LangChain has published a case study on why Fleet uses both general-purpose chat and specialized agents. The setup is designed to support quick ad hoc tasks while also delegating recurring responsibilities to more focused agent workflows.
That distinction is becoming important as teams move beyond a single chatbot interface. Some work benefits from open-ended conversation, while repeated operational tasks need tighter instructions, memory, and evaluation.
The case study points to a practical pattern for AI adoption: use chat for exploration, but build specialized agents when the same job needs to be done reliably over time.