TECH & AI ABB’ s response combined anomaly detection, remaining useful life estimation and prescriptive analytics into a single programme. Early proofof-concept work had already shown it was possible to spot abnormal operating conditions and give engineers better visibility of emerging risks.
The most valuable element, Anindya suggests, was not flagging a problem but explaining it. Prescriptive analytics and causal analysis helped teams understand the likely root causes behind emerging issues, so decisions could be prioritised by condition, risk and business impact rather than guesswork.
The same principle can support defect detection more broadly. Because defects rarely arise from a single variable, ABB models individual
78 August 2026