AI Portfolio Management at a Food Corporation

Starting point

A family-owned German food corporation with around 2,500 employees and several specialized subsidiaries wanted to systematically embed AI into its value chain. Initial in-house developments had failed, the data landscape was fragmented across dozens of source systems, and there was no operating model to manage AI initiatives in a structured way. The executive board set clear guardrails: no parallel organization, no major investments, leverage existing teams. Maximum ambition with minimal resources.

What we did

Over 18 months, we served as an embedded AI strategy advisor at executive level, building an AI portfolio of 49 use cases — from identification through prioritization to implementation. In parallel: GenAI rollout with 450 accounts, establishment of a decentralized AI coach network to enable business units, technology scouting and platform evaluation, and a portfolio of small-scale automations that delivered immediate impact. Scope: strategy, portfolio management, enablement, implementation support. Team: 1 external advisor, 1 internal counterpart.

Results

What we learned

The real bottleneck was never the technology. It lay in data provisioning — specifically in the missing link between product data management and the AI systems. We see this pattern across industries: companies systematically underestimate how much of AI's value creation depends on the data layer, not the model.