OpenAI and PwC Build Finance Agents for Forecasting, Payments, and Treasury
OpenAI and PwC announced a partnership Tuesday to develop AI agents for finance operations, focusing on forecasting, planning, reporting, procurement, payments, and treasury functions.
The collaboration centers on real-world applications rather than theoretical models. OpenAI and PwC are already building a procurement agent within OpenAI's own finance organization and plan to apply those lessons to other finance processes.
Finance Teams Shift From Process Work to Agent Oversight
Under this model, finance professionals move from executing processes to overseeing and improving AI agents. They remain accountable for judgment, controls, and outcomes while helping define standards that let agents perform more effectively.
Sarah Friar, OpenAI's finance chief, said the shift enables finance leaders to "see around corners and act faster." She framed the opportunity as extending beyond efficiency gains to giving finance teams tools for "greater foresight, agility, and strategic impact."
Tyson Cornell, PwC's U.S. advisory leader, said organizations are moving "from process efficiency to intelligent, decision-centric operations." The partnership aims to embed agentic AI into the core of the finance function, enabling proactive insights and stronger controls.
Data Quality Remains a Barrier
The partnership comes as businesses adopt AI for finance more broadly. However, adoption faces real constraints in accounts receivable and accounts payable functions.
Most B2B contracts and AP/AR workflows were designed for human judgment, not machine execution. These documents are often dense and ambiguous, making it difficult for agents to extract structured data reliably.
Organizations that can feed AI agents high-quality, structured data will have an advantage over those relying on document-centric processes.
Your membership also unlocks:
