Permanent volatility is forcing manufacturers to rethink supply chain planning

Supply chain disruption has become a permanent feature of manufacturing rather than an exceptional event to be managed until conditions return to normal. The harder question is how quickly manufacturers should react. Respond too slowly and customer demand, material availability or transport constraints may move on before the plan catches up. Respond too aggressively and the supply chain can push instability directly into production.

For Simon Bowes, CVP, Supply Chain Advisory at Blue Yonder, that tension is becoming more pronounced as manufacturers are expected to make decisions over shorter timescales. External shocks remain important, but their impact is magnified when customers and competitors increasingly expect businesses to respond while there is still time to change the outcome.

“Everyone talks about volatility, whether that is tariffs, the Straits of Hormuz or the Red Sea, and all of those things are creating more disruption,” Bowes says. “But lead times and frozen periods, where the production plan is effectively locked and cannot easily be changed, are reducing, which means that volatility is felt more directly in the planning process. If you have a two-month frozen period, there is not much you can do about it. Now there is competitive pressure to be much more reactive and responsive.”

Those shorter decision cycles change the purpose of supply chain planning. Traditional processes were often built around relatively stable handoffs between demand planning, supply planning, production, warehousing and transportation, with overnight or weekly batch processes moving information from one function to the next. When decisions need to be reconsidered within hours, those delays become part of the problem.

Planning cannot stop at the factory gate

The greatest risk is that supply chain responsiveness becomes somebody else’s production problem. A planning system may identify an opportunity to meet additional demand, rebalance inventory or shift production between sites, but the factory still has changeover times, labor constraints, maintenance schedules and operating targets that cannot simply be optimized away.

A more responsive supply chain can expose tensions that remain hidden when planning cycles are long and stable. A decision that makes sense for the wider business may require an individual plant or function to accept disruption that it would normally try to avoid.

“Production is typically based on efficiency, overhead recovery and all of the things that mean you do not want to change over,” Bowes says. “Changing over is an inefficiency, and that impacts the KPIs of the production plant. Companies are recognizing that they need to be more responsive and deal with volatility, but they are still slow to reflect that in the management processes and the way the business works.”

Decision orchestration becomes important because the best answer for one function can be the wrong answer for the business as a whole. Increasing production may appear feasible until warehouse capacity, transportation availability or additional changeover costs are considered. Conversely, a plant may reject a late customer request because it disrupts its schedule even when the revenue opportunity outweighs the manufacturing penalty.

The objective is not to let an algorithm continually rewrite the production schedule. It is to make the consequences visible early enough that different functions can evaluate the same scenario using current information. Production can see the effect on changeovers, while supply chain teams can assess service and revenue implications before anyone commits to the change.

Knauf provides an example of that direction. The building materials manufacturer, which operates more than 300 production sites across over 90 countries, is deploying Blue Yonder Demand Planning as part of a 2032 strategy to digitalize, orchestrate and automate supply chain processes across its three global business units. The program is intended to combine real-time data with scenario planning, market intelligence and machine learning so that changes in demand can be assessed more quickly and reflected in planning decisions. Knauf expects the approach to improve forecast accuracy, optimize inventory at SKU level, reduce excess stock and working capital, and shorten forecasting cycles so the business can respond faster to changes in customer demand.

Autonomy starts with better decisions

As planning cycles accelerate, manufacturers need ways to explore alternative responses without committing production to every possible change. The practical question is how much of that analysis AI can perform before responsibility passes back to the planner.

“What we are enabling at this point is agent-supported, what-if planning and scenario-based planning,” Bowes adds. “We are not seeing completely autonomous planning trusted yet by companies. They want to understand what would happen if demand changed, if they ran a promotion or if production capacity was taken down, and then decide what they want to do about it.”

A supply chain decision rarely ends where it starts. An unexpected request for significantly more product may require someone to check plant capacity, material availability, transportation requirements and warehouse space before a decision can be made. That manual chain can consume valuable time while the commercial opportunity continues to change.

“Instead of you working out the synthesis of your problem into all the different things you need to do, you can give the problem to the agent, and the agent will do those things,” Bowes continues. “The promise is that it can make those changes, but I do not think we are yet at the point where everyone says, ‘Do not bother asking me.’ There is still a desire for confirmation before the action is taken.”

The immediate benefit is reducing the latency created when planners manually navigate between applications, run reports and pass information between colleagues before the organization can understand its options. Faster analysis also gives manufacturers more time to consider whether a proposed response is genuinely worth the disruption it may create elsewhere.

Reducing the time needed to investigate a supply chain problem also creates an opportunity to bring changing market conditions into the decision earlier. Bowes points to demand models that can incorporate shifts in economic indicators, market behavior and other external influences rather than waiting for conventional statistical processes to identify a trend from accumulated historical data. When demand and supply planning are connected, manufacturers can then see the operational implications of those changes at the same time instead of waiting for one planning function to hand its revised assumptions to another.

Resilience requires different operating rules

Faster planning can expose trade-offs that were previously hidden by long planning cycles. A plant may be asked to make smaller batches, perform more changeovers or manufacture products that are less attractive from a local efficiency perspective because the overall business benefits from responding to a customer opportunity.

Greater responsiveness can therefore create friction inside the business as functions are asked to behave differently from the way they have been managed historically. Unless that tension is addressed at leadership level, local teams have little reason to sacrifice their own performance for a wider supply chain objective. “You can say that the business strategy is to become more adaptive, responsive and resilient, but you cannot then refuse to challenge the status quo of how management is run, how KPIs are set and how people’s bonuses are paid,” Bowes says. “You have to revisit all of that.”

The willingness to delegate more consequential decisions will also depend on what happens when a recommendation is wrong. Planners remain accountable for inventory, service and production performance, giving them good reason to be cautious about handing control to a system they cannot interrogate.

“Explainability is really important,” Bowes says. “You learn to trust it because you have tried it, it came up with the right answer and it explained why it did that. If it is a total black box that simply says, ‘Trust me, you should accept that order,’ it is very hard to build trust because it only needs to go wrong once for everyone to say they cannot trust it.”

The future of supply chain autonomy may therefore be less about handing control to a single intelligent system and more about steadily reducing the friction between decisions. Common data can allow demand, supply, warehouse, transport and production teams to work from the same plan even when they view it at different levels of detail and over different time horizons.

Bowes expects those boundaries to continue blurring as planning becomes more connected and agentic systems replace some of the work currently performed through menus, workbenches and manual handoffs. Over time, manufacturers may also rely less on internally maintained representations of suppliers and customers if information can move directly across a more interconnected supply chain.

Permanent volatility does not mean manufacturers should change production every time a signal moves. It means they need the ability to understand their options quickly, test the consequences across the wider operation and act when the business case justifies disruption. The manufacturers that become more resilient will be those that connect planning decisions without losing sight of the physical and organizational constraints that determine whether the factory can deliver.