Cases & Failure Modes:** How should the system behave when the AI is uncertain or performs incorrectly? (e.g., "If the AI cannot confidently suggest articles, display a