The Talent Revolution of Artificial Intelligence

Artificial intelligence is rapidly reshaping how work is performed within the banking industry, altering roles, skills, productivity expectations, and operating models across the industry. While much of the public discussion has focused on technology adoption, leading research suggests that the more dynamic transformation is occurring within the banking workforce itself.  

Over the next three to five years, AI is expected to automate a meaningful share of routine and rules-based tasks across customer-facing roles, back-office operations, and administrative functions. In most cases, jobs are not eliminated outright; rather, they are redefined. Human roles increasingly shift toward judgment, exception handling, and strategy while AI agents perform more research, analysis, monitoring, and execution of tasks at scale.  

This transition introduces both opportunity and risk. Productivity gains are material, but uneven. Entry-level roles and traditional talent pipelines are altered. Governance, accountability, and regulatory expectations are evolving faster than many workforce models.  

As a result, talent strategy and operating model design are emerging as critical determinants of whether banking institutions realize AI’s promised value. Whether this evolution occurs at a measured, predictable, and manageable pace, or is fraught with unanticipated disruptions, remains to be determined. 

One thing is clear: AI’s most significant impact in banking is not the technology itself, but how it impacts how work is structured, roles are defined, and productivity is realized across  
the workforce.  

What we know. Leading research consistently finds that AI automates tasks rather than entire jobs, reshaping how work is performed across nearly all banking functions. In regulated industries, AI adoption progresses fastest where work is high-volume and repetitive; rules-based or data-intensive; and highly document-driven. As AI takes on more analysis, monitoring, and execution tasks, human roles increasingly shift toward higher-order business planning and system orchestration functions, and tasks requiring creativity, judgment, learned experience, decision oversight, people leadership, relationship-building skills, and performance governance.  

What is new. Workforce and operating model design—not AI model performance or data integration—is emerging as the primary constraint on AI value realization. There are generally considered to be four stages of AI performance readiness: novice, experimenter, practitioner, and expert. Today, fewer than one percent of those working in potentially AI-impacted functions are considered expert; whereas a full 96 percent are novice or experimenter, and according to AI workforce transformation specialists Section AI, most of those workers will not move beyond the experimentation phase in 2026.  

Institutions risk missing the workforce productivity implications and organizational performance opportunities of AI.  

Why this matters. AI adoption decisions made today will implicitly define future banking job roles, talent pipelines, and governance structures–often before those choices are explicitly discussed or agreed. These risks and opportunities are shaping the collective future of the industry.  

AI’s Impact on Existing Employees 

Today there are roughly two million employees at work within the U.S. banking industry. While the numbers of employees and distribution of roles can range widely across industry tiers and segments, the functional categories of work remain relatively consistent across institutions. Most banking institutions—both retail and commercial—employ a mixture of customer service personnel, back-office operators, product and domain specialists (including risk, compliance & fraud), technologists, marketers and data analysts, executives, and relationship managers. 

Predictions about the impact of AI on this diverse workforce during the next five years vary widely, ranging from less than 10 percent to greater than 40 percent in some estimates. Regardless of which level of impact proves to be more realistic, it is generally agreed that workers who operate in rules-based, high-volume, data- and document-centric jobs that can be more easily automated are the most susceptible to the impacts of AI. Impact risks are likely to be greatest when there is little capacity or incentive for workers and leaders to evolve these roles to rely more on uniquely human talents like creativity and judgment.  

ProSight compiled the chart below. It shows the predicted impact of AI across various banking roles. The roles are plotted on two axes, with the X showing the ease of AI automation and augmentation, and the Y-axis showing the likelihood that the role can adapt to AI in new, performance-enhancing ways.  

The impacts illustrated here are already being experienced today in several areas:  

Contact centers were among the first to realize efficiency gains, as AI agents now routinely support basic inquiries, enabling staff to provide more complex customer care.  

In operations, AI-based workflows are prevalent in payments processing, reconciliation, servicing, and settlement.  

Fraud & anti-money laundering teams benefit from AI’s massive pattern recognition and anomaly detection capabilities for roles like transaction monitoring, alert triage, and investigations.  

Credit & underwriting teams use AI effectively for pre-screening, document analysis, and scoring.  

Other areas that show promising results from effective AI application include legal & compliance for document analysis, summarization, and monitoring; data & analytics for self-serve analytics and automated machine learning; and information technology for coding, quality assurance, and ticket management.  

The daily workforce efficiency gains are proving to be significant. In a recent interview with Business Insider, Bank of America CEO Brian Moynihan pointed out that AI-based coding advances have already reduced the bank’s engineering cycle times by 30 percent.  

According to Moynihan, “That saves us about 2,000 people.” In that same publication, Citi’s Jane Fraser noted that AI-based code reviews were applied more than one million times last year, leading the bank to save “around 100,000 hours of weekly capacity.”  

Most banking institutions cannot approach the level of savings realized by these top U.S. money centers, yet the opportunity to redistribute capacity is yielding significant productivity gains across the industry. According to Chris Nichols, Director of Capital Markets at SouthState Correspondent Division, “[Banking institutions] routinely see 347-480 hours of savings per month … even before broad training or workflow optimization was introduced.”  

The good news for bankers and banking employees is that most jobs, even those with high potential for automation and augmentation, are expected to evolve and adjust—becoming more strategic, human-centric, creative, and business-focused—as AI becomes more pervasive in the workforce. Under the best circumstances, this shift will unlock a substantial productivity boon as talent can be refocused on delivering more projects more swiftly, because time-to-deliver and cost of execution will present less of a limitation. 

