The moment: Representation without power

On January 1, 2026, Dame Debbie Crosbie, CEO of Nationwide Building Society, was appointed HM Treasury’s Women in Finance Champion. She now plays a leading role in the Women in Finance Charter, whose more than 400 signatories represent approximately 1.1 million employees in financial services. Firms commit to targets for female representation in senior leadership, link executive compensation to progress, and report annually.

The Charter has made a difference. Women now hold about 43% of Financial Times Stock Exchange (FTSE) 350 board positions. In 2011, that number was 10%. About 8% of FTSE 350 companies have a female CEO, despite women holding 36% of senior leadership roles. The numbers tell a roughly familiar story: Representation may have improved, but power has not kept pace. We have more women in the room compared to 2015, but we have not yet changed who runs it.

Written in collaboration with

Written in collaboration with

---

The arrival of artificial intelligence (AI) is occurring at precisely this moment of “representation without power.” AI is not the first disruption to reshape work in general and women’s work in particular. Industrialization, the Great Depression, the World Wars, the 2008 financial crisis and Covid‑19 all rewired labor markets and household economics. Each shock created new opportunities for women and, at the same time, exposed how fragile those gains were when systems were not designed with them in mind. AI sits squarely in that tradition, with a crucial difference: the speed and scale of automation.

For Dame Crosbie and the Charter, AI raises a central question:

Is AI going to close the gap between representation and executive power, or will it industrialize the structural biases we have spent decades trying to dismantle?

Without intentional leadership and reskilling strategies, AI could amplify progression and pay gaps rather than close them. Used intentionally, it can support more inclusive leadership pipelines by enabling transparency. The issue is no longer representation alone but how close female leaders are to the algorithmic levers of the firm and to the profit and loss (P&L) lines where AI value is created.

Will AI be the tool that hands women the algorithmic keys to the firm, or will the design, deployment, and engagement with these systems remain a new frontier of old-school exclusion?

The lens: Exposure, agency, attribution

To lead in an increasingly AI-intensive sector, we must look beyond upskilling and focus on three dimensions of executive power that will determine the next decade of leadership:

Exposure: Whose tasks are being automated? In high-income countries, approximately 9.6% of female employment is at high automation risk, compared to 3.5% for men. Women are more concentrated in back-office operational and administrative roles—the very tasks AI is already replacing.⁴

Agency: Who sits at the AI decision table? Who controls the research and development budgets? Power resides in the architectural roles—those who decide where AI is deployed, what data it learns from, and which risks are acceptable. Women represent only about a quarter of the global AI workforce and roughly 15% of tech leadership roles worldwide.

Attribution: When AI-enabled projects succeed, who is recognized? Is credit flowing to the technical teams or to the leaders who translated technology into business outcomes?

At board level, this exposure–agency–attribution lens is a practical way to think about gender‑intentional AI governance. It shifts the focus from “Do women have access to AI?” to “Whose work is being reshaped, who sets the priorities and budgets, and who gets credit and reward when AI delivers value?”

Issues to put on the agenda include:

1. The “broken rung” or a “missing ladder”? As AI takes over first-draft work, the analyst roles that served as the first rung of the ladder are thinning. Yet, used well, AI could also accelerate apprenticeship rather than erase it: Junior staff can use AI tools to interrogate models, run scenarios, explore edge cases, and prepare higher‑quality recommendations earlier in their careers.

What will replace the traditional analysts and graduate roles that serve as a learning ground for early-career women?

How do we ensure early-career women are gaining fluency as commanders of AI and driving change and adaptations rather than just supervising automated workflows?

2. The adoption gap: Research shows that men are 20% more likely to experiment informally with Generative AI (GenAI). Early evidence suggests this engagement gap is driven by a mix of behavioral, organizational, and cultural factors: differences in confidence and risk‑taking around new tools; patterns in who is given early access, sponsorship, and informal coaching; and the “double burden” of unpaid care and mental load that leaves many women with less discretionary time to experiment. None of these are about aptitude, but all of them affect who accumulates AI‑related advantage.
   
   In an environment where output and productivity are increasingly mediated by AI tools, lower adoption among women is likely to show up in performance metrics, utilization figures, and promotion decisions, even when underlying capability is identical. Without deliberate intervention, a usage gap risks hardening into a productivity gap, then a promotion and pay gap.

How will we measure AI adoption and usage by gender in financial services?

How do we prevent a confidence gap from becoming a promotion gap as productivity becomes increasingly mediated by these tools?

