The Challenge
QVC Group's Core Retail transformation required connecting transactional retail platforms - SAP S/4HANA, IBM OMS, and STIBO PIM - with advanced AI capabilities. The organization needed a cohesive architecture that could support large-scale ML, generative AI, knowledge graphs, and autonomous agent frameworks while ensuring end-to-end interoperability across data, processes, and decision systems.
My Approach
I defined and continue to evolve the enterprise AI and Agentic Architecture roadmap, ensuring that AI capabilities are not bolted on but deeply integrated into the retail technology landscape.
Architecture Integration
The roadmap addresses integration across three core enterprise platforms:
- SAP S/4HANA: ERP backbone for finance, supply chain, and operations
- IBM OMS: Order management and fulfillment orchestration
- STIBO PIM: Product information management and data governance
Each integration point is designed for interoperability - AI systems can consume and enrich data across these platforms without creating silos.
Agentic Framework
The architecture incorporates autonomous agent frameworks for retail-specific use cases, enabling AI systems that can reason, plan, and execute multi-step workflows with appropriate human oversight.
Key Decisions & Trade-offs
Interoperability first: Rather than building isolated AI capabilities, I prioritized end-to-end data flow across all enterprise systems. This added architectural complexity but ensures AI capabilities compound rather than fragment.
Agentic + traditional ML: The roadmap balances autonomous agent capabilities with proven ML approaches (forecasting, recommendations, classification), choosing the right paradigm for each use case.
Impact
The architecture provides a unified foundation for AI-driven personalization, operational optimization, and enterprise automation across QVC Group's Core Retail operations.