KPMG LLP, the U.S. member firm of the KPMG International network, is folding its artificial intelligence, innovation, and ecosystem functions into a single new unit designed to spin up AI-native businesses at start-up speed — and installing a vice chair who reports directly to the firm's chief executive to run it. The structural move amounts to a bet that the biggest disruption to the professional-services model in a century cannot be managed through the usual committee layers. KPMG is one of the Big Four accounting firms — alongside Deloitte, EY, and PwC — whose core product for more than a century has been expert human judgment, exactly the kind of knowledge work generative AI is increasingly able to perform.

The U.S. audit, tax, and advisory firm's new Client Technology & Innovation (CT&I) group, to be announced Sept. 22 and effective Oct. 1, will be led by Todd Lohr, a 15-year KPMG veteran who told Fortune he had been "pushing for something like this for several years" — specifically, a structure with "a singular leader that reports directly to the CEO, which I do, because that is how we're going to drive the transformative change that is required at KPMG."

CT&I's launch coincides with the departure of Steve Chase, vice chair and global head of AI & Digital Innovation, who is to retire from the firm on Sept. 30 — one day before the new unit takes effect. Chase built KPMG's original AI and Digital Innovation group starting in 2023 and more recently held a KPMG International mandate as Global Head of AI, a role he moved into in August 2025.

The reporting line elevates AI and innovation strategy out of the layer of vice chairs and functional heads and puts it in the same room as CEO Tim Walsh's own agenda — a signal that KPMG views this less as a technology upgrade than as a net-new group focused on building what the firm will sell in the future.

"There are things in the future that are, you know, a little existential for any organization, including our own," Lohr said, "but I certainly think that we are sitting at this kind of pivotal moment in history, where we're going to be able to take advantage of some of these swings in the market, and be able to look very differently at the future of KPMG, which is super exciting."

Why now, and why Lohr

Lohr, 49, is not an outside hire brought in to inject Silicon Valley DNA. An internal operator who has spent years building the exact relationships CT&I is designed to commercialize, he helped start the firm's early data and AI practices, ran its U.S. technology business including cyber, then moved into a role overseeing third-party technology partnerships before a stint leading the firm's clients-and-markets commercial organization.

Lohr was one of the leaders involved with architecting KPMG's alliance with Anthropic. Along with the Anthropic tie-up, existing relationships with Google Cloud, Microsoft, OpenAI, and Databricks all now roll into CT&I's ecosystem mandate.

His path to KPMG is a bit unusual for Big Four leadership. Before joining the firm, Lohr was employee No. 3 at Confiance, a financial-services consulting start-up he co-founded in 2008 — "not the best year to do a start-up," he said — and worked as a business architect at Zurich Insurance. That start-up scar tissue, he suggested, is precisely what qualifies him to run what he described as an internal venture studio: "I'm taking my design cues from Silicon Valley incubators."

What CT&I actually does

Lohr's new organization has four pillars that go well beyond a typical "AI office":

• Products and platforms strategy — owning the products KPMG delivers to clients, not just internal tooling.
• Commercial infrastructure — new go-to-market and deal structures "in very different commercial terms than what we have historically done."
• Firmwide AI and data strategy — the strategic mandate previously housed under the AI & Digital Innovation group.
• Innovation/incubation — an internal venture studio incubating a portfolio of "edge disruption plays" with varied capital structures and routes to market, explicitly aimed at disrupting KPMG's own core service lines.

Each "edge company" carries a mandate to build and scale with multiple exit paths — including being folded back into KPMG's core business, spun out with outside capital, or run as a standalone joint venture with a technology partner.

Colleagues have floated even bigger stakes. Lohr called the group an incubator modeled on Silicon Valley design and said he hopes a billion-dollar company comes to fruition as a result. "I think KPMG and professional services in general have some of the greatest — we are a talent factory for most of the world on a lot of business skills," Lohr said. "We have a lot of ideas. What we have not done is … bring those ideas through commercial lens — we are not as good at that, because we know how to do the model that we've done for over a century."

That talent question is also why Lohr said KPMG has spent the past several years deliberately educating its management committee and board on disruption, including regular management committee meetings held in Silicon Valley, where the leadership team participates in sessions with frontier AI labs. "Our next one is in the next few weeks," Lohr said, "and I'm going to bring the entire management committee to one of the AI frontier companies… We're going to meet with their chief scientist, we're going to meet with their CISO."

Lohr said he is hoping to realize the "great ideas that we have in the firm," but operating more "on the edge" than the traditional approach of the Big Four consulting giant. He said he plans to rotate rising KPMG staff through "edge companies" as founders and operators, arguing the experience will make them "better accountants" and "better consultants" when they eventually return to core roles — mirroring, he said, the traditional Big Four model where talent cycles in and out of the firm but stays part of its alumni network.

An awesome responsibility

In June 2026, AI-detection firm GPTZero found that a flagship KPMG International report, "Total Experience: Redefining Excellence in the Age of Agentic AI," cited case studies involving UBS, Swiss Federal Railways, and the U.K.'s National Health Service that simply did not exist. Forty of the report's 45 citations were later found to be fabricated, and roughly half of its factual claims were either false or misattributed.

Lohr revealed that KPMG has recently launched an AI tool called Hawk that looks at all of its content and assesses AI hallucinations, and it has launched for most KPMG content — sort of an in-house GPTZero. A KPMG spokesperson clarified that Hawk is mainly used by the marketing department to vet thought leadership pieces, whereas CT&I's remit goes beyond internal transformation and internal tooling and is much more focused on the future of how KPMG serves clients.

GPTZero founder Edward Tian, a Princeton researcher and former Bellingcat team member, has since documented the same pattern at every other member of the Big Four: at EY, 60% of the citations in a December 2025 cybersecurity report were found to be hallucinated; at Deloitte Australia, a University of Sydney academic flagged 19 hallucinations in a single report; and at PwC, GPTZero put the odds that one Middle East governance report was entirely AI-generated at 84%.

In the interview, Lohr stressed that KPMG's AI acceptable use policy stretches back to his first work on the topic a decade ago. "At the end of the day," Lohr warned, "it's your output. You need to own the work product." He said he wants the entire workforce to understand that "AI is awesome, but it's also an awesome responsibility that everyone needs to own."

"As one of the folks that really started our AI business almost 10 years ago," he added, "the first offering was AI strategy, the second offering was trusted AI. That was something we've been adamant [about]. It's part of our culture."

Editor's note: This report has been updated to clarify that the new division will be announced Sept. 22 and effective Oct. 1.

This story was originally featured on Fortune.com.