Agent Model Fit
Right agent. Right model. Right problem.
Agent Model Fit is a living research and implementation publication for digital and IT leaders deciding where AI belongs in real work.
Start here
Building Agent Model Fit: Right Agent, Right Model, Right Problem
A documented publication build shows how objectives, decision rights, versioned handoffs, and corrections shape agent work—and what the record cannot prove.
Latest articles
All PostsContext, Memory, and Retrieval: What Does an Agent Know?
Separate what an agent can use now, what persists between runs, and what needs a fresh lookup before trusting its answer.
ReAct, Reflection, or Planner–Executor: Which Agent Pattern Fits?
Choose agent patterns by what the task needs: adaptive tool use, feedback-driven revision, or explicit planning. Combine them only where each adds value.
Tool Access and Agent Permission: Where Approval Belongs
Tool discovery and connected credentials establish technical access. A separate, enforceable boundary determines which actions an agent may take.
Sections
Explore all sections- Framework
4 Articles
ways to reason about agent, model, and workflow decisions.
- Fieldwork
2 Articles
documented implementation choices, corrections, and lessons.
- Guides
6 Articles
practical procedures a reader can adapt to a defined task.
- Evaluations
4 Articles
how to assess outputs and interpret comparison evidence.
- Business
1 Article
organizational value, readiness, operating costs, and practical tradeoffs.
- Case Studies
0 Articles
bounded accounts of a problem, intervention, and outcome when evidence supports them.