Why your API infrastructure isn’t ready for AI (and what to do about it)

Financial institutions have developed extensive API infrastructures but often lack a structured context for effective AI integration, as highlighted by Mark Boyd at LEAP 2026. The challenge lies in understanding user needs, market dynamics, and organizational capabilities, rather than just the technology itself. Boyd advocates for a context-first approach, where intent is clearly documented to prevent issues like API drift and sprawl, enabling AI systems to function reliably. This method is exemplified by companies like ABN AMRO and Klarna, who successfully implemented agentic AI by focusing on user context and interoperability. Boyd suggests starting with low-risk, high-value AI use cases, particularly in financial operations, and emphasizes the importance of mapping business capabilities and regulatory environments. By aligning infrastructure decisions with business goals through a framework that includes FinOps modeling, organizations can effectively manage costs and achieve cross-functional buy-in.