Webinar Recap: Reducing Waste in the Age of AI — Smarter Inventory Decisions Across Industries

Summary

In our Future.Now webinar, “Reducing Waste in the Age of AI: Smarter Inventory Decisions Across Industries,” we explored how organizations can dramatically reduce excess and obsolete inventory by making faster, smarter decisions across the supply chain. The opportunity is significant. Most enterprises already have the data they need, but the gap between insight and action is wide. Forecasting tools, dashboards, and control towers have improved visibility, yet disruptions still trigger slow responses, exception queues never empty, and excess inventory quietly accumulates. The problem isn’t data or people. It’s the speed of decision-making.

We examined how agentic decision intelligence closes that gap by continuously sensing signals, generating explainable recommendations, and driving action in real time. A live platform demo showed how the Excess and Obsolete (E&O) Reduction Skill works in practice, from surfacing root-cause analysis and financial impact projections to enabling autonomous execution and continuous learning. A real-world case study from The Hershey Company illustrated how organizations are achieving measurable ROI in as little as 90 days, with fast, composable implementations that don’t require replacing existing systems.

Key Takeaways

• The execution gap is where waste lives.
  Most supply chains have improved their planning, but planning is a snapshot while reality is continuous. When reality diverges from plan, whether due to a late shipment, a demand spike, or a supplier failure, it often takes more than 72 hours to respond. Every day a decision is delayed, waste gets locked in. This execution gap is the single biggest source of value leakage in most supply chains.
• Inventory waste is a cross-industry problem, and it compounds over time.
  Excess and obsolete inventory tends to be top of mind for companies with perishable products, but organizations without hard expiry dates are often the ones with the biggest E&O problems. Because the issue isn’t urgent, it gets overlooked and snowballs until it becomes a major write-off or working capital burden that was entirely preventable.
• Decision intelligence sits between planning and execution, and doesn’t replace either. Rather than adding more dashboards or customizing ERP and APS systems beyond the point of diminishing returns, decision intelligence operates as a layer on top of existing tools. It continuously senses signals from across the enterprise, evaluates options using prediction, optimization, simulation, machine learning, and agentic reasoning, and then surfaces explainable recommendations or executes decisions autonomously before writing outcomes back to source systems.
• Explainability and the confidence score are what make automation possible.
  Every recommendation comes with a transparent confidence score, built from weighted contributors that show exactly how a recommendation was generated. The system tracks projected versus actual outcomes, and confidence improves over time as more decisions are made. Organizations can set thresholds, for example automating any recommendation above 95% confidence, so that human time is reserved for decisions that genuinely require it.
• Decisions should be made in dollars, not just quantities.
  A key shift that decision intelligence enables is moving from quantity-based recommendations to financially grounded ones. When every recommendation is dollarized, expressed in terms of margin, cost, working capital, or E&O impact, teams immediately understand the business value of each action. This drives adoption, builds trust, and makes it easier to demonstrate ROI to leadership.
• E&O reduction is an enterprise-wide effort, not just a planning problem.
  Preventing and reducing excess inventory requires coordination across procurement, sales, and marketing, not just supply chain planning. Decision intelligence enables that cross-functional view: recommending promotional strategies to accelerate demand, identifying rebalancing opportunities across the network, and surfacing component-level risks driven by finished goods underperformance, all in real time, without waiting for the next S&OP cycle.
• Start fast, prove value, then expand.
  Rapid, composable implementations, typically delivered in weeks rather than months, allow organizations to deploy in a targeted use case, accumulate measurable outcomes, and then expand into adjacent decision domains. The pattern is consistent: begin where the pain is loudest, demonstrate value quickly, and let the results build the business case for scale.

Speakers

Vincent Wicker, Senior Manager, Data & AI, Autonomous Supply Chain, Accenture Vincent specializes in autonomous supply chain transformation and brings over five years of experience helping organizations leverage AI-driven insights to optimize operations and drive efficiency. Based in Boston, he works with enterprise clients to align data, decisions, and workflows across fast-moving consumer goods, retail, healthcare, and high-tech industries.

Matthew Bunce, Senior Engagement Principal, Aera Technology Matthew focuses on enterprise transformation and decision intelligence at scale. He works across industries to help organizations shift from reactive processes to AI-driven, intelligent execution, bringing two decades of supply chain experience to each engagement.

Suraj Ramalingam, Senior Solution Engineer, Aera Technology Suraj brings extensive experience in global supply chain management, having held senior roles at Salesforce and Procter & Gamble. At Aera, Suraj focuses on enabling customer success through technology-driven transformation, helping organizations operationalize AI-driven decision-making at scale.

Full Recording

Access the full webinar recording here.

Q&A

At the end of the presentation, we held a short session dedicated to answering attendees’ questions. Below are several of the key questions and answers.

Q: What types of recommendations appear in the planning module?

A: The planning module surfaces recommendations focused on inventory rebalancing across the network. When there is excess inventory at one location, the system identifies other parts of the network where demand is stronger and recommends transferring products there. The goal is to move inventory closer to where it can be sold faster, reducing E&O risk without waiting for a manual review or the next planning cycle.

Q: How is the confidence score calculated, and how accurate is it?

A: The confidence score measures two things: the likelihood that a user will take the recommended action, and the likelihood that the action will deliver the projected value, such as an E&O reduction. Confidence improves over time as more decisions are made in the platform. 

Because most supply chain decisions operate on a planning horizon of several weeks, it takes time to register the actual outcome and compare it against what was projected. In a typical enterprise implementation, confidence scores start below 90% and increase as the machine learning model accumulates more data. If confidence scores aren’t climbing, the right approach is to investigate why users are dismissing recommendations and refine the underlying rules and heuristics accordingly.

Q: How does Aera handle tariff volatility and geopolitical risks?

A: Tariffs and geopolitical disruptions increase the velocity at which supply chains need to make decisions, and that is precisely where decision intelligence adds value. Aera supports real-time scenario modeling that allows organizations to evaluate alternative routes, sourcing options, and cost impacts quickly, without waiting multiple planning cycles to act. For example, if a routing change is needed, the system can immediately notify logistics providers of lead time changes. 

Aera also helps organizations track key metrics like time to detect supplier region exposure by SKU, the speed from a news event to a ranked SKU impact list, landed cost freshness, and the number of active resiliency scenarios with named triggers. When tariffs affect certain SKUs, substitute SKUs that aren’t on a blockage list may need to absorb the volume shift across tens of thousands of material locations. Aera handles the underlying calculations so that humans can focus on evaluating and executing the right decisions