Data & AI skills for financial services

Financial services has had data at its core for decades. Risk models, pricing, capital requirements: the quantitative side is mature, often more mature than the sector gets credit for. That is exactly why capability building here looks different from anywhere else.

The gap is rarely in the data team. It sits with everyone around it. The mortgage advisor reading a risk score, the claims handler working next to an AI assistant, the product owner deciding whether an AI use case is worth building at all. They do not need to build models. They need to understand enough to challenge what they are shown, and to know where the line runs.

And that line is real. Model risk management, supervision by DNB and the ECB, and the EU AI Act's high-risk classification for creditworthiness assessment all mean the same thing: a programme that treats governance as an afterthought is a programme your second line will stop. We build it in from the first cohort instead of bolting a compliance module onto the end.

We start from what has to change in the work, not from a catalogue. The behaviour we are after in this sector is specific: a colleague who reads a model output, notices it does not match the file in front of them, and says so — with a reason the second line accepts. That is a skill, and it is teachable.

Getting there needs three things at once. Knowledge of what a model can and cannot carry, and where the regulator draws the line. Skill in the tools your people actually work in, practised in the browser on your own reports and cases, because hands-on practice is roughly six times more effective than watching video (Carnegie Mellon). And the habit of asking for the number before the opinion, which only forms if people practise it on real work rather than a case study.

Adoption is the part most programmes skip. Cohorts run with a kick-off, live sessions and Data Champions from your own teams, and we hand ownership to your L&D and community at the end so it keeps running without us. Growth is measured from a baseline taken before the start and benchmarked against other organisations, so you can show a change in capability rather than a completion list.

We work with financial institutions including Rabobank, ING, PGGM and Allianz Partners.

Across all our clients, cohorts finish at 79% completion over the last twelve months and more than 10,000 learners, against an industry standard of roughly 12%.

Our co-founder Wouter Neef worked at ING before starting Data Booster, so the constraints on this page are not theoretical to us: model risk management, a second line that has to sign off, and colleagues who need to challenge a score without ever building one.