The fundamental unit of analysis has shifted. We no longer design for users; we design for collaborative entities consisting of a human and an AI agent working together, evolving over time.
Traditional requirements methodologies—user stories, use cases, and personas—were designed for a world where humans acted alone and software was a passive tool. As AI agents become active collaborators capable of learning, adapting, and taking initiative, these frameworks fail to capture the dynamic, evolving nature of human-agent partnerships. Agentic Analysis provides a complete toolkit: methods to understand current work (Agentic Replica), envision optimized futures (Agentic Reengineering), and artifacts that model how trust develops, responsibilities shift, and emergent capabilities arise when humans and agents work together over time.
Two complementary approaches for understanding work and designing human-agent collaboration.
Map the as-is workflow of human employees. Treat the job as a "black box" to catalog inputs, transformations, and outputs—creating a factual baseline for AI agent development.
What exists today
IPO diagrams
Envision how job areas could be optimized with AI agents. Restructure workflows for efficiency, scalability, and innovation—designing the to-be state.
What could be
Reengineered workflows
From traditional user stories to collaborative HAP Plans, these artifacts capture the full spectrum of human-agent requirements.
Best suited for conventional apps without agent collaboration.
For autonomous agents with triggers, skills, workflows, and collaboration patterns.
The primary requirements artifact for collaborative systems. Maps how the pair works together, builds trust, and evolves.
See how methods map to Light and Full artifact formats