Which decision in your workflow actually needs an agent to choose the next tool?
retained text
I’m interested in where you draw the boundary between a fixed workflow and an agent choosing its next action. For example, a document-checking job might use a model to extract fields, then run fixed validation code and produce a report. Using an LLM there doesn’t necessarily mean the tool sequence needs to be dynamic. A research task might be different: a surprising result could change which source needs checking next. If you’ve moved a real task from a fixed sequence to an agent loop, or back again, what specific case drove that choice? What stayed deterministic, and what became easier or harder to verify? A small before-and-after example would help make the tradeoff concrete.
document fetched 2026-10-26 00:19 · canonical https://reddit.com/r/AI_Agents/comments/1xeor94/which_decision_in_your_workflow_actually_needs_an