Agents of change: How AI is helping automotive supply chains recover value and reduce waste

Artificial intelligence is rapidly moving beyond general-purpose tools toward specialised systems designed for specific industries and workflows. Gartner predicts that by 2028, more than half of enterprise generative AI models will be domain-specific, reflecting growing demand for AI systems built around the realities of a particular business rather than generic use cases.

Few industries illustrate the need for that specialisation better than automotive.

Automotive manufacturing operates through a vast $4 trillion supplier ecosystem spanning dozens of specialised manufacturing and component domains. Small changes to demand, production schedules, engineering requirements or logistics constraints can ripple across suppliers, inventory, labour and operations networks, often creating hidden costs and profit leakage throughout the value chain.

For many organisations, the greatest opportunities to improve profitability no longer reside solely within factory walls. They exist within the complex interactions between customers, suppliers, logistics providers and planning teams.

“The reason we went after automotive in the first place is because automotive has a very large network with a distributed problem,” says Gaurav Palta, founder and CEO of SupplyWhy.ai. “But most importantly, from a profitability and value perspective, we chose automotive primarily because of its inherent ‘hair-on-fire’ problems.” 

That challenge resonated with the team at Yazaki Innovations (YII), which has been exploring how AI can improve planning, service, responsiveness and operational performance across automotive supply chains.

“We've always believed innovation has to start with solving real operational problems,” says Jeff Gallo, director of program management, Strategic Partnerships Group at YII. “The opportunity with AI is not simply automation. It's helping teams understand signals faster, make better decisions and respond more effectively when conditions change.”  

That shared view of AI as a decision-support capability became the foundation for the collaboration between Yazaki and SupplyWhy.ai, bringing together Yazaki's operational expertise and SupplyWhy.ai's domain-specific AI capabilities. 

Multi-agent intelligence for EBITDA performance 

SupplyWhy.ai's AI-powered platform uses a multi-agent system called Jenae that acts as a decision intelligence layer alongside ERP systems, EDI transactions and traditional planning tools. The platform is designed to identify distributed problems not just within a specific facility, but as they propagate across multiple tiers of the supply chain.

Automotive supply chain decisions require industry-specific knowledge, commercial rules and operational context. Jenae was purpose-built for automotive, incorporating the workflows, data structures and decision patterns that drive supply chain performance.

Jenae acts as a supervisory intelligence system overseeing supply chain planning, risk response, claims recovery and profit intelligence. Her objective is to identify patterns of EBITDA leakage across the network, explain root causes and help organisations take action before losses compound.

Palta argues that every automotive sub-sector has its own profit leakage profile. Whether it is wire harnesses, interiors, electronics or heavy manufacturing, operational behaviours influence profitability in different ways.

Wire harness manufacturing illustrates the challenge. A small demand change can quickly cascade into labour inefficiencies, inventory exposure and operational cost, while other automotive sectors experience different forms of profit leakage.

A key challenge is understanding what changed, who caused the change and how the financial impact should be allocated. Historically, this has required significant manual effort, making root-cause analysis difficult and often delaying corrective action.

The importance of understanding why

Tier suppliers receive production targets and align resources, inventory and labour accordingly. Yet customer forecasts often differ from actual consumption. When demand changes, suppliers must determine what happened, why it happened and what response is appropriate. Historically, this process has relied heavily on spreadsheets, meetings and manual analysis.

The scale of the challenge becomes clear in large automotive operations. At YII, planners manage a large number of finished goods parts and often thousands of supporting components. Demand signals can experience significant volatility, creating downstream implications for inventory, labour, purchasing and production planning. In this environment, understanding why a change occurred is often more important than detecting that it occurred.

SupplyWhy.ai addresses this challenge by combining internal operational data with external factors such as tariffs, market dynamics and geopolitical disruption. Through its Profit Intelligence capabilities, Jenae helps users understand not only what is likely to happen, but why.

“The demand signal has to be very focused on the ‘why’. The why is a lot more important than the ‘what’,” Palta explains.

Jenae aligns stakeholders around shared reasoning and continuously evaluates the operational and financial consequences of decisions, identifying patterns of margin erosion from inventory, freight, labour inefficiencies, tariffs and under-recovered claims.

The objective is not simply to improve forecast accuracy. It is to help leaders understand where operating profit is leaking from the business and prioritise interventions based on financial impact.

Claims recovery and financial forensics

Jenae's Claims Recovery capability directly links operational events to financial accountability. Engineering changes, release volatility and production disruptions frequently create recovery opportunities, yet many organisations struggle to identify and pursue them effectively.

As Palta explains: “Everyone is incentivised to understand the reason behind a change and whether it is adding value or subtracting it from the network. A lot of this is done through root-cause analysis which digs into the context.”

The platform's Financial Forensics capabilities help establish the causal chain linking operational decisions to financial outcomes, creating a clear justification trail and accelerating recovery.

SupplyWhy.ai also supports customers through its FireFight capability, designed for the frozen planning horizon where changes are not supposed to occur but often do. At that stage, the challenge is no longer prediction, it is response.

“Agentic AI becomes less about predicting what is going to happen and more about responding very quickly to a change, seeing what is feasible and what it is going to cost,” Palta notes.

The greatest value often comes from avoiding the cost altogether. Early identification of demand volatility, material shortages, capacity constraints and production risks helps organisations prevent premium freight, excess inventory, overtime labour, extraordinary expense and line-down risk before they occur. Claims recovery remains important, but preventing EBITDA leakage in the first place is even more valuable.

Jenae helps quantify impact, identify causality and document supporting evidence, enabling both proactive response and faster resolution.

Better planning at Yazaki

Following a successful 12-month pilot, YII has implemented Jenae across demand planning and operational decision-support workflows, with capabilities extending into claims recovery and disruption response.

“One of our business units is using it fully implemented,” says Gallo. “We are noticing that from the demand side signal, we are seeing a significant improvement in our teams’ decisions on what we think the signal should be.”

For Yazaki, better signals are not the outcome. They are the starting point. Improved signal quality enables better labour deployment, tighter inventory control, more effective capacity utilisation and faster response to customer changes.

“We are seeing a reduction in overall inventory levels and better planning of manpower based on the signals,” says Gallo.

Production planners use Jenae to compare release data, analyse changes and assess impacts before making decisions.

“All the same people are involved doing the work. They just have better data at their fingertips and are able to make better decisions more quickly,” says Gallo.

The exchange of reasoning

Data quality, governance and standards remain important, but both SupplyWhy.ai and Yazaki believe the next evolution of supply chain technology is not simply the exchange of data, it is the exchange of reasoning.

As agentic AI matures, organisations will increasingly rely on intelligent agents not only to report events but also to explain causes, coordinate actions, recover losses and continuously improve supply chain profitability.

“At the end of the day our goal is to have agents sitting across two enterprises being able to resolve and come to an agreement,” says Palta. “That is what we need to get to if we are to solve this network-wide harmonisation problem.”

By helping organisations detect, explain and respond to patterns of EBITDA leakage, SupplyWhy.ai and Yazaki are demonstrating how agentic AI can move beyond visibility and become an active participant in supply chain performance, resilience and profitability.