Designing Human + AI Banking Teams  

As banking institutions expand their use of artificial intelligence, a clear distinction is emerging between AI agents and traditional automation. Traditional automation focuses on rule-based task execution, such as moving data, triggering workflows, or enforcing predefined business logic.  

AI agents, by contrast, are designed to interpret context, reason across information sources, and take action in support of specific goals. These agents increasingly function as digital collaborators rather than tools, assisting human employees in complex, judgment-heavy environments such as fraud investigation, compliance monitoring, customer service, and credit analysis.  

In practice, banking institutions are beginning to define AI agent “roles” that align with existing business functions. Examples include Fraud Analyst Assistants that triage alerts and surface investigative insights, Compliance Monitors that scan transactions and communications for emerging risks, and Service Agents that support frontline staff by generating responses or summarizing customer histories. Importantly, these roles are not monolithic; they vary in scope depending on whether the AI is used for automation, augmentation, or autonomy. 

• Automation replaces discrete tasks.  

• Augmentation enhances human performance by providing recommendations or insights.  

• Autonomy allows AI systems to act independently within defined guardrails, often with escalation pathways to humans.  

These distinctions are reflected in emerging human-to-AI workflow patterns. In  
human-in-the-loop models, AI supports analysis or recommendations while humans retain decision authority, a pattern commonly used in credit underwriting, fraud resolution, and regulatory reporting.  

Human-in-the-loop models involve AI systems operating at scale with periodic human supervision, such as continuous transaction monitoring or operational risk surveillance. In more advanced deployments, banking institutions are experimenting with AI-first workflows in which AI handles routine interactions or analyses and escalates exceptions to human experts—an approach increasingly seen in customer service, AML alert handling, and internal knowledge management.  

Early lessons from banking institutions and other regulated industries suggest that successful human-to-AI partnerships depend less on technical sophistication and more on clarity of roles, escalation protocols, accountability, and culture.  

In retail banking, AI is being used to support frontline staff through real-time customer insights, personalized product recommendations, and conversational service agents that handle high-volume inquiries. In commercial banking, AI is often deployed to assist relationship managers and credit teams by synthesizing financial data, monitoring portfolio risks, and accelerating deal preparation.  

Across the organization, use cases vary by function. Front-line workflows emphasize speed, consistency, and customer experience, while back-office applications focus on scale, accuracy, and risk reduction in areas such as operations, finance, and compliance. At the executive and management level, AI is increasingly used to support strategic planning, workforce analytics, and performance monitoring by aggregating insights across the enterprise.  

Taken together, these patterns underscore a central reality: AI’s greatest impact in banking is not as a replacement for human expertise, but as a partner that reshapes how work is organized, supervised, and executed within a highly regulated environment.  

AI Agents as Digital Teammates  

As AI takes on more autonomous and persistent responsibilities, banking institutions are beginning to confront a novel question: should AI agents be treated, in operational terms, as a distinct category of “worker”? While AI agents are certainly not employees, they increasingly perform tasks that resemble human roles, raising questions about accountability, supervision, and control. For example:  

Should AI agents be regulated, monitored, and measured more like traditional software or more like employees?  

Considering they require humans to guide and oversee their work, is there a role for AI-in-the-loop to monitor and even manage their human team members?  

Some leading institutions are responding by defining AI agent job descriptions that specify scope of authority, permitted actions, escalation requirements, and performance expectations. This approach helps clarify where human oversight is required, how risks are managed, and who is accountable for outcomes produced by AI systems. In regulated environments such as banking, this clarity is becoming essential rather than optional.  

Over time, the ability to effectively integrate AI agents into the workforce and operating models will become a differentiator. Banking institutions that can clearly articulate how human and AI roles are expected to interact and perform together, rather than allowing AI responsibilities to emerge informally, are likely to achieve greater scale, consistency, and trust in their AI deployments. 

The AI Talent Risk 

As AI adoption accelerates across banking, talent risk is becoming one of the most significant, and least evenly distributed second-order effects. While AI promises meaningful productivity gains, it also introduces new vulnerabilities related to skill relevance, workforce engagement, governance, and institutional resilience.  

Many banking institutions are discovering that the greatest near-term risk is not job displacement, but misalignment between the pace of AI adoption and the organization’s ability to retrain, redeploy, and support its existing workforce.  

One emerging risk is skills obsolescence, particularly in roles where AI rapidly absorbs routine analytical or administrative tasks. Employees whose roles are partially automated may experience erosion of core responsibilities faster than new, higher-value tasks are defined.  

This reality creates uncertainty, morale challenges, and uneven performance unless accompanied by deliberate job redesign and new learning pathways. At the same time, banking institutions face heightened execution risk if they become overly dependent on a small cohort of AI-literate employees, creating concentration risk similar to what many institutions experienced during earlier digital transformations.  

Human resources executives are grappling with questions about the evolving workforce. According to Susan Lamonica, CHRO at Citizens Bank, “We need people who can quickly learn, adapt, and change. Our technologists need to develop their business acumen, and our business folks need to develop their digital and technical fluency. The lines are blurring.”  

AI also creates meaningful talent opportunities for banking institutions that act early. Institutions that invest in reskilling, transparent workforce communication, and structured human-AI collaboration models are finding that AI can improve employee satisfaction by reducing low-value work and enabling greater focus on judgment, relationships, and innovation. People leaders are finding traction by increasingly leveraging “T-shaped employees,” who combine specialized vertical skills and technical depth with more horizontal business savvy, to assist in translating AI outputs into business context and cross-functional applications.  

Over time, banking institutions that treat workforce transformation as a strategic capability rather than an HR exercise should gain a durable advantage in speed, adaptability, and talent retention.