Armanino Q&A: AI is Reshaping Manufacturing Supply Chains

As tariff uncertainty continues, supply chain leaders face mounting pressure to balance operational efficiency with risk mitigation. Traditional approaches such as stockpiling inventory tie up working capital and create bottlenecks, making agility essential.

AI is becoming crucial for just-in-time supply chain management. The technology combines historical data with real-time intelligence on pricing, geopolitics and supplier performance, enabling supply chain teams to make more precise decisions on ordering timing, quantities and resource deployment.

The goal is to shift from reactive purchasing to strategic supply chain planning, moving beyond instinct-led decisions towards data-driven clarity.

Bryan Graiff, M&D Midwest Industry Leader, and Amy Julian, Strategy and Transformation Partner at Armanino, explain how AI is reshaping inventory models. The technology translates insight into ROI through improved demand forecasting, tighter cost control and more resilient supply chain planning.

For supply chain professionals, this means enhanced visibility across the procurement cycle and stronger supplier relationship management, ultimately building more adaptive networks capable of responding to market volatility.

How should just-in-time evolve amid tariff volatility?

Amy Julian

Just-in-time was designed to keep carrying costs low by limiting how much inventory a company holds. This approach began even before the recent tariff volatility, during the COVID-19 pandemic, when companies got a harsh wake-up call about how quickly those lean inventories turn into a massive liability when supply chains slowed, transportation costs surged and production shut down. The tariff noise added to a broader period of uncertainty, pushing companies to reassess what just-in-time should look like today.

One of the most powerful ways just-in-time supply chains can adapt to tariff volatility is by strengthening demand signals. Having a clear and accurate view of what customers actually want is critical right now. This means improving forecasting, gaining closer visibility into customer orders and developing pricing strategies that can adjust quickly as costs fluctuate.

Instead of relying on a single baseline plan, teams should run scenarios and identify where it makes sense to hold additional inventory, allowing them to maintain steadier prices and avoid stockouts. This is especially beneficial for high-velocity SKUs or items with volatility-prone inputs. For these items, the industry trend is moving toward the use of strategic buffers (JIC). This helps balance availability and service levels for the right items. 

Another way companies can redesign just-in-time is by diversifying their supplier base. This isn’t about companies necessarily reshoring production back home; heavy localisation can reduce resilience and increase supply risk. Supplier diversification has improved resilience, especially for bottleneck inputs. To understand and optimise this, organisations are investing in mapping at least two Tier-2 suppliers and beginning to track risks based on supplier locations.

This combination of stronger demand planning and supply diversification allows companies to keep many of the cost-saving benefits of the original just-in-time model for stable/high velocity SKUs with diversified supply bases with JIC buffers for long-lead or volatile items, building in the kind of resilience they need when tariffs or other global disruptions mess up the flow of goods.  

When policy shifts, which external signals should AI prioritise to recalibrate supply/inventory decisions?

Bryan Graiff

AI can help companies monitor the external signals that influence demand, but the “right” inputs really depend on the industry and the company’s position in the supply chain.

We're seeing a lot of organisations already using fairly simple AI agents that run on a daily basis. They scan for relevant news and automatically notify teams when anything pops up that could affect their supply, inventory, or demand.

Interest rate movement is one of the most important signals to track, since it influences. how willing consumers are to buy. Labour and jobs data also matter because shifts in employment usually give us the first clue about how fast customers might pull back on discretionary spending.

If companies have access to more advanced AI, they can actually feed policy-related signals directly into their demand forecasts and build full scenarios showing how those changes might affect sales. Others might use AI to monitor publicly available competitor pricing to understand how the market is responding, although ensuring compliance for competitor-price intelligence is critical to review prior to implementing any model.

The best part is that you don't always need some massive, complex system. Often, simple agents paired with well-designed prompts can automate all this monitoring.

That's the real win—it helps teams react significantly faster when policy or market conditions shift unexpectedly. Industry leaders incorporate these initial AI-driven insights with strong data models where supplier IDs are aligned across all systems (ERP/WMS/TMS) and there’s a supplier base mapped down to at least a Tier 2 supplier base to enable faster decision-making. 

