Coupa’s push toward fully autonomous spend management took a significant step at its Inspire 2026 conference, where the company revealed it is buying AI document specialist Rossum. The deal underscores intensifying competition to automate back-office finance and procurement workflows with domain-specific AI.

The relationship between the two firms began as a partnership in 2024, when Rossum’s intelligent document processing (IDP) was first embedded in Coupa’s platform for accounts payable tasks.¹ Over the next two years, customers used Rossum’s tools to handle complex invoices, setting the stage for a deeper integration.

At Coupa Inspire 2026 in Las Vegas, the company announced it would acquire Rossum outright, aiming to extend these AI-powered IDP capabilities “across Coupa's product portfolio.”² Rossum’s technology is built on a transactional large language model (T‑LLM) trained on tens of millions of documents, designed to move beyond traditional optical character recognition and continuously learn from each customer’s document set.²

From Coupa’s perspective, the acquisition is about accelerating the shift from process automation to decision automation in spend management. CEO Leagh Turner said “Rossum changes the game entirely,” arguing that combining Rossum’s AI-first technology with Coupa’s data could help customers save “the next US$300bn in five [years] with a system of decision and intelligence that is unrivalled.”²

Industry observers place the move within a broader wave of AI-driven enterprise deals. AIMA Magazine listed Coupa’s purchase of Rossum alongside other major AI developments, including OpenAI’s reworked Microsoft agreement and SAP’s controversial API policy, as one of the week’s top AI stories, highlighting how core business systems are being reshaped by generative and transactional AI.¹

By integrating Rossum across direct and indirect spend scenarios, Coupa and Rossum pitch the deal as a way for enterprises to gain tighter control over data in source-to-pay operations, potentially reducing costs and speeding complex invoice processing.²