Act

An analysis is useful when it helps you decide what to check or do next. Start from the evidence in your notebook, distinguish observations from hypotheses, and keep enough context for someone else to assess the conclusion.

​

Record what you found

Ask the assistant to summarize the investigation:

Summarize the observations, the evidence for each possible explanation, and what remains uncertain. Suggest the next checks that would distinguish the explanations.

Review the summary against the underlying data. Save the notebook, then record confirmed findings and relevant context in the sample or project’s notes. Include the notebook URL so colleagues can inspect the analysis.

​

Choose the next check

The next step might be reviewing a measurement setup, inspecting maintenance records, comparing another growth, or planning a follow-up experiment. Prefer a check that distinguishes the remaining explanations. A correlation or model output is evidence to assess, not proof that a particular adjustment will improve the process. Follow your team’s instrument and process procedures when deciding on recipe changes, interventions, or run disposition.

​

Reuse the investigation

If you repeat the same kind of analysis, create a reusable template. Replace one-off assumptions with clear inputs and check it on another appropriate growth before using it routinely. As new evidence arrives, update the analysis and save a version so its reasoning remains inspectable.

​

Monitoring and automatic detection

Use Monitor to review active and completed growths. Automatic detection must be enabled for your deployment. If it is not enabled, new streams and uploads do not generate detections. A notebook investigation does not itself configure continuous monitoring or notification channels. For enabled RHEED changepoint detection, see the reference. Contact Atomscale support to discuss detection availability or a process-control integration for your setup.

​

Learn from other applications

Case Studies describe specific applications and their conditions. The earlier-detection result in the RHEED example and the accuracy result in the MOCVD uniformity study come from different investigations; they are not general performance guarantees for an analysis or alert.