SBS’ answer is to bring the AI runtime to the data, rather than move sensitive records into an external model environment.
Why sovereign AI matters
The platform combines NVIDIA NIM microservices, NVIDIA NeMo Guardrails and open Nemotron models within SBS’ banking software estate.
It connects data from core banking, lending, digital banking and regulatory reporting, creating a governed intelligence layer for customer engagement, decision-making and risk monitoring.
SBS says Scottish Building Society is already achieving approximately 30% to 40% faster response times, with more consistent outputs and the explainability and audit trail regulators expect.
The deployment is significant because it’s the first live use of NVIDIA’s full AI stack inside a core banking system, the companies say.
- Data sovereignty: Banks can deploy AI on private cloud, managed cloud or on-premises infrastructure
- Regulatory control: NeMo Guardrails keep AI responses within each institution’s policy and compliance boundaries
- Operational intelligence: Natural-language tools convert complex structured and unstructured data queries into actionable insight
- Customer personalisation: AI-guided onboarding uses a bank’s own records and product portfolio to match customers with suitable products
- Risk and auditability: Risk monitoring surfaces compliance exposure while generating a complete regulatory audit trail
- Agentic roadmap: SBS plans to extend the platform to autonomous agents that execute multi-step banking workflows under human oversight.
SBS’ banking proposition
SBS positions itself as more than an AI vendor.
The Paris-headquartered company provides core banking, lending, payments, compliance, open banking and asset finance software to institutions across 80 countries, with its client base including Santander, Société Générale, BNP Paribas, Crédit Agricole, HSBC and Nationwide.
That installed base gives SBS a distribution advantage: it can embed AI where banks already hold customer, lending and regulatory data.
The company says its composable architecture is designed to modernise banking operations without requiring institutions to replace every underlying system at once.
“Banks set the standard for protecting customer data, and holding that standard shouldn’t mean having to wait on AI,” Eric Bierry, CEO of SBS, tells FinTech Magazine.
“Until now, most AI tools required that data to leave the bank’s environment. SBS already runs the core systems where that data lives, and working with NVIDIA, we’ve built a way to bring AI directly to it.
“Banks can now move forward on AI while keeping the control their customers expect.”
Eric’s argument reflects a wider industry gap.
Leave a Reply