The Model Training Methodology Behind Blue Yonder's Supply Chain AI

Not all supply chain AI is built for the work supply chains demand. As AI moves from answering questions to driving workflows, the difference between a generic LLM and a domain-trained model becomes critical. In this webinar, we will show how Blue Yonder, in partnership with NVIDIA, is building a repeatable Model Training Factory to fine-tune specialized AI agents for real supply chain decisions, with the precision, speed, and cost profile autonomous operations require.

This session is designed for supply chain, operations, technology, and innovation leaders who want to understand what it takes to move from AI experimentation to trusted execution in complex environments.

Here's what you'll gain from this on-demand session:

• Understand why domain-trained models outperform generic LLMs in high-stakes supply chain workflows
• See how the Model Training Factory fine-tunes, tests, and evaluates specialized agents for specific operational tasks
• Explore how agentic AI enables faster, more precise decision-making across warehouses, planning, transportation, merchandising, and network operations
• Learn why a hybrid model approach helps balance scale, cost, and workflow-level accuracy
• Examine early real-world applications in warehouse workflows, including allocation shorts, inventory exceptions, due-time urgency, and yard and receiving decisions

We also discuss the latest developments in the Blue Yonder and NVIDIA partnership, including how NVIDIA Nemotron and NVIDIA NeMo agent Toolkit support trusted agent development and repeatable model improvement.

What makes this session especially relevant is its focus on how AI gets operationalized, not just discussed. You will leave with a clearer view of how trained models, evaluated workflows, and governed guardrails can help build the next generation of autonomous supply chains.