The collaboration, announced Tuesday (May 5), will see the company build artificial intelligence (AI) agents centered on the “core operating rhythms of finance,” including things like forecasting, planning, reporting, procurement, payments and treasury.

“Finance has always been about judgment, trust, and making decisions in environments filled with complexity and constant change. AI gives finance leaders a much deeper ability to see around corners and act faster,” Sarah Friar, OpenAI’s finance chief, said in a news release.

“I believe we’re now entering a moment where the finance function itself gets reimagined to shape decisions in real time. The opportunity here is far bigger than efficiency, it’s about giving finance leaders the tools to operate with greater foresight, agility, and strategic impact across the business.”

The companies say the collaboration is focused on real-world uses rather than “designing in theory.” For example, the two firms are building a procurement agent inside the OpenAI finance organization, and applying those learnings to other agents.

“Finance is at an inflection point, where organizations are moving from process efficiency to intelligent, decision-centric operations,” said Tyson Cornell, PwC’s U.S. advisory leader. “Through our collaboration with OpenAI, we’re helping clients embed agentic AI into the core fabric of the finance function, enabling more proactive insights, stronger controls, and a more adaptive operating model.”

Under this new model, the release added, the role of finance professionals is changing from carrying out processes to overseeing and improving AI agents. While still accountable for judgment, controls and outcomes, finance teams also help define the standards that let agents perform more effectively and responsibly, the companies said.

The partnership comes as the business world continues to adopt agentic AI. But as PYMNTS wrote last month, there are places where this adoption can only go so far, such as in accounts receivable (AR) and accounts payable (AP) functions.

“As AI agents become more capable, the organizations that can feed them high-quality, structured data could be better positioned to gain a decisive advantage over those that cling to document-centric and human-optimized processes,” that report said.

“The problem is that most B2B contracts and AP/AR workflows were never designed for an agentic reality. These fundamental operational artifacts are frequently dense, ambiguous and optimized for human judgment rather than machine execution.”