Topic 1
Who chooses the next step?
Every multi-step AI system answers one question: who decides what happens next, your code or the model? Anthropic's "Building effective agents" (2024) uses that question to split agentic systems in two:
- Workflows: models and tools orchestrated through predefined code paths. The model does individual steps (classify, extract, draft); code decides their order.
- Agents: the model dynamically directs its own process and tool use, deciding how to accomplish the task.
Predict first
Scroll sideways to see the whole diagram
Anthropic: workflows are systems where models and tools are orchestrated through predefined code paths. The model does steps; code decides their order.
Text version
- Anthropic: workflows are systems where models and tools are orchestrated through predefined code paths. The model does steps; code decides their order.
- Workflows offer predictability and consistency for well-defined tasks.
- Agents are systems where the model dynamically directs its own process and tool use.
- That suits open-ended problems where the number of steps can't be predicted, at the price of higher cost and the potential for compounding errors.
The two aren't better and worse; they suit different problems. Anthropic says workflows offer predictability and consistency for well-defined tasks, while agents are the better option when flexibility and model-driven decisions are needed. Agents suit open-ended problems where the number of steps can't be predicted, and their autonomy means higher costs and the potential for compounding errors.
Most real systems mix them: a fixed workflow with one agentic step, or an agent whose risky actions run through fixed code. The useful habit is to decide, for each step, whether it needs a model's judgment or just needs doing.
Quick check