How can AI cut risk without tying up cash in inventory?

Amy Julian

The core problem here is that companies want to reduce risk in their supply chains, but the old solution, holding excess inventory, ties up cash. 

AI can help bridge that gap by acting like an analyst that can surface issues and opportunities much faster than a human team could. Instead of manually reviewing production data, inventory turns and demand forecasts over several days, AI can scan that information routinely and highlight what matters most. For example, it can flag the SKUs that are selling quickly and recommend ordering more before they become a problem, helping companies avoid stockouts without overbuying. 

AI can also spot inventory that’s aging or at risk of becoming obsolete. By detecting those patterns early, the system can suggest pricing adjustments or other actions to move slow-moving items before they become a loss. It doesn't matter if the data is buried in an ERP system or scattered across spreadsheets; AI can consistently pull it all together, giving teams clearer and more timely visibility into exactly where they face risk. The ability to implement is dependent on a strong data model.

In short, these AI strategies can help companies carry the right products at the right time. It reduces the likelihood of surprise shortages or excess inventory, all without forcing them to tie up extra cash on the shelves.

What guiding principles ensure trustworthy, auditable AI in supply chains?

Bryan Graiff

For AI to be trustworthy and auditable in supply chains, companies need clear visibility into how the tool arrives at its conclusions. Essentially, it's about asking yourself: Can I trace my data back to what’s actually happening in my business?

Inventory processes already require auditors to match what’s in the system to what exists on the warehouse floor and the same expectation applies when AI assists with forecasting, ordering or reporting. Auditors aren’t just going to look at the final output, they’ll also want to understand the data lineage and logic behind it. Modern tools make this easier because AI can “self-report,” outlining its model versioning, tracking approvals or human reviews and show its reasoning so teams can trace results all the way back to the source system.

Equally important is replicability. Companies must be able to take the AI’s data and logic and replicate the result themselves. Replicability means that given the same inputs and configuration, the model will consistently provide the same results. Even when AI provides a clear explanation, teams shouldn’t rely on it at face value; they should be able to double-check the math, assumptions and outcomes. Baking this transparency into agents or internal models ensures the AI remains auditable and aligned with expectations for financial reporting.  

Which KPIs best measure resilience rather than pure efficiency?

Amy Julian

Over the past few years, supply chains have all experienced a level of uncertainty. Between COVID-19 and tariff volatility, it’s become clear that resilience isn’t just defined by efficiency, but by how quickly a supply chain can adapt when conditions shift. The buzzword for this right now is resiliency. Traditional KPIs, such as inventory turns or tight just-in-time performance, are still key metrics, but they don’t tell the full story today. The metrics that reflect true resilience focus more on flexibility, reactiveness and supplier strength.

A big part of this comes down to supplier health. Metrics such as supplier concentration index, supplier risk scores and the number of qualified alternative suppliers for core components show whether a company can pivot when costs or policies change. 

Another key indicator is speed. How quickly can a company adjust order volumes, shift suppliers, or realign inventory levels when demand signals move? Metrics such as Time-to-Recover (TTR) and Time-to-Survive (TTS) have evolved to highlight what this could look like. 

Inventory-related measures also help offer insight into resilience. Tracking near-overage or aging inventory, or the cost to expedite it per unit of revenue, shows how well demand was forecasted. Ultimately, fewer stockouts and fewer cases of excess product indicate stronger demand planning and better responsiveness to volatile periods.

How can organisations take action?

Amy Julian

Organisations that have completed this journey have shown double-digit improvements in product stoppages by identifying and acting on risk signals days earlier. We recommend three key steps that can be managed at any size organisation.

1. Start with mapping the supply chain and identifying key risk points. This should be at least a level 2 and potentially extend deeper into the supply chain for critical components. Identify the types of risks that each supplier node is exposed to and incorporate them into your strategy. Identify indicators or sources of data that can help monitor for risk. 
2. Build or strengthen your internal data capabilities. This could involve setting clear data governance rules or building a data warehouse focused on managing and enriching only the information needed to make informed business decisions quickly.
3. Set decision frameworks and a governance plan. This should incorporate resilience KPIs that monitor information and data, notify the team and provide options for action.