Better Decisions Under Uncertainty: A Leadership Approach to Industrial AI

A leader gets a call: the budget has changed, a deployment is moving, or a location may no longer be available. The team needs to explain what those changes mean for its ability to deliver. Having the data is a starting point. Evaluating the options is the harder task.

Retired U.S. Air Force Major General Brent Baker recalled situations like these in a recent interview I hosted with him and Dr. Sanjay Jagdale, Vice President of Servigistics Technical Solutions at PTC. Their discussion offered a practical approach to industrial AI: prepare the digital foundation, use tools designed for the planning problem, and keep human judgment connected to the decisions.

Get the digital house in order

Baker emphasized that organizations need their digital house in order to benefit from AI. Information must be dependable enough to support analysis, and systems must be connected so that teams can understand the operation end to end.

That principle has particular weight in service supply chains. Inventory, repair activity, asset configuration, demand, and replenishment conditions together influence what the organization can deliver. Incomplete or disconnected information can undermine the analysis before a recommendation reaches a planner.

Preparing the foundation means understanding which information the planning process requires, how it is maintained, and how it represents the operation. It is work that gives AI something useful to reason about and planning tools something credible to evaluate.

Turn what-if questions into decision support

Baker described how difficult it could be for his teams to answer questions about changes in funding or deployments without the necessary tools. Scenario planning provides a way to investigate those questions before committing to a course of action.

What if aircraft deploy to a different location? What if a base becomes unavailable? What if funding increases or decreases? A useful analysis examines how each alternative affects support requirements and the outcomes leaders care about.

Jagdale explained the foundation for this work: a predictive digital twin represents the service network, supply and product hierarchies, and relevant demand patterns. Planning also needs to account for future fleet deployment and changes such as upgrades or retrofits.

Servigistics provides a platform for evaluating alternatives through modeling and what-if analysis. The analysis helps leaders understand the consequences of choices. Baker was clear that the information may not always be perfect; its value lies in helping people make more informed decisions under uncertainty.

Make AI useful through operational context

Jagdale distinguished between asking a general question and receiving an explanation grounded in a particular planning model. Two apparently similar parts may be stocked differently for reasons that become clear only when the model’s assumptions and relationships are considered.

He described how the Servigistics AI Assistant can draw on the context of a customer’s implementation to explain recommendations. This makes the answer more useful to a planner who needs to understand a specific result, rather than receive a general explanation of inventory management.

Language models can help interpret and explain information. The calculations needed to evaluate network dependencies, asset configurations, and fleet-level objectives require purpose-built forecasting and optimization tools. Bringing those capabilities together helps connect a user’s question to the underlying planning analysis.

Use automation to concentrate human attention

The discussion also looked ahead to agents taking actions on a planner’s behalf within a predefined context. Jagdale described the opportunity to automate suitable routine work while freeing planners to address exceptions that require creativity or information beyond the mathematical model.

A planner may know about an upcoming customer requirement or a supplier shutdown that has not yet been reflected in the system. That knowledge needs to influence the plan. Human expertise remains essential to defining priorities, introducing new context, and evaluating consequential choices.

Baker reinforced the importance of managing by exception. When someone is responsible for thousands of parts, the ability to identify what needs attention becomes a practical leadership issue. Automation can help focus that attention on the decisions most likely to affect operational performance.

For leaders considering AI, the discussion suggests a clear test: can the system use dependable operational context, evaluate alternatives with appropriate tools, explain its recommendations, and support the organization’s chosen level of human control? Those capabilities connect technology adoption to better decisions.

Industrial AI earns its value when it helps an organization respond thoughtfully to uncertainty. The opportunity is to give leaders and planners a stronger analytical foundation and more capacity to apply their expertise where it matters.

Watch the full conversation with retired U.S. Air Force Major General Brent Baker and Dr. Sanjay Jagdale to hear their perspectives on digital transformation, decision support, and the future of service supply chain planning.