Module 01 Knowledge Check
5 questions. Pass mark 4 out of 5. Answer every question before checking the answer key below, then retry after reading the feedback.
1. What problem does a graph solve that a single agent loop does not?
- A. It makes the model smarter
- B. It makes control flow explicit and inspectable instead of hidden inside model reasoning
- C. It removes the need for tools
- D. It reduces token cost
2. A 'while the model decides what to do next' loop is hard to operate mainly because…
- A. It is slow
- B. Its behaviour is non-deterministic and hard to reproduce, test, or bound
- C. It cannot call tools
- D. It uses more memory
3. When is a graph overkill?
- A. Always
- B. When the task is a single prompt with one output and no branching or actions
- C. When there are tools
- D. When there is state
4. 'Agentic' should be understood as…
- A. A model capability
- B. A spectrum of autonomy - from suggesting, to acting with approval, to acting alone
- C. Any use of an LLM
- D. Multi-agent only
5. The main operational risk that grows with autonomy is…
- A. Latency
- B. The system can take consequential actions before a human can intervene
- C. Token cost
- D. Prompt length
---
Answer Key and Explanations
Check these only after attempting every question.
1. B - It makes control flow explicit and inspectable instead of hidden inside model reasoning
The value is visibility. When routing lives in a graph you can see, test, and constrain it; when it lives in a prompt you cannot.
2. B - Its behaviour is non-deterministic and hard to reproduce, test, or bound
You cannot write a reliable test for a path that is re-invented on every run.
3. B - When the task is a single prompt with one output and no branching or actions
Orchestration has a real cost in complexity. Reach for it when branching, actions, or recovery genuinely exist.
4. B - A spectrum of autonomy - from suggesting, to acting with approval, to acting alone
Autonomy is a design choice you set per step, not a property the model has or lacks.
5. B - The system can take consequential actions before a human can intervene
A wrong answer misleads; a wrong action changes the world. Controls must scale with what the workflow can do.
Practical Check
Apply this module to your own work: complete the module activity for *Why Graph Orchestration Exists*, then write one sentence naming what your result shows and one naming what it does not.
Strong Answer Pattern
A strong answer names the task, the evidence used, the check performed, and the remaining limitation. It avoids "proved", "guaranteed", or "always" unless the evidence genuinely supports it.