Financial institutions are often the earliest adopters of cutting-edge technology. When milliseconds mean millions, technological innovation is essential to continue to generate value. At the same time, Finance is the transactional backbone of every industry on the planet, and it runs on troves of personally identifiable information in a massively complex regulatory environment, with real money at stake when things go wrong.
A regional bank recently told us that the table stakes for building anything new were that "everything works–fast, consistently, and all the time." That need to balance the tension of innovation and safety is why financial institutions like JPMC are becoming the new thought leaders in AI security, and a reason we've had so many successful partnerships in the industry.
Once baseline service and security guarantees are met, finance is ultimately a relationship business. The biggest differentiator often isn't an application feature or even the services offered, but the customization, packaging, and delivery of those services in a way that leverages customer data without compromising security.
A mandate for innovation, and absolute need for accuracy, and decades of private data that needs to be incorporated with extreme caution is the perfect storm for RAG, and it's why the financial industry is leading the way. Whether it's creating efficiencies in customer support organizations, launching personalized services grounded in real-world customer behavior, or improving fraud identification, RAG is the layer of the stack that keeps LLMs honest while keeping data secure and compliant.
See us in person at the Open Source in Finance Forum (OSFF)
We recently highlighted how ReBAC (and specifically, SpiceDB) is the easiest path to a governable Agentic RAG system. Next week, at OSFF London, you can see it in action when Evan Corkrean presents "Fine-Grained Authorization for RAG Embeddings Retrieval With Relationship-Based Access Control." It's happening Thursday, June 25, 2026, from 12:00–12:15 BST in the Churchill Room on the ground floor. It's a digestible, concentrated discussion of the architectural decisions necessary to get RAG running properly, led by someone with implementation experience spanning multiple Financial Services RAG companies, as well as RAG implementations at Zoom and OpenAI.
Add the session to your OSFF London schedule, get your questions ready for Evan, and if you'd like to prototype before the event, try the AuthZed Cloud Starter Program for $700 in credits to get you going!
Add the Session to Your Schedule