Transformation is an operating change, not a software purchase

The operating reality

The evidence is visible in every stalled program. Data moves to the cloud, yet inventory is still defined three ways across business units. An AI pilot ships, yet plant managers ignore its output. ERP is modernized, yet the broken processes survive the migration intact.

Three implications for leadership

First, ownership cannot sit with the technology team alone. A data and AI program touches finance, supply chain, engineering and operations within weeks. The programs that scale share accountability across all of them. Leaders modernize core systems, centralize governance and embed AI into production and planning, while laggards stay stuck in isolated pilots.

Second, sequence from the decision, not the platform. The debate between data platforms is real but secondary. The first question is which business decision has to improve: predicting equipment failure, sharpening demand forecasts, isolating a quality defect, giving finance and operations one trusted number, integrating an acquisition faster. Define the outcome, then derive the data, architecture and talent.

Third, the team has to match the ambition. A few general data engineers will not deliver an enterprise program. Most require a transformation leader who ties investment to outcomes, cloud and security architects, ERP specialists, data engineers, data scientists and ML engineers, manufacturing domain experts, governance leaders and change leaders. The mix depends on the current team, the platform choice and the stage.

The leadership test

Strong leaders sequence the work. They set the foundation, pick a first problem that matters, build credibility by delivering and create a model they can repeat. They also know where to build internally and where a partner accelerates, without letting the manufacturer become permanently dependent on outsiders for its own architecture and operating knowledge.