From Blind Spots to Crucial Insights.
Case Study: Hitachi Energy
At a Glance:
Industry:
Energy and Power Systems Industry
Use Case:
Detecting supply chain risks by using advanced Data Science to analyze existing market data and uncovering hidden dependencies.
Technologies Used:
- Statistical Analysis
- Clustering and Correlation
- Machine Learning
- LLM (Generative Al)
Impact:
Web-based application giving Hitachi Supplier Chain Managers deep insights into anticipated stock movements
Demonstrated ability to ingest and analyze daily stock levels of 28,000 parts and over 150,000 news sources
Demonstrated ability to predict stock, find behaviour clusters, interpret and correlate global news to stock events
The Challenge
A critical chip, sourced from China five tiers deep, was suddenly unavailable due to geopolitical influences, exposing a lack of transparency in Hitachi Energy’s multi-tiered supply chain. This disruption highlighted the urgent need for better visibility and risk mitigation to prevent unforeseen shortages.
Our Approach
We developed an Al-driven system, Bocca, that predicts supply chain disruptions by analyzing market data, stock levels, and global news. Using machine learning and Large Language Models, Bocca uncovers hidden dependencies, anticipates shortages, and enables proactive decision-making for greater resilience.
What we uncovered
“Black Box” nature of modern supply chains
By tracking daily stock level changes, the Spur Insights team surfaced signals of shifting supply and demand. Deeper analysis of price and inventory trends across components revealed hidden interdependencies-pointing to shared suppliers, raw materials, production sources, or regions. These insights helped expose the opaque links within complex supply chains.
Hidden risks in supply chains through global news
A key part of our methodology involves processing global news feeds using a Large Language Model (LLM) to identify critical events, like factory closures or geopolitical shifts. These events are linked to stock level changes. For example, when a major chip manufacturer closes a factory, the system flags the event, assesses the potential impact on related components, and strengthens internal correlations if stock levels drop, enabling quicker detection of similar risks in the future.
The Results
Web-based application giving Hitachi Supplier Chain Managers deep insights into anticipated stock movements
Demonstrated ability to ingest and analyze daily stock & news data on 5000 key parts
Demonstrated ability to predict stock, find behaviour clusters, interpret and correlate global news to stock events
“Bocca uncovered hidden supply chain dependencies, enabling us to address risks proactively and strengthen our resilience. We are now better equipped to navigate disruptions and make smarter, data-backed decisions.”
— Nikolaus Gaebler, Head of Supply Chain Management
What’s Next?
While supply disruptions in the semiconductor market have eased post-COVID, the risk of future disruptions remains elevated. The ongoing decoupling from China is expected to eventually impact the electronics supply chain. Bocca has been instrumental in providing deeper visibility into the market dynamics operating “beneath the surface.” Hitachi plans to continue expanding and investing in the Bocca solution to de-risk their supply chain and strengthen their resilience moving forward.