Is the product being sold the setup of the bespoke models from their bag of what they’ve done before? Regressions seem like the obvious way to detect anomalies

Is the product being sold the setup of the bespoke models from their bag of what they’ve done before? Regressions seem like the obvious way to detect anomalies to me since it should be 100% repeatable and make sense according to amount of heat being generated and removed , how to apply ML/AI to it I am not so sure