1. The driver prompts our agent with details of the sample.
  2. The model interprets the prompt and passes on sample information to our code.
  3. Our code calculates parameters.
  4. The model presents the proposed parameters, explains them, and waits for approval.
  5. The driver approves, or adjusts the setup directly.
  6. Our code acquires images and analyzes them.
  7. The model explains the analysis results.
  8. The driver accepts the results, or restarts the loop.