We ' ve heard a lot over Inspire about the sheer amount of data – global spend data – that Coupa has. How do you ensure a platform with that much data stays fast and reliable as the network continues to grow? We deal with about 7,000 requests per second in terms of our APIs. We handle about 1.5 million approvals a day and about three terabytes of data per day in terms of our data analytics. With such a humongous amount of data and approvals that we handle, we have to ensure that what we do scales. Twenty years ago, our founders had this thought: we need to enable everything in the cloud. And it started with that cloudfirst approach – multi-tenant, multi-region and multi-data centre. Because of that, we are able to horizontally scale the application layer, the data layer and now the agentic layer.

How is Coupa actually building AI into the core of its platform so it helps people work faster and achieve better results? A lot of companies bring AI in on top and bolt it in, then call themselves an AI company. In our case, we thought completely differently. We went through a few iterations, but the core of that was: we want to build AI from within the platform. There ' s a difference between an AI-powered company and an AI-first company. Our vision is to be the AI-first company.

For a global enterprise thinking of introducing AI to their workflows, trust would be everything. Howdoes your team of engineers create a platform with the security and resilience that makes customers feel safe when they make that transition to autonomous operation? For us, the three key tenets of trust are control, security and transparency. With control, what we mean is the admin has absolute control of what data and what access they can grant to each and every agent or API – whether within their own company or externally to partners. With security, everything is protected by what we call RBAC, role-based access control. Data is secured at rest and in transit, so at no point does anyone have access to the data other than our customers alone. And finally, transparency. Every AI decision is logged, so there is very clear auditability and traceability. Because of that, our customers know exactly why certain actions are taken, what resulted in a particular outcome and they can go back and audit every step of the way.

84 July 2026