Tackling your board's next big question
WTF are the Big Four up to with AI?
Jul 30 | 6 min read | By Tim Cooper
TLDR;
Large consultants are coupling up with AI models faster than contestants on a dating show. It’s their shortcut to relevance in the $1 trillion agentic AI market. But also necessary if they are to defend their business model from the 1,000 lb gorilla in the room… AI.
These partnerships promise wide-ranging benefits, including much faster, cheaper implementations. But they also bring new things for CFOs to think about, like conflict of interest risks and data governance issues.
The Big Four are going all in with AI model partnerships and using themselves as guinea pigs to test AI efficiencies and use cases.
The financial terms of these partnerships are largely a black box so far, leaving CFOs to guess at what it all means.
There are still looming questions around compliance, security, and conflicts of interest as the Big Four embrace AI model tie-ups.
CFOs don't adopt AI they can't explain to their auditors.
Campfire's built-in AI teammate shows its work. Every answer carries full source attribution - the transactions behind the number, visible and traceable. That's 95% accuracy, human-in-the-loop review, and a complete audit trail baked in.
It shows up every month end: pulling variance explanations, flagging anomalies, and drafting commentary automatically.
There’s been a spraycan of news about Big Four consulting firms partnering up with AI models. PwC is holding hands with OpenAI, while KPMG is going to the big dance with Anthropic. And EY is now going steady with Microsoft. But the business case behind these couplings isn’t quite clear yet. So, what gives?
The Big Four firms are trumpeting the benefits to clients in shiny press releases. But they’ve also been tight-lipped about the commercial terms of these agreements. And with so many partnerships being announced, it can be hard to keep up with who is doing what.
“CFOs are doing more research, experimenting with tools before they speak to consultants, and asking a lot more questions about the pros and cons of Big Four/AI model tie-ups,” said Geraldine Garaud, senior director analyst at Gartner.
Before we can get into the implications of these deals, we have to ask: WTF are they actually up to?
PwC and OpenAI go big into finance
PwC is taking one of the biggest swings through its partnership with OpenAI, aiming for an AI-driven transformation of the finance function.
PwC first linked up with OpenAI in May 2024. In that original deal, PwC became OpenAI’s largest enterprise customer, buying over 100,000 enterprise licenses for employees and rolling out an early version of ChatGPT Enterprise. PwC also became the first authorized reseller of OpenAI enterprise products as part of the 2024 deal. The partnership was touted as a first step in PwC’s plan to invest $1 billion into AI. But the firm stayed mum on what that deal cost.
In May of this year, PwC expanded its collaboration with OpenAI and narrowed its focus to building real-world AI finance agents to “reimagine the office of the CFO.” The companies announced that they’d be jointly building agents for procurement, forecasting, treasury, tax compliance, contract review, and the monthly close, starting first inside OpenAI’s finance organization. They also said that they are applying learnings from the internal rollout to additional agents across core finance workflows.
OpenAI CFO Sarah Friar said that "the finance function itself gets reimagined" around foresight and agility rather than efficiency alone. Financial terms were not disclosed in this deal, either.
KPMG dates around
KPMG is partnering with multiple AI models. On May 19, Anthropic gave KPMG the exclusive rights to a Claude-powered tax product, built directly into KPMG’s client delivery platform, Digital Gateway.
Anthropic also named KPMG as a preferred consultant for private equity. That means that when PE firms want AI built into the companies they own, Anthropic points them to KPMG instead of doing the work itself.
KPMG also launched a product called KPMG Blaze, which helps portfolio companies fix aging IT systems and quickly build AI products. But that part of the KPMG/Anthropic deal comes with a twist: around the same time, Anthropic partnered directly with Blackstone and Goldman Sachs on a similar venture to do much of the same work. In effect, Anthropic named KPMG an 'official partner' while simultaneously backing a competing operation that circumvents them entirely.
But KPMG hasn’t been entirely faithful, either. On July 21, the firm announced a separate deal with OpenAI that runs alongside its Anthropic partnership. Just like PwC, KPMG’s deal with OpenAI is testing new AI tools and systems, particularly around supply chain and government software, within OpenAI itself.
KPMG has not released commercial terms of any of its deals with the frontier AI models.