The structural risk: Alpha vs. hygiene split in AI deployment

A structural split is emerging in how AI is deployed across financial services, mirroring the old front‑office/back‑office divide.

• AI for “alpha” (male‑dominated): Deployed in trading, origination, and product coverage; positioned as innovation and revenue growth.
• AI for “hygiene” (female‑concentrated): Deployed in risk, compliance, and HR; positioned as efficiency, cost-saving, and regulatory necessity.

In financial services, this is not just a cultural split; it is an economic one. AI deployed in trading, origination, product, and advisory is directly tied to revenue growth and P&L ownership. AI deployed in risk, compliance, and HR is framed as cost control and regulatory hygiene. If women are concentrated in the latter and absent from the former, they are structurally positioned in efficiency layers rather than value‑creation roles. AI then becomes a force for concentrating economic power, not distributing it.

Question to put on the agenda:

Women are often the architects of AI governance. Does that governance come with real budget authority and P&L responsibility, or is it kept at the margins as a brake on “real” innovation? Which AI P&L lines are led by women?

If AI is used primarily to strip out junior work in female‑dominated functions while augmenting high‑judgment work in male‑dominated ones, the power gap will widen by design.

The pipeline challenge: Entry and mid-career disruption

Mid-career is where progression often stalls. AI accelerates this through role compression. As models take over intermediate analysis, roles tilt toward stakeholder management and human‑in‑the‑loop oversight—work that is essential but often categorized as “glue work” and under‑rewarded.

Over the next two to three years, many firms will face a stark choice: Treat mid-career roles as redundancy risks or deliberately redeploy talent into AI-augmented roles in risk, product, advisory, and client strategy. The impact will be particularly significant for many women in mid-career, where progression already tends to slow, with lower promotion rates into senior, P&L-linked roles.

Questions to put on the agenda:

• Reskilling vs. redeployment: Are we treating mid-career women as redundancy risks or as prime candidates for AI-augmented roles in client strategy and advisory? Are we sponsoring mid-career women into AI‑intensive, P&L‑relevant roles, or into oversight roles that are essential but low‑visibility?
• The burden of churn: If operational roles are restructured more frequently, how do we ensure women aren’t trapped in a cycle of constant reselection while the rewards of productivity gains flow elsewhere?

The opportunity: Why women are critical to AI outcomes

Rather than seeing AI purely as a threat, it can be framed as a catalyst that values the leadership qualities often found in the current cohort of female executives.

As AI commoditizes routine analysis and first-draft work, the premium on human judgment, contextual reasoning, and stakeholder trust increases. Women currently hold a disproportionate share of the roles where these capabilities are most critical: risk, compliance, audit, and complex client relationships. This is not a soft advantage. In financial services, these are precisely the functions where AI failure is most costly, where regulatory consequence is most severe, and where client trust is hardest to rebuild once lost. Women in these roles are not peripheral to AI deployment; they are essential to its responsible governance.

There is also a direct commercial case. Research consistently shows that diverse leadership teams make better-calibrated decisions, identify a broader range of risks, and are less likely to exhibit the groupthink that has preceded the most significant failures in financial services. As AI amplifies the consequences of both good and poor decision making, the cost of homogeneous leadership at the architectural level rises accordingly.

AI also gives financial institutions the tools to audit their own biases. Pay-gap analytics, promotion-pattern detection, and language analysis in performance reviews can surface the structural barriers that informal processes have historically obscured. Used deliberately, AI becomes an organizational diagnostic, not just a productivity tool.

Finally, women represent a significant and underserved market. Firms that design AI-enabled products and services with women in leadership roles are better positioned to serve women as savers, investors, and business owners—a market the UK financial services sector has systematically underleveraged.

The question is not whether women belong at the AI frontier. The question is whether institutions will act before the window closes.

Implications: Questions for boards and policymakers

The transition is not without friction. Several areas require deliberate boardroom intervention to ensure the “male default” is not hard‑coded into the future. First, the automation risk gap—women’s higher concentration in operational and support roles—is a predictable consequence of decades of occupational sorting. Without deliberate intervention, this structural disadvantage will deepen as AI accelerates the very patterns of role concentration it inherits.

Second, in environments where informal experimentation is how people discover AI use cases and build confidence, lower engagement with AI risks translating into lower productivity and, over time, slower progression. To prevent an engagement gap from quietly becoming a productivity and promotion gap, firms need to create protected time, access, and sponsorship for women to experiment.

