AI in manufacturing: Building resilience for a disruptive world

Manufacturing leaders aren’t dealing with one challenge. They’re managing several at once.

Geopolitical uncertainty is reshaping trade flows and supplier networks. Supply chains remain unpredictable, while ESG and sustainability expectations continue to rise. Legacy operational technology and production systems are being pushed beyond what they were designed for, exposing new cyber and operational risks. At the same time, experienced workers are retiring, taking critical process and operational knowledge with them, leaving organizations to close growing capability gaps.

For many organizations, this can feel like a constant game of catch-up—adjusting plans, responding to disruptions and trying to keep operations stable.

But reacting isn’t enough anymore. Manufacturing leaders are starting to design operations that can handle disruption from the outset. The priority is shifting toward resilient operations that can absorb shocks, support growth aligned with ESG goals and respond faster to changes in demand.

Increasingly, AI is playing a key role here—not as a standalone solution, but as part of how teams plan, monitor and act across the value chain.

Resilience by design starts with unified visibility

When teams talk about improving forecast accuracy and planning capacity, the first question often is: Which data points matter most?

The most significant gains don’t come from accumulating more data. They come from connecting the right signals and using them in context.

Manufacturers today need a clear view across multiple fronts. Customers expect hyper-personalized products, shorter lead times and visible ESG and sustainability commitments. Employees expect intuitive digital experiences that match their experiences as consumers. And partners across complex ecosystems expect trust, transparency and reliable, secure data exchange.

Forecast and planning accuracy improves when organizations stop chasing volume and focus on decision-grade data. Leading indicators matter far more than lagging KPIs.

• Internally, this includes demand signals such as order patterns and volatility; supply signals such as supplier lead times and capacity constraints; and operational signals like asset availability, maintenance status, labor constraints and yield.
• Externally, geopolitical risks, weather events, energy availability and logistics congestion shape outcomes.

Connecting these signals changes how teams respond. When demand signals are viewed alongside real constraints, teams can see issues earlier and respond with options.

Our experience shows that real value comes from linking the right signals across the ecosystem and using AI to turn them into early warnings, rather than after-the-fact analysis.

Scenario planning: From the “right plan” to making better trade-offs

When disruption hits, planning quickly becomes a balancing act.

Do you protect service levels or reduce inventory?
Do you prioritize speed or manage cost?

Traditional planning tends to treat these as separate decisions. In reality, they’re connected.

AI helps bring those trade-offs into the open. It allows teams to test different scenarios—what happens if demands spikes, if a supplier fails to deliver or if lead times stretch—and see the impact of decisions across service, cost and risk at the same time.

For this to work, teams need clear guardrails upfront: minimum service levels, maximum working capital exposure and acceptable risk thresholds. AI does not replace decision-makers. It helps surface the consequences of choices early and clearly.

This shift changes the conversation from “What is the right plan?” to “Which trade-off are we willing to make?”

In a crisis, speed and alignment matter more than theoretical accuracy. AI enables teams to rapidly model the impacts of disruption, compare alternatives, and act based on risk appetite rather than intuition.

For a deeper look at how teams are tackling more complex supply chain decisions, this blog from my colleagues, “Optimization, generative AI and quantum - Making complex supply chain decisions faster and more accessible,” explores how these approaches come together in practice:

Smarter sourcing: Where AI delivers fast, and lasting value

Sourcing today is becoming more data-driven, not in theory, but in practical, day-to-day operations.

AI is delivering some of the fastest results in procurement, especially where decisions are frequent, rules-based and supported by existing data. This is particularly true in intake-to-pay processes and in compliance and risk sensing.

These are operational areas where impact shows up quickly. AI helps reduce cycle times, cut errors and ease manual effort through capabilities like natural-language intake, automated classification, intelligent approval routing and invoice anomaly detection.

AI also improves spend classification and visibility by normalizing data across ERP and procurement systems and highlighting leakage in near-real time. At the same time, continuous risk monitoring scans suppliers and sub-tiers for early signals related to financial stress, geopolitical exposure, lead-time variability and ESG compliance issues.

In many cases, teams describe this as “found money”—value that was already there but hard to see.

Supplier discovery is another area that benefits significantly, especially in constrained categories or high-risk regions. AI can rapidly scan long-tail suppliers, match capabilities, pre-screen for risk and ESG criteria, and generate shortlists in days instead of months.

Quick wins matter, but leadership looks at long-term value.

Blending AI, ESG and human judgment

Effective decision-making relies on clearly defined roles. Buyers provide context, policy sets the boundaries, and AI offers recommendations. AI should support decisions, not act as an automated gatekeeper.

AI can continuously score and monitor risk and ESG indicators, while buyers interpret context, manage relationships and override recommendations where necessary. Policy defines thresholds, escalation rules and governance.

Initiatives such as Catena-X and the Digital Product Passports, demonstrate how federated data ecosystems support trust, end-to-end supply chain visibility, ESG traceability and compliance reporting—all while keeping accountability with people.

From pilot to production: Making AI part of everyday work

AI scales when it delivers value in daily work, not when models are technically impressive.

The most successful organizations start with use cases that solve real problems and fit within existing workflows—for instance, supplier discovery, demand forecasting, lead-time prediction, risk sensing, and inventory optimization.

They also define success upfront, as business outcomes:

• Fewer disruptions
• Improved on-time delivery
• Shorter lead times
• Lower total cost of ownership
• Reduced inventory buffers without increasing service risk.

Executives are no longer asking whether to use AI; they are asking how it protects revenue, margins and continuity.

Embedding AI for long-term success requires three conditions:

1. AI is integrated into core systems such as ERP, planning and sourcing systems.
2. Decision ownership is clear. AI delivers speed and consistency. Humans own judgment and exceptions.
3. Feedback loops improve model performance over time.

Governance as an enabler, not a barrier

There is a common concern that governance slows down AI. In practice, the opposite is true.

When governance is clear—who own what, how models are monitored and how decisions are traced—organizations move faster because they avoid rework and build trust early.

AI in manufacturing is becoming part of core operations, and like any critical capability, it needs the right structure around it. This includes clear ownership, secure deployment, explainability, and human oversight at key points.

Trust doesn’t come from the model itself; it comes from how it is used.

Looking ahead

Over the next decade, disruption will remain constant in manufacturing. What will change is how prepared organizations are to respond.

The manufacturers that move forward will be the ones that treat AI, data, and emerging technologies as levers for renewal, not just incremental improvement. They’ll focus on unifying their foundations, scaling AI responsibly and staying ready for what comes next. The result will be stronger customer value, more resilient ecosystems and growth that lasts.

If you’re looking to explore how AI can strengthen resilience in your manufacturing operations, I’d be glad to connect for a discussion.