EY and Microsoft
Instead of going after the whole enterprise, EY is getting into the weeds with Microsoft. In May, EY and Microsoft announced a five-year deal that pairs EY consultants with Microsoft engineers in joint teams for “rapid, hands-on agent-building and deployment.”
Sidestepping company-wide AI rollouts, the teams are focusing on specific functions, including finance, tax, risk, HR, and supply chain. EY also set itself up as a use case for Microsoft’s Copilot, building agents directly into its own finance functions.
This deal did come with an announced $1 billion investment.
In March, EY rolled out a partnership with Chamath Palihapitiya's 8090 company, an AI software development platform, called EY.ai (PLDC). The new venture claims to accelerate both new software development and modernizing old software at scale through a “collaborative mesh” of AI agents on 8090’s platform. EY said it will deploy PDLC to tens of thousands of consultants firm-wide.
EY also joined forces with NVIDIA in December 2025 to launch EY.ai Labs, a project to develop “physical AI” that can control physical objects like drones and robots. Yes… You read that right.
Deloitte and Google Cloud
Deloitte has taken a less splashy route so far. In April, Deloitte expanded its deal with Google and created a new team inside the firm whose only job is to help clients use Google's AI tool, Gemini, to build and run AI agents. Rather than starting from scratch with every client, Deloitte built over 1,000 ready-made AI agents for specific industries that companies can plug in and customize, running on Deloitte's own delivery platform.
Neither company have said much about the terms of the deal.
The competitive landscape
For the AI companies, building on these collaborations is an aggressive land grab for clients and market share. Google forecast agentic AI will create a $1 trillion global market and they all want to cement their share.
The Big Four understandably want in. First and foremost, they want to stay ahead of what they see as a generational threat to their business models.
It also makes economic sense for the Big Four. Partnerships are revenue multipliers – for example, Google claims that for every dollar a customer spends on Google Cloud, partners capture up to $7.05 in services revenue. Big Four firms claim their expertise in areas such as deployment and change management should help organizations embed AI solutions faster and more efficiently to generate more value.
By linking with AI providers, they’re also likely seeking:
Margin share on licenses and co-developed solutions
Preferential pricing
Revenue from implementation, change management and other downstream services
Deeper internal expertise by rolling out tools across their workforces. This could help strengthen controls and deployment capabilities, and drive greater insights and adaptive operations for clients.
Leigh Coney is founder of WorkWise Solutions, which provides AI advisory and implementation in financial services. He said Big Four/AI deals now come up in every procurement conversation he’s in.
“An AI partner sitting behind the Big Four firm is a procurement variable CFOs didn't have to consider before. That triggers internal data governance and vendor risk reviews. Automation projects in EU companies are stalling because firms can't credibly answer whether sensitive data leaves the jurisdiction once it hits the model vendor's infrastructure,” he said.
Portability and what happens to a fine-tuned model at termination are also live questions, but most contracts are silent on these topics unless you query them. For example:
Which model touches what information?
Vendor agnostic clauses.
Portability and who owns workflows at engagement end?
“None of these were covered in service agreements a year ago. They are now,” said Coney.
Consultants are aware of these anxieties and have used recent announcements to try and reassure CFOs on AI trust and safety, and help address topics such as data governance.
One thing is for sure… engagement letters are going to get a whole lot more complicated.
Reading the room…
Answering your board’s next big question.
What do these partnerships unlock for us that we could not achieve as quickly, cheaply, or effectively alone?
Where could early adoption create real competitive advantage, rather than simply adding another expensive layer to our technology stack?
How is the adviser paid, and could these commercial relationships compromise its independence, including across existing audit or advisory work?
What does the engagement letter say about the ownership of agents, workflows, code, data, outputs, and other work products?
Can we move the solution to another model or adviser, or are we creating long-term lock-in?
What data reaches the AI provider, where is it processed, and how is confidential information protected?
What measurable business outcomes justify the cost, risk, and management attention required to make this work?
Boardroom Brief is presented by The Secret CFO Network
Last week’s Playbook was the grand finale of our series on building FP&A. This time, we look at the (uncertain?) future of FP&A.
If you found this helpful, please forward it to your fellow finance leaders (and maybe even your Board). If this was forwarded to you, make sure you receive the next edition by subscribing here.
