Practice for Handling Batch Quality Degradation in Cognite CDF

Hi Everyone,

I have a question regarding how Cognite Data Fusion is typically used in a real industrial batch manufacturing environment.

Consider a batch production process in an FMCG or chemical plant where process parameters such as reactor temperature, pressure, agitator speed, flow rate, or pH begin to deviate from their normal operating ranges, indicating a potential degradation in final batch quality.

I would like to understand the typical Cognite workflow in this scenario.

1. How does Cognite detect or identify the potential quality degradation? Is it through rules, analytics, ML models, or contextualized data?
2. Can Cognite automatically trigger workflows, create events, or send notifications to the relevant production or process engineers?
3. Is the next recommended action generally reviewed and approved by an operator/engineer before execution, or are there real-world implementations where Cognite integrates with control systems to initiate corrective actions automatically?
4. How are Atlas AI, Workflows, Events, and Notifications typically used together in this type of use case?

I'm interested in understanding the best practices followed in actual industrial deployments rather than a theoretical workflow.

Any examples or insights would be greatly appreciated. Thank you!