When a client hires PwC, they expect a team of humans with deep expertise to scrutinize their numbers, flag their tax exposure, and guide them through complexity. Starting March 19, 2026, PwC One — the firm's newly launched AI platform — will in some cases run that review before a PwC person ever looks at the file.
"The new tax and consulting tools could be accessed in the first steps, without a PwC person in the loop," PwC US Senior Partner and CEO Paul Griggs told the Financial Times. The Guardian, reporting on the FT interview, noted the quiet inversion of the consulting engagement model: not humans assisted by AI, but AI checked by humans. The question PwC's competitors and clients are now working through is what that means for the rest of the industry.
PwC One launches with six automated services: a financial statement analyzer, a tax incentives explorer, a transfer pricing policy compliance reviewer, a transfer pricing intercompany agreement reviewer, a GHG emissions navigator, and a financial due diligence assistant. The platform runs on a multi-agent architecture combining traditional AI, generative AI, and machine learning, with a conversational interface backed by a dynamic structured canvas that surfaces data and workflows in real time. Clients can subscribe to those services without engaging a PwC team for the initial steps — a departure from how consulting engagements have historically been scoped and billed.
Griggs was blunt about the cultural shift that underpins it. "An employee who thinks they have the opportunity to opt out of AI is not going to be here that long," he told the Financial Times, as reported by The Guardian. "I don't think anyone gets a free pass here. Anyone." PwC cut 5,600 staff last year, bringing global headcount under 365,000. Griggs confirmed the firm is hiring fewer traditional accountants proportionally, favoring engineers and data specialists instead. The message to the partnership is that the firm's future involves fewer humans doing what the AI now does — and the door is not held open for those who want to wait it out.
The pricing model is changing too. PwC One will involve subscription or consumption-based elements rather than the hourly billing that has defined professional services for decades. "Over time, it will move more and more of our work to outcomes pricing," Griggs said. That matters beyond optics: it signals that PwC is willing to take on some of the risk and uncertainty of its own delivery efficiency, which is a meaningful shift in how the firm structures its economics.
Scott Likens, global and US Chief AI Engineer at PwC, described the platform as the culmination of a longer arc. PwC began experimenting with AI assistants in 2019 but found early chatbots too limited. By 2022, the firm announced a multibillion-dollar commitment to AI. The Agent OS that preceded PwC One — connecting agents across Anthropic, AWS, GitHub, Google Cloud, Microsoft Azure, OpenAI, Oracle, Salesforce, SAP, and Workday — now has over 250 AI agents deployed internally across the firm, according to PwC's own reporting. Clients are seeing measurable effects: a hospitality client reduced review times by up to 94% through the system, and a healthcare client cut staff administrative burden by nearly 30%.
The competitive context matters. By 2025, all four of the largest global accounting and consulting firms had launched multi-agent AI platforms: PwC Agent OS, Deloitte's Zora AI, EY's ai platform, and KPMG's Workbench. EY went first among the Big Four in March 2025, deploying more than 150 specialized tax agents to support 80,000 professionals. Deloitte built Zora AI in partnership with Nvidia, targeting finance and procurement bottlenecks. The race is real — and it is moving faster than the firms themselves can fully characterize in their own announcements.
Matt Wood, PwC's US and Global Commercial Technology and Innovation Officer, described PwC One as part of the firm's effort to embed AI directly into its methodologies rather than bolting it onto existing workflows. "PwC is integrating AI directly into its methodologies and professional workflows — extending the reach of its expertise while keeping human judgment firmly at the center of every critical decision," the press release states. The tension between that framing and the "no PwC person in the loop" claim is the story's honest core: the human judgment stays central in theory, but the initial analysis no longer requires it in practice.
That gap — between the firm's rhetorical commitment to human oversight and the operational reality of fully automated initial steps — is where the real questions live. For clients, it raises questions about accountability when an AI flags an issue that a human reviewer then endorses without independent re-examination. For the industry, it raises the question of what the consulting partner of 2030 actually does.
According to K2 Consulting Research, global consulting grew 5.5% in 2025, doubling the previous year's growth rate — a buoyant market that PwC is using to signal that its AI investment is paying off. Whether that growth is a tailwind the firm is riding or a result of the efficiency gains it's now openly advertising is, appropriately enough, one of the questions PwC One is not yet designed to answer.
