Data Engineering Meets Excel: Building Explainable and Reliable Decision Models with River Solutions

In this episode of The Data Playbook, Kris Peeters talks with Amaury Anciaux, founder of River Solutions, about turning messy, error-prone Excel files into explainable and reliable decision models.

They dive into why Excel will never die, how River Solutions augments analysts instead of replacing them, and what happens when you bring engineering best practices (tests, data quality checks, documentation) into the world’s most abused business tool.

You’ll learn:
Why 99% of decision models are still built in Excel
How to make Excel models explainable to stakeholders
How River replaces fragile formulas with visual, flow-based modelling
Built-in data quality checks to kill “silent errors”
How AI copilots (like GitHub Copilot) accelerate both engineers and analysts
When to stay in Excel vs. when to move to production data platforms

🎧 Listen to more episodes of The Data Playbook for real-world stories on data platforms, GenAI, data products and cloud independence from Europe’s leading data practitioners and leaders.

🌐 More at https://www.dataminded.com/resources and subscribe to our channel:    / @dataminded  

Chapters:
00:00 – Intro & episode setup
00:45 – Amaury’s background & consulting career
02:00 – The hidden reality of Excel decision models
04:00 – Why “just get it out of Excel” doesn’t scale
05:10 – What River Solutions does in Excel
06:40 – Visual model maps for explainable models
08:40 – Removing formulas & adding data quality checks
10:50 – Why Excel errors are so risky for big decisions
13:15 – Who River is for: analysts, Excel gurus & managers
16:05 – Why Amaury started River now & building with Copilot
19:00 – Will AI copilots replace River and Excel modelling?
22:51 – How River works as an Excel add-in (UX & interactivity)
26:25 – How River changes the analyst role (less debugging, more thinking)
28:10 – Roadmap: community, cloud, AI & connecting to data warehouses
31:14 – Biggest lesson learned: software is easy, change is hard