Skip to content

PwC, OpenAI Prep AI Treasury Agents

Chimera readability score 72 out of 100, Expert reading level.

After years of skepticism, agentic AI is reshaping how CFOs run their organizations.
Working in conjunction, global accountancy and advisory firm PwC and OpenAI are bringing agentic AI to CFOs and their organizations. They promise that their agents can deliver benefits to the planning, forecasting, reporting, procurement, payments, treasury, and tax functions of financial organizations.
The technology is no longer seen as emerging—it is now widely accepted as an essential tool for optimizing operations and driving long-term growth.
As recently as October 2025, AI remained controversial. Deloitte in Australia faced a reported $290,000 judgment after it submitted a report to Australia’s Department of Employment and Workplace Relations that included a range of generative AI hallucinations, prompting litigation. Such incidents made accountants wary of the technology and its shortcomings.
Nevertheless, appreciation for AI input has rapidly evolved, with a little help from human touch. PwC and OpenAI have clearly defined roles: AI agents execute and coordinate work, while PwC employees supervise—a structure designed to reduce the risk of hallucinations.
Proposal Relies on Real-World Experiences
OpenAI is presented as “customer zero.” The company uses its ChatGPT AI chatbot and Codex software coding agent in its own financial organization, where they “monitor payments, review contracts, update forecasts, and prepare reporting materials,” according to a prepared statement. Meanwhile, PwC implements that know-how in other companies. The lessons learned at OpenAI will help other CFOs.
Some of the complex corporate workflows that AI agents have managed, according to OpenAI officials, include processing five times more contracts without adding professionals to the existing team, and managing more than 200 investor interactions during a fundraising event.
PwC and OpenAI appear to have mastered the path to deploying agentic workflows.
Nevertheless, in this rapidly evolving new world, PwC doesn’t work exclusively with OpenAI. The firm recently announced another collaboration with OpenAI rival Anthropic. PwC is offering its large client portfolio access to Anthropic’s Claude AI assistant. Financial services, pharmaceuticals, and life sciences clients are particularly interested in Claude’s efficiencies, according to PwC. In the insurance sector, underwriting cycles could be reduced from weeks to days. In cybersecurity, agents respond to threats in minutes rather than hours. The reimagining of the CFO’s office is just beginning.

Sentinel — Human

Confidence

This text is a well-contextualized report that successfully bridges cutting-edge AI theory with specific industry risks and professional service partnerships.

Signals Detected
low severity: Human text variance in sentence structure and pacing is present; the flow shifts between reporting facts and forward-looking commentary.
low severity: The text maintains a coherent, slightly cautious, yet optimistic tone consistent with financial industry reporting. The framing balances technological excitement with documented risk.
low severity: Specific data points (e.g., $290k judgment, 5x contracts) are presented alongside qualitative partnership information, suggesting editorial curation rather than raw LLM generation.
low severity: The specific naming of partnerships (PwC/OpenAI/Anthropic) and the historical reference to the Deloitte judgment anchor the narrative in verifiable, specific events, increasing credibility.
Human Indicators
The inclusion of a specific, dated legal incident ($290,000 judgment) acts as an idiosyncratic point of reference that anchors the generalized AI discussion in concrete reality.
The transition between discussing high-level agent capabilities and specific client sector applications (insurance, cybersecurity) shows contextual depth typical of human industry analysis.
The nuanced discussion regarding the split of roles (AI executes, PwC supervises) reflects an understanding of organizational risk management philosophy rather than pure technological enthusiasm.