OpenAI and PwC just announced a partnership to build AI agents specifically for CFOs. Not copilots. Not chatbots. Autonomous agents that execute entire finance workflows from start to finish — procurement, contract review, month-end close, treasury, tax, and reporting. This is agentic AI for finance, and it's already running inside OpenAI's own finance organization.
Here's why this matters for technical and business leaders.
What OpenAI and PwC Are Building
The partnership targets what PwC calls "practical, high-value workflows where AI agents can execute and coordinate work under human supervision." Translation: agents that don't just assist — they act.
Specific workflows include:
- Procurement agents that handle vendor selection, contract negotiation, and purchase order processing
- Contract review agents that accelerate legal review and flag risks
- Close agents that speed month-end and quarter-end activities
- Risk assessment agents that identify financial exposures
- Reporting agents that generate customized dashboards and financial reports
- Treasury agents that manage cash flow, payments, and liquidity
OpenAI's finance organization is serving as "customer zero" — testing these agents internally before rolling them out to other CFOs. According to OpenAI, they're already using AI agents across investor relations, treasury, tax, reporting, corporate development, and contract review workflows.
PwC's U.S. advisory leader Tyson Cornell framed it this way: "Finance is at an inflection point, where organizations are moving from process efficiency to intelligent, decision-centric operations. 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."
Why This Is Different: Agentic AI vs. Copilots
Most CFOs are familiar with AI copilots — assistants that respond to prompts, summarize information, or suggest next steps. Agentic AI is fundamentally different.
According to the CFA Institute, agentic AI systems "run entire financial processes autonomously, end to end." Instead of waiting for a human to ask a question, agents operate with intent. They monitor transactions in real time, validate controls, flag anomalies, and suggest remediations — all without needing a month-end trigger or manual review.
The key difference:
- Copilots assist: You ask, they answer. You decide, they execute.
- Agents act: They understand outcomes, execute multi-step processes, and hand off edge cases to humans only when necessary.
Gartner projects that by 2030, more than 80% of finance functions will embed AI-driven autonomy in core processes. The shift from "assisting" to "acting" is already underway.
The Technical Implementation: What CIOs Need to Know
From an IT perspective, this isn't just another SaaS integration. OpenAI and PwC are building what they call an "AI-native finance function" — a system where agents are embedded directly into core finance workflows, not bolted on as external tools.
Key technical considerations:
-
Enterprise-scale workflows: OpenAI is testing governance models, runtime controls, and human-agent collaboration patterns at scale. This isn't a proof of concept — it's production-grade.
-
Token consumption visibility: PwC explicitly states that CFOs will need visibility into AI usage, token consumption, and projected spend. This means real-time cost monitoring is built into the architecture.
-
Compliance and audit trails: In regulated finance environments, explainability is non-negotiable. Agents need to produce audit-ready logs and justify every decision.
-
Integration with existing ERP systems: These agents won't replace your Workday, Oracle, or SAP stack. They'll integrate with it. PwC is building reference architectures, APIs, and compliance frameworks that other finance platforms can adopt.
-
Human-in-the-loop design: Agents execute autonomously, but edge cases and final approvals still route to humans. The goal is augmentation, not full automation.
The Competitive Landscape: Big Four Firms Are All In
OpenAI and PwC aren't alone. The Big Four consulting firms are racing to deploy agentic AI in finance:
- KPMG + Google Cloud: Launched an AI assistant for month-end close in April 2026, powered by Gemini Enterprise and integrated with Workday.
- Deloitte + Google Cloud: Announced an "agentic transformation practice" focused on finance workflows.
- Deloitte + HPE: Testing agentic AI in finance workflows at Hewlett Packard Enterprise.
This is a land grab. Consulting firms see agentic AI as the next multi-billion-dollar finance transformation wave, and they're moving fast to lock in clients.
What CFOs Should Do Now
If you're a CFO, here's the playbook:
1. Identify high-value, repeatable workflows
Start with processes that are:
- Time-consuming but rules-based (e.g., invoice matching, contract review)
- High-volume and predictable (e.g., expense approvals, vendor payments)
- Audit-sensitive but formulaic (e.g., reconciliation, compliance checks)
2. Benchmark your current costs
Before deploying agents, you need baseline metrics:
- How many hours does your team spend on month-end close?
- What's the cost per invoice processed?
- How long does contract review take on average?
Without baselines, you can't measure ROI.
3. Demand token-level visibility
If you're paying per API call or per token, you need real-time cost monitoring. Ask vendors:
- What's the cost per workflow execution?
- How do token costs scale with transaction volume?
- What's the projected annual spend at our scale?
4. Start with a pilot, not a full rollout
OpenAI's own finance team is testing these agents internally before deploying to customers. Follow their lead. Pick one workflow (e.g., contract review), run a 90-day pilot, measure results, then scale.
5. Prepare for board questions
According to recent surveys, 97% of CFOs say their boards expect regular readouts on AI investment and progress, with a focus on cost savings, ROI, and productivity gains. Document everything from day one.
The Bottom Line
Agentic AI for finance is no longer theoretical. OpenAI is running it in production. PwC is packaging it for enterprise deployment. KPMG and Deloitte are building competing offerings.
The strategic question for CFOs and CIOs isn't "Should we deploy AI agents?" — it's "Which workflows do we automate first, and which vendor do we trust to execute?"
Finance is moving from process efficiency to intelligent, decision-centric operations. The CFOs who move early will gain a measurable cost and speed advantage. The CFOs who wait will spend 2027 explaining to their boards why competitors are closing books faster and operating leaner.
The race is on.
Continue Reading
- AI Metrics for CFOs: The Four-Layer Framework Beyond Tokens and Seats — How to measure AI ROI when seat-based pricing doesn't apply
- Gartner: By 2030, 80% of Finance Functions Will Embed AI-Driven Autonomy — What the next generation of CFO technologies looks like
- KPMG Launches AI Assistant for Month-End Close with Workday and Google Cloud — How the Big Four are racing to deploy agentic AI
About the Author: Rajesh Beri is Head of AI Engineering at a Fortune 500 security company and publishes THE DAILY BRIEF, a newsletter for enterprise AI leaders. Follow him on LinkedIn and Twitter/X.
