Roopali Anand Thapar, IFPUG president
This presentation introduces a practical framework for estimating AI agents using Function Point Analysis (FPA). While traditional FPA suits data-driven systems, AI agents require a nuanced approach due to their learning, reasoning, and interaction capabilities. The session explores how to map AI functionalities to FPA components, adjust complexity weights, and incorporate value adjustment factors. It also contrasts the effort involved in building versus embedding AI agents and provides actionable recommendations for accurate, scalable estimation in AI-driven solutions.
Key Topics
- Introduction to Function Point Analysis (FPA)
- Understanding AI Agents in Software Systems
- Adapting FPA for AI Workloads
- Complexity Weighting and Value Adjustment Factors (VAFs)
- Effort Estimation: Building vs. Embedding AI Agents
- Sources of Estimation Data
- Estimation Strategy and Best Practices
- Recommendations and Expert Tips
Benefits for Participants
Participants will gain:
- A clear understanding of how Function Point Analysis can be adapted for AI agent estimation.
- Practical techniques to map AI agent components to FPA elements.
- Insights into effort drivers for building vs. embedding AI agents.
- Strategies to estimate AI workloads using standardized, technology-agnostic methods.
- Expert tips for improving estimation accuracy and traceability in AI projects.