At ICON 2026, Blue Yonder’s SVP of Generative AI, Chris Burchett, emphasized the critical engineering challenges in deploying AI for supply chains, advocating for a hybrid model approach that combines large frontier models with proprietary, smaller AI models tailored to supply chain needs.
- Hybrid AI architecture blends large frontier models and Blue Yonder’s owned domain models.
- Domain-specific models improve reliability by reducing dependence on shifting general-purpose models.
- AI frees human workers for higher-value tasks, introducing new roles like ‘agent wranglers’.
What happened
Blue Yonder’s AI leadership presented a candid overview of their generative AI development during ICON 2026, focusing on building dependable intelligence layers for supply chains. Chris Burchett, SVP of Generative AI, outlined the rationale behind creating smaller, domain-specific models in partnership with NVIDIA to handle supply chain tasks more reliably than general-purpose frontier models.
The hybrid AI framework uses large frontier models to interpret human intent and proprietary, specialized models to execute detailed supply chain operations. This approach addresses reliability and context sensitivity issues driven by continuous retraining and quantizing of large models, which can cause inconsistent performance for enterprise applications.
Why it matters
The supply chain sector requires consistent AI behavior to maintain operational integrity, especially as disruptions or unique orders frequently break rigid deterministic workflows. Blue Yonder’s approach to ‘owned intelligence’ ensures the AI can absorb shocks in these complex scenarios without degrading service quality or violating compliance requirements.
By reducing reliance on external frontier models alone, Blue Yonder gains control over the AI’s interpretive and decision-making accuracy, which is crucial for mission-critical supply chain functions. Additionally, this shift signals a broader change in workforce roles, with AI managing routine tasks and humans focusing on higher-order customer and relational decisions.
What to watch next
Industry observers should monitor the adoption and efficacy of Blue Yonder’s hybrid AI models in real-world supply chains, particularly how they handle unstructured or novel disruptions compared to deterministic alternatives. Observing customer feedback on the consistency and reliability of AI-driven planning and operations will be key.
Additionally, the evolution of workforce dynamics around AI agents will be important. Blue Yonder foresees new roles emerging to oversee and manage AI agents, reflecting a shift in organizational operating models and human-AI collaboration that could redefine supply chain labor markets and skills requirements.