Customer Case Study: Inventory Optimization
A leading manufacturer of IT products sees $19mn in savings from matching inventory levels to customer demand more accurately
A global technology conglomerate
Manufacturing
Inventory Optimization
$19 million savings
The Challenge
Demand is driven by many different factors, both at a macro level and with different drivers impacting individual products. Capturing the true drivers of demand for the necessary range of scenarios was not possible for demand planners when adapting to changing landscape conditions.
In some cases, the insufficient accuracy of demand planning models led to lost revenue through shortage of product supply, while overstocking led to unnecessary costs in manufacture and storage
Business stakeholders lacked trust in their existing AI model results due to the lack of clarity in how different models’ inputs impacted the results. Combining the best of their expertise, with the power of Causal AI to ensure the AI models met their requirements and addressed their demand planning limitations.
-
1Explainability
Business stake holders lacked trust in the existing model results
-
2Learning from correlations
Trend based approaches proved to be innacurate
-
3Simulating scenarios
Demand planners wanted to understand changes to demand under different conditions
The solution
Unable to produce trusted models effectively using traditional approaches, the customer’s data science team turned to causaLens to help build a more robust set of causal models that could identify the true causal drivers of demand and make accurate predictions. Leveraging the power of decisionOS the customer gained access to:
- Models that combine data derived insights and knowledge from domain experts
- Fully transparent and explainable demand models
- Powerful decisionApps that allow business stakeholders to interact with and interrogate the models
Results and Benefits
The customer no longer has to rely on large spreadsheets that are hard to manage and difficult to read, making the demand planning process time-consuming and opaque. Using the apps built with decisionOS, the decision-makers were able to visualize the entire process and the relationships between different factors, providing increased transparency and explainability throughout the organization, and time-savings of up to 80% for data analysis.
The accuracy and robustness of the causal models powered by decisionOS enabled the company to improve its decision-making throughout the supply chain, leading to inventory levels that better match demand each quarter, resulting in $19M in savings annually
