Most AI investments don't fail in implementation. They fail earlier — in the assumptions nobody checked before the budget was committed.

I run due diligence on AI investments for Mittelstand companies, based in Cologne, Germany. Before you sign with a vendor, build on a model, or restructure a process around AI, I test what you're actually betting on: the vendor's claims, the compliance exposure, the assumption that this problem needed AI at all.

One example. The EU AI Act sorts systems into risk tiers, and the obligations land on you as the deployer - not only on the vendor. Most vendors don't mention their risk classification in the sales process. Some don't know it. You find out at audit time. Or earlier, if someone asks first.

Asking first is my job.

What I bring: 30 years across financial services, risk management, and technical architecture — the last 16 running my own independent practice, client-funded since 2010. Certified Financial Risk Manager (GARP), quantitative finance background (CQF), and three decades of watching confident projects answer the wrong question expensively.

What I don't bring: an AI product to sell you, a partnership with any vendor, or an interest in telling you what you want to hear.

Engagements:

• Second Opinion — written review of a decision on your desk, 48–72h (from €2,500)
• AI Investment Assessment — 5–7 day deep check of a planned investment (€3,000–€5,000)
• Due Diligence Mandate — 8-week engagement for major commitments (from €15,000)

If you want confirmation, hire a cheerleader. If you want to know what you're missing while it's still cheap: andredaus.com.

Founder — https://andredaus.com

Technical consulting & WordPress infrastructure — LittleBig.Co (https://littlebig.co)

Co-Host, The Human Perimeter (cybersecurity podcast) — https://human-perimeter.com

Host, Iconoclast Insights (solo podcast)