Welcome to The Wersel Labs Briefing. This is the inaugural issue — short, candid, and intended to set out what you can expect from this dispatch and from the firm behind it.
Why we founded Wersel Labs
We are senior practitioners from banking, analytics, data engineering and applied AI who have spent careers building the models, pipelines and governance artefacts that regulators inspect. We founded Wersel Labs in February 2026 because the next decade of regulated analytics is not going to be served by the legacy consulting playbook — pyramids of junior consultants, billable hours scaled against headcount, and a thin layer of AI bolted onto an old delivery model.
What it needs is AI-native engineering applied by senior practitioners to defensible quantitative craft. That is what we are building.
What we believe
- AI is a model-risk problem, not a chatbot problem. Inside a regulated bank or insurer, every prompt template, retrieval pipeline and agent workflow is inventory-eligible, governance-eligible and validation-eligible from day one. The institutions that recognised this in 2024 are shipping at meaningful volume now; the ones that deferred are retro-fitting under deadline.
- Lineage is captured, not mapped. A BCBS 239 programme that hand-maps the lineage of every critical risk report produces a snapshot that decays inside eighteen months. Sustainable lineage is engineered into the data platform.
- Defensibility is engineered, not narrated. A model that survives supervisor inspection at first review is one where the documentation, validation evidence and ongoing monitoring were designed in from the outset — not assembled into a binder before the meeting.
- Senior practitioners on day one. The person the client meets at pitch is the person who ships the work. We are not building a pyramid.
What to expect from this dispatch
The Wersel Labs Briefing is a quarterly synthesis of what is shifting in regulation, modelling standards and applied AI for senior decision-makers in financial services. We aim for short, candid and useful. No vendor talking points, no recycled commentary, no thought-leadership-for-its-own-sake.
If you have made it this far — we are glad you are with us, and we look forward to the conversation.