The Big Four Are Partnering With AI.

The Scale of the Commitment

The partnership activity across the largest professional services firms over the past two years represents a structural repositioning, not an incremental technology upgrade. In May 2026, KPMG announced a global alliance with Anthropic to embed its Claude model directly into Digital Gateway, the platform KPMG’s 276,000 employees across 138 countries use to deliver client work, beginning with tax and legal services. The alliance also names KPMG a preferred partner for deploying AI into private equity portfolio companies and extends into cybersecurity (Anthropic, 2026a).

PwC expanded its own alliance with Anthropic in the same month, committing to roll out Claude tools across a workforce of hundreds of thousands, establishing a joint Center of Excellence, and launching a programme to train and certify 30,000 of its professionals. PwC also created a dedicated finance business group, an Office of the CFO, built on the technology, and reported delivery-time improvements of up to 70% across its early production deployments (Anthropic, 2026b).

These commitments sit within a wider shift. The Cambridge Centre for Alternative Finance, in its 2026 Global AI in Financial Services Report produced with the Bank for International Settlements, the IMF, and the World Economic Forum, surveyed 628 financial institutions, vendors, and regulators across 151 jurisdictions. It found that 81% of financial services firms are now deploying AI at some level, but that only 14% currently regard AI as transformational to their strategy, a result the report describes as a significant execution and business integration gap (Garvey et al., 2026).

The Execution Gap the Announcements Obscure

Investment figures and partnership announcements measure intent, not capability. The CCAF report found that while deployment is widespread, 55% of industry respondents and 63% of surveyed regulators report difficulty measuring the actual value of that deployment. Among large financial institutions, that figure rises to 76%. The productivity gains that do exist are concentrated in technology, data and product functions, at 79%, while back-office and front-office functions report lower impact at 75% and 69% respectively. The report identifies talent shortages, data quality, and legacy systems as the three core constraints on scaling AI (Garvey et al., 2026).

The professional body evidence reinforces this. The AICPA and CIMA surveyed 1,446 senior finance and accounting leaders globally in late 2025. While 88% identified AI as the most transformative technology trend facing their function over the next 12 to 24 months, only 8% described their organisation as very well prepared to manage it, and 56% identified generative AI as their most prominent skills gap. Half of respondents cited a lack of human capital, skills, and talent as the single biggest barrier to adoption (AICPA and CIMA, 2025). The Financial Services Skills Commission places the AI skills demand gap across the UK financial sector at 35 percentage points, with an estimated 160,000 professionals requiring urgent upskilling (Financial Services Skills Commission, 2025).

Taken together, the data points to a capability that has been widely procured but not yet widely converted into measurable value. Adoption has outpaced the organisational and human capacity to use the technology well.

What Is Genuinely at Risk

The nuanced position is not the comforting one. The Financial Services Skills Commission identifies financial services as the sector most exposed to AI of any industry, with routine and language-heavy tasks the most automatable. These are often concentrated in entry-level and back-office roles, which have historically served as the apprenticeship layer of the profession, where early-career professionals build the judgment that more senior work depends on. Changes to that layer carry structural consequences beyond individual roles.

At the same time, the evidence does not support a simple displacement narrative. The Financial Services Skills Commission concludes that AI will mainly alter existing roles rather than create entirely new ones, with only an estimated 0.5 to 1.5 percent of the workforce requiring specialist AI skills, while the large majority require a working understanding of the technology rather than technical expertise (Financial Services Skills Commission, 2025). The CCAF report reaches a similar conclusion at a global level, finding that reskilling and role transformation, rather than outright displacement, is currently the dominant workforce expectation across the institutions it surveyed (Garvey et al., 2026). The question for someone entering the profession is therefore less whether roles will exist, and more whether the entry points through which expertise has traditionally been built will continue to function in the same way.

The Competence the Partnerships Cannot Buy

The most consistent finding across these sources is that the binding constraint on AI in finance is not the technology. It is the people and the data around it. The CCAF report names talent shortages, data quality, and legacy architecture as the core barriers to scaling AI, the same bottlenecks identified in its 2020 predecessor (Garvey et al., 2026). McKinsey research, cited by the Financial Services Skills Commission, found that 46% of global business leaders identify skills gaps in their workforces as a significant barrier to adoption (McKinsey, 2025, cited in Financial Services Skills Commission, 2025). The AICPA and CIMA found the same: half of finance leaders rank a lack of skills and talent as their single largest obstacle (AICPA and CIMA, 2025).

What none of the partnerships or investments can manufacture is the professional who understands both sides at depth: who can evaluate a financial model and interrogate the system that generated it, who understands governance frameworks and can apply them to AI deployment, and who can operate at the boundary between domain knowledge and technical infrastructure. The technology is becoming a commodity. The judgment to deploy it well is not.

The constraint is not the tool. It is the capacity to use it with judgment, and that is not something a partnership announcement can create.

A Note on Where I Am Writing From

This piece is partly a response to the anxiety that these announcements produce for anyone early in their career. The instinct when reading them is to ask whether the window has already closed. The evidence suggests a more accurate question: whether the conditions that made the old entry points valuable still hold, and what the new ones require.

The consistent conclusion across the institutional research — from Cambridge to the AICPA to the Financial Services Skills Commission — is that the scarce resource is not access to AI but the combination of domain knowledge, judgment, and the ability to work at the boundary between finance and technology. That profile is built deliberately, over time. Whether the implementation gap narrows quickly or slowly, the people who can stand in it remain in short supply. That is the more useful thing to take from the headlines.

References

AICPA and CIMA (2025) Future-Ready Finance: Technology, Productivity, and Skills Survey. New York: Association of International Certified Professional Accountants. Available at: https://www.prnewswire.com/news-releases/ai-transformation-opens-door-for-finance-professionals-to-build-future-ready-skills-aicpa-and-cima-survey-finds-302644839.html (Accessed: 19 May 2026).

Anthropic (2026a) KPMG integrates Claude across its core business and workforce of more than 276,000 in strategic alliance. Anthropic News, 19 May. Available at: https://www.anthropic.com/news/anthropic-kpmg (Accessed: 22 May 2026).

Anthropic (2026b) PwC is deploying Claude to build technology, execute deals, and reinvent enterprise functions for clients. Anthropic News, 14 May. Available at: https://www.anthropic.com/news/pwc-expanded-partnership (Accessed: 22 May 2026).

Financial Services Skills Commission (2025) Unlocking AI’s Potential: The Skills That Matter. London: Financial Services Skills Commission, May. Available at: https://financialservicesskills.org/wp-content/uploads/2025/05/Artificial-Intelligence-Report-2025.pdf (Accessed: 21 May 2026).

Garvey, K., Zhang, B., Roberts, I., Abdou, F., Barrie, H.M., Arner, D.W., Nolens, B., Frost, J. and Segoviano, M. (2026) 2026 Global AI in Financial Services Report: Adoption, Impact and Risks. Cambridge: Cambridge Centre for Alternative Finance, Cambridge Judge Business School, University of Cambridge, in partnership with the Bank for International Settlements, International Monetary Fund, and World Economic Forum. Available at: https://www.jbs.cam.ac.uk/faculty-research/centres/alternative-finance/publications/2026-global-ai-in-financial-services-report/ (Accessed: 30 May 2026).

McKinsey and Company (2025) cited in Financial Services Skills Commission (2025) Unlocking AI’s Potential: The Skills That Matter. London: Financial Services Skills Commission, May. Available at: https://financialservicesskills.org/wp-content/uploads/2025/05/Artificial-Intelligence-Report-2025.pdf (Accessed: 21 May 2026).