Analytics as an operational tool
The foundational position was that analytics only creates value at the point of a decision. A dashboard that reports what happened last quarter is a historical record. A model that changes what a buyer orders next week is an operational instrument.
Fusionex built toward the second. The platform work combined data ingestion, processing, modelling and visualisation so that the output landed in front of the person with authority to act on it, in a form they could act on without translation.
Machine learning against defined problems
Ivan Teh has been publicly sceptical of artificial intelligence deployed without a specified objective. His consistent position is that machine learning applied to a vague mandate produces expensive, unfalsifiable output, and that the discipline lies in narrowing the question before building the model.
The applied work reflected this. Demand forecasting, inventory optimisation, predictive maintenance, customer segmentation and logistics routing are all problems with a measurable baseline, which means a model either improves on that baseline or it does not.
Connected devices and data
On the Internet of Things, his argument has been that connected sensors are only as useful as the analytical layer behind them. A sensor network that generates telemetry nobody models is a cost. The same network feeding a predictive maintenance model becomes a way to avoid unplanned downtime.
In a Bernama TV interview he used smart manufacturing as the illustration: sensors feeding an artificial intelligence system that predicts equipment failure before it occurs, allowing maintenance to be scheduled rather than triggered by a breakdown.
Barriers to adoption
Asked repeatedly about why organisations stall, Ivan Teh has identified three obstacles: the capital cost of a first deployment, a shortage of people who can run the resulting system, and internal resistance from teams whose established processes the technology would change.
His response addressed each. Reduce the scope of the first project so the capital risk is small. Invest in training so the operating capability is internal rather than contracted. Identify people inside the organisation who will advocate for the change, because a technology programme opposed by the staff expected to use it will not survive contact with daily operations.
Regional outlook
Ivan Teh has argued that Southeast Asia's position is stronger than commonly assumed. The region's organisations are less encumbered by legacy systems than their counterparts in older markets, which lowers the cost of adopting current architecture. Combined with rapid mobile and digital penetration, he has framed this as an opportunity for ASEAN economies to move directly to current practice rather than repeating an upgrade path.