Built from Real Financial Challenges
Stellix Max started when a group of financial analysts got tired of spending weeks on risk assessments that still missed critical patterns. We'd sit through endless meetings, crunching numbers manually, only to find out later that we'd overlooked something obvious in hindsight.
The turning point came during a project in 2019 where manual analysis completely failed to predict a client's exposure. That's when we decided to build something different — tools that could see patterns humans naturally miss, but still explained decisions in ways that made sense.
We're based in Toulouse, working with businesses across France and beyond. Our focus isn't on replacing financial expertise. It's about giving professionals better tools to do what they already do well, just faster and with more confidence in the results.
How We Actually Work
Most risk assessment tools either oversimplify everything into red-yellow-green indicators, or they're so complex that you need a PhD to interpret the results. Neither approach helps someone who needs to make a decision by Friday.
Our approach sits somewhere in between. We train models on historical financial data — not just numbers, but the stories behind those numbers. Market conditions, industry shifts, regulatory changes. Then we show you not just what the risk level is, but why the system thinks that way.
The goal is transparency. You should be able to challenge the assessment, add context the system doesn't know, and ultimately trust your own judgment with better information backing it up.
Pattern Recognition
We analyze thousands of financial scenarios to identify risk indicators that aren't obvious on surface-level review. Things like unusual transaction timing, correlation shifts, or subtle changes in cash flow patterns that often precede bigger issues.
Contextual Analysis
Raw numbers don't tell the whole story. Our systems factor in industry benchmarks, seasonal variations, market conditions, and business lifecycle stages. A cash flow dip means something different for a retail business in January versus July.
Liselotte Vandenbroeck
Chief Risk Analytics Officer
Liselotte joined us after spending over a decade building credit risk models for European financial institutions. She's the one who insists that every algorithm we deploy needs to pass the "explain it to a board meeting" test — if we can't justify why the system flagged something, it doesn't go into production.
Adaptive Learning
Our models improve as they process more data from your specific business context. What works for manufacturing doesn't work for professional services, so the system adjusts its baseline assumptions over time.
Multi-Source Integration
We pull data from accounting systems, bank feeds, invoicing platforms, and market indicators. The more complete the picture, the more reliable the assessment becomes.
Explainable Results
Every risk score comes with clear reasoning. You'll see which factors contributed most to the assessment and how changing specific variables would affect the outcome.