Third, if AI models are trained uncritically on historical data, they will faithfully reproduce the biases of the past; without women in the architectural roles that select, audit, and challenge these data sets, we risk building a financial infrastructure that is structurally blind to women’s financial realities in hiring, promotion, and customer decisions.

Finally, as AI changes how work is designed and distributed, the career journeys of women become more fragile. Entry‑level rungs can disappear, mid-career roles can be compressed into under‑recognized “glue work,” and senior roles can remain concentrated in AI‑enabled P&L lines led by men. Safeguards are needed to keep the leadership pipeline flowing and to prevent the gains of the past decade from being eroded by the next wave of automation.

A global inflection point: The UK’s leadership opportunity

The UK is uniquely positioned to lead this conversation. With a decade of the Women in Finance Charter and a robust AI regulatory framework, we can decide if the next generation of financial services will be built with women or merely used on them.

Compared with peers in the US, EU, and parts of Asia, the UK combines relatively high female board representation with an emerging, principles‑based approach to AI regulation. That combination gives the UK both the responsibility and the opportunity to set gender‑intentional standards for AI‑enabled finance that others can follow or adapt.

If the UK is serious about leading on women, AI, and finance, it must treat AI deployment and skills as deliberate design choices, not as technical details delegated to specialist teams. That means ensuring women are present in AI‑intensive P&L roles as well as in risk and compliance; aligning businesses’ Women in Finance Charter commitments with AI‑related metrics such as AI adoption, budget ownership, and promotion outcomes by gender; and investing in the AI and STEM capabilities that equip women to act as architects of the next financial system, not just its supervisors.

Past disruptions—from the 2008 financial crisis to Covid‑19—have often rolled back women’s gains when gender was treated as an afterthought. AI gives the UK a chance to do it differently: to design systems that recognize women as architects of financial innovation, not just as workers whose tasks can be automated or customers whose data can be harvested.

For CEOs, executive committees, and boards, the next two to three years will involve a set of nondelegable decisions:

• where to deploy AI for growth vs. efficiency and whether women are present in the value‑creation P&L lines;
• whether to treat mid-career women in operational roles as redundancy risk or as the core talent pool for AI‑augmented roles in risk, product, advisory, and client strategy;
• who owns AI budgets, who chairs AI decision forums, and how executive pay is linked to gender‑balanced outcomes in AI‑intensive areas; and
• how to close the AI engagement gap so that women are not left behind as productivity, performance, and promotion become increasingly AI‑mediated.

What should the UK do differently in the next 12 months? What specific policy or regulatory change would make the difference?

What comes next?

This paper is intended as a catalyst for senior‑level reflection and dialogue, not a set of prescriptive solutions. The questions it raises are deliberately open: What should the UK do differently in the next 12 months, and what specific policy or regulatory change would make the difference?

The immediate aim is to surface the tensions that leaders need to engage with now: who owns AI budgets, where women sit in the AI value chain, how career pipelines are being redesigned, and what “gender-intentional AI” might mean in practice.

These are not abstract questions. They map directly onto the exposure–agency–attribution framework set out in this paper: measuring where women are most at risk of displacement, ensuring they hold decision-making power over the AI systems being built and deployed, and creating the structures through which their contributions are recognized and rewarded.

This paper was first shared with the UK CEO group on April 14 to anchor discussion on how AI is already reshaping women’s roles, power, and progression in UK financial services. It then was brought to the Global CEO group at the Stratos × Cambridge Global Executive Symposium, held April 23–24 in Cambridge, to place the UK experience in a wider international context and invite comparative perspectives.

Looking ahead, there is an opportunity to take this work further in a more structured way. Potential next steps could include:

• developing a practical toolkit for gender-intentional AI in financial services over the coming months, with an initial version by August 2026, aligned with the Women in Finance Charter and grounded in measurable commitments on AI exposure, agency, and attribution; and
• piloting a diagnostic and measurement framework with a small group of institutions, building toward initial results by December 2026, to track AI exposure, agency, and attribution by gender, generating the evidence base needed to move from aspiration to accountability.

Alongside this, we will convene a small group of financial services CEOs, regulators, and practitioners to define leading practice and shape UK and global standards for the sector.

The direction and pace of this work will be shaped by the priorities that emerge from these conversations. But the central ambition is clear and will not change: to ensure that, as AI becomes the operating system of financial services, women are architects of that future, not afterthoughts in it.