Idea
An AI-powered platform generating guided follow-up questions to improve requirements elicitation interviews for product managers and analysts
Research Paper
Core Innovation
This paper introduces a method using GPT-4o to generate follow-up questions during requirements elicitation by leveraging a framework of common interviewer mistakes. It uniquely guides question generation based on mistake types to improve relevance and quality. The approach is validated by comparing LLM-generated questions to human-authored ones, showing comparable or superior performance in real-time interview settings.
Market Size (TAM)
$2–10B TAM, $500M–$1B SAM; assumption: software development and product management tools market with growing AI integration.
Potential Customers & Pain Points
- Product Managers Struggling with Domain Knowledge
- Business Analysts Facing Cognitive Load in Interviews
- Software Teams Needing Better Requirements Clarity
Business Model
Subscription-based SaaS platform with tiered pricing for enterprises and integration APIs for software tools.
Competitive Landscape
- Refract
- Gong
- Chorus
Implementation Challenges
- Domain-specific adaptation challenges
- User trust in AI-generated questions
- Integration with existing interview workflows
Validation Strategy
- Pilot with product teams to measure interview quality improvements
- A/B test AI-generated vs human-generated follow-up questions
- Collect user feedback on question relevance and usability
Research Paper Overview
Requirements Elicitation Follow-Up Question Generation
Summary
This paper explores using GPT-4o to generate follow-up questions during requirements elicitation interviews, addressing challenges like domain unfamiliarity and cognitive load. It builds on a framework of common interviewer mistakes to guide question generation and evaluates LLM-generated questions against human-authored ones, finding LLMs perform comparably or better when guided by mistake types, enhancing interview quality in real time.