Financial services is one of the most AI-exposed industries in the world, and one where career paths have traditionally been rigid: analyst, associate, manager, director. A PwC survey of 1,004 senior executives at US firms with at least $500 million in revenue, published on 3 August 2026, shows how AI is rewriting that ladder from both ends.
Key takeaways:
- 91% of executives say they are increasing compensation for employees with AI skills; 58% will tie pay directly to AI-enabled productivity.
- 86% agree that AI skills training is more valuable than an MBA for many new hires.
- Nearly 8 in 10 expect their workforce to shrink by at least 20% over the next five years. 30% name entry-level roles as most vulnerable, followed by middle management (26%).
- 77% say most AI investments are not yet delivering measurable return.
- 62% plan to hire for AI-specific skills in the coming year; 61% plan to upskill existing staff.
- Shadow AI is widespread: 90% say staff using tools outside governed systems creates regulatory risk.
The two-sided message
Read together, the numbers describe a workforce that is smaller, better paid and more AI-fluent.
| Executives say... | Share |
|---|---|
| Increasing pay for AI skills | 91% |
| AI training more valuable than an MBA for many new hires | 86% |
| Workforce will shrink ≥20% within five years | ~80% |
| Entry-level roles most vulnerable | 30% |
| Middle management most vulnerable | 26% |
| Employees concerned about job security from AI | 44% |
| Employees use AI only when required | 43% |
PwC's own commentary is pointed. Firms, it says, are often "planning for a smaller workforce, and hoping AI fills the gap" rather than designing the workforce they will actually need. Only 42% have done enterprise-wide modelling of AI's effect on labour capacity, and of those, just half have looked at what redesigning processes around AI would mean.
The entry-level pipeline problem
Thirty per cent of leaders point to entry-level roles as most exposed. PwC flags the consequence explicitly: fewer juniors today means a smaller pool from which to develop tomorrow's managers and leaders. The firm urges deliberate investment in career pathways and leadership development precisely because the old apprenticeship model is thinning.
This echoes the Stanford payroll data on software developers. The pattern is now visible in two very different industries.
What firms want from candidates
The survey gives an unusually direct answer:
- AI fluency over credentials. 86% rate AI skills training above an MBA for many new hires.
- Continuous, role-specific learning. PwC recommends peer-to-peer and on-the-job programmes over one-off enterprise-wide courses.
- Agent oversight as a growth role. The report suggests people displaced from automated tasks could be upskilled into the growing demand for managing and supervising AI agents.
The governance angle
Nearly nine in ten executives say their firm has clear accountability for AI-agent decisions, yet there is no consensus on who owns the risk of material harm: 27% say CEO and board, 16% a technology leader, 15% risk and compliance, 12% a business-unit leader, 8% the human who acted on the output. That disagreement, in a heavily regulated industry, is itself a career signal: AI governance and risk roles are being invented in real time.
Concept to take away: the pay premium and the headcount cut are the same phenomenon. Firms expect AI to let fewer people do the work, so each remaining person is worth more, provided they can direct the tools. The premium is not a bonus for enthusiasm; it is the market price of being one of the people kept. Position yourself accordingly.
Why learners should care
If you are heading into banking, insurance, asset management or fintech, this is the clearest statement yet of what employers will pay for and what they plan to stop hiring for. The advice writes itself: arrive AI-fluent, aim above the tasks agents will take, and consider governance and oversight as a growth path.
Note: this is a US survey of large firms and reflects executives' expectations, not outcomes. PwC itself warns against "knee-jerk workforce reductions based on models that may not pan out", noting some firms are already projecting more modest gains as they account for the real cost of AI usage.
