All case studies

Food & Ingredients · Food Corporation

Two AI Platforms, One Decision

Starting point

An internationally positioned food corporation with several thousand employees and over a dozen operating companies faced a classic knowledge problem: R&D findings, formulations, and process documentation were scattered across SharePoint, local drives, and the heads of individual experts. New employees needed months to get oriented. Technical questions were resolved via email or hallway conversations — or not at all. The executive board wanted to know whether an AI-powered knowledge platform could solve the problem — and if so, which one.

What we did

makematiq led the project as interim management: product ownership and technical steering sat with us, while implementation was handled by a specialized service provider with roughly six developers. In Phase 1, we ran a 30-day pilot with an enterprise search solution, identified 29 use cases, and prepared a decision brief. When a second platform provider entered the European market with an integrated agent-based approach, we set up a full comparison test in Phase 2: 186 data sets across 4 use cases, systematically benchmarked. Result: executive approval, rollout concept, production deployment.

Results

186

data sets compared across 4 use cases

78 %

of responses at or above human level

29

identified use cases

12 Monate

from evaluation to production

What we learned

The technology selection was not the hardest decision. The hardest was to run the comparison test when the market shifted mid-project. Anyone evaluating in a rapidly moving technology market must be willing to reopen a decision already made — it costs time but prevents strategic mistakes.

This is the summary. How we approached it methodologically — which architectural decisions we made, what we discarded and which patterns can be transferred to other contexts — we discuss in a personal conversation.

Not because we want to sell you something. But because this depth is what our clients engage us for — and it does not belong on the open internet.