When Banking Changes Faster Than Bankers

Banking has traditionally relied on qualifications, experience and seniority as indicators of professional competence. A CA, MBA or M.Com provides a foundation. Experience adds judgement. Seniority brings institutional knowledge and authority. For much of modern banking, this model has worked reasonably well because the profession changed gradually. Knowledge acquired early in a career could remain relevant for many years.

Now, that assumption is becoming harder to sustain. Artificial intelligence is entering customer service, fraud detection, credit assessment, compliance, risk management and other important areas of banking. In many cases, AI is not simply automating an existing activity but is changing how that activity is performed, what information is used, and what a professional must understand to exercise sound judgement.

This raises a difficult question: has the competence of bank decision-makers kept pace with the changing nature of banking?

This is not an argument against qualifications or experience. Both remain essential. The issue is that a qualification, or even past professional success, may become a less reliable measure of competence over an entire career. The resulting problem can be described as competence lag. It occurs when the requirements of a professional role evolve faster than the knowledge and capabilities of the person performing it.

In banking, competence lag is not merely a human-resources concern. It can affect credit allocation, compliance costs, operational resilience, customer treatment and, ultimately, trust in the financial system.

Shorter Shelf Life

Consider a credit executive with 25 years of underwriting experience. That experience may provide deep judgement about borrowers, industries, economic cycles and credit behaviour. Now suppose the executive is presented with a credit recommendation generated or materially influenced by an AI system using hundreds of variables.

The executive does not need to become a machine-learning specialist. The decision, however, requires additional capabilities. What information is the system using? What are its limitations? How does it behave outside normal conditions? Could any particular variables introduce bias? Can its recommendation be explained? When should human judgement override the model?

The executive’s banking experience has not become less valuable. The environment in which that experience must be used has changed. The relevant variable, therefore, is not age but current capability.

The wider workforce is changing too. The World Economic Forum’s Future of Jobs Report 2025 estimates that 39% of workers’ core skills are expected to change by 2030. It also reports that 97% of employers in financial services expect AI and information-processing technologies to transform their businesses by then. These are global rather than India-specific estimates, but they indicate the scale of the challenge facing financial institutions.

Banking has an additional complication. When AI contributes to a financial decision, responsibility cannot simply be transferred to the machine. Someone still must decide whether its recommendation should be accepted. Formal authority may remain with the banker even when part of the informational basis for the decision has shifted to the machine.

That makes the competence of the human decision-maker more, not less, important.

Literacy Is Not Competence

Completing an AI training programme should not automatically be treated as proof of readiness. An employee can understand the terminology and receive a certificate but still be incapable of challenging an AI-supported recommendation in a real business situation.

Banks therefore need to distinguish between AI literacy and AI competence. While literacy will give an understanding of what AI can and cannot do, competence will allow employees to use, question, escalate, override or govern an AI-supported decision within their professional responsibilities.

The level of competence required will depend on the role played by the system. The closer AI comes to determining an outcome, the stronger the requirements for human understanding, oversight and accountability should become.

This changes how banks should assess readiness. A credit executive could be asked to review an AI-generated recommendation and identify the questions that must be answered before approval. A risk professional could be given an unusual model output and asked when human intervention would be appropriate. A senior manager could assess the risks of deploying an AI system supplied by an external vendor.

Such exercises test judgement rather than mere knowledge.

The same principle has implications for recruitment and promotion. The answer is not to replace banking qualifications with technology certificates or experienced bankers with younger technology professionals. It is to bring different forms of expertise together so that decisions become stronger.

Governance Competence

The issue becomes more serious when attention shifts from employees using AI to the people approving or overseeing consequential decisions. A bank may have a capable technology team and sophisticated models, but responsibility still rests with managers, committees and boards.

Senior decision-makers need enough understanding to ask informed questions. They should be able to examine the data a system uses, the assumptions underpinning it and the reliability of its output. They should also ask whether bias could affect the decision, whether the reasoning can be explained, who is accountable when the system causes harm and when human judgement should intervene.

These are not merely technology questions. They concern risk, ethics, governance and accountability. An AI system may produce a statistically sound recommendation and still raise concerns about fairness, explainability or suitability. Responsibility for examining those concerns cannot simply be delegated to a technology team or transferred to an algorithm.

This is where competence lag becomes a governance risk. A bank may have sophisticated technology without having equally sophisticated human oversight.

Existing international standards do not prescribe a single test of AI competence, but they increasingly point towards the capabilities institutions will need. Basel’s Core Principles for effective banking supervision and Corporate governance principles for banks emphasise effective board oversight, appropriate knowledge and experience, robust risk-management capabilities and the continuing development of directors and senior personnel.

Work by the Bank for International Settlements, including Humans keeping AI in check and Intelligent financial system: how AI is transforming finance, has similarly highlighted expertise, accountability, transparency, fairness and continuing human responsibility. The Financial Stability Board’s June 2026 consultation report proposes 12 sound practices covering organisation-wide AI governance, different stages of AI development and deployment, and AI-related cyber, information and third-party risks.

India is moving in a similar direction. The RBI’s FREE-AI Committee Report, released in August 2025, contains seven foundational principles and 26 recommendations covering areas including capacity, governance, protection and assurance. The RBI’s Annual Report 2025–26 records recommendations including capacity building within regulated entities, a board-approved AI policy, AI-system governance, AI audit and an AI inventory.

The FREE-AI report is a committee framework and set of recommendations rather than binding regulation. Its broader message is nevertheless important: institutions need sufficient internal capability to understand, govern and manage the risks created by the technologies they deploy.

The AI Diffusion Matrix, developed by AI4India and the Centre for Study of Science, Technology and Policy in consultation with NITI Aayog, makes a related point by encouraging organisations to assess technology readiness, governance, risk management and value creation. Although it is not a banking standard, it reinforces the principle that AI readiness is not simply the acquisition of technology.

Knowledge Risk

There is another institutional dimension to competence risk. Banks may invest heavily in small groups of employees who understand AI systems, data architecture, model-risk controls or technology vendors. If this knowledge becomes concentrated in too few people, their departure can create operational risk. The technology may remain in place while the institution’s ability to understand and govern it weakens.

The same risk applies to traditional banking knowledge. Much of the judgement developed through unusual borrowers, difficult fraud cases, stressed situations or complex regulatory issues is never fully captured in manuals or systems.

Banks with significant concentrations of experienced personnel approaching retirement may face one form of succession risk, while those dependent on small groups of technology specialists may face another. In both cases, the issue is whether critical knowledge is identified, shared and preserved before the people who carry it leave.

Banks have already built redundancy into critical technology systems. Increasingly, they may also need redundancy in knowledge. Documentation, cross-training, internal mobility and development of the next level of talent can reduce dependence on scarce individuals, teams and external vendors.

Important experience should not remain only in the memory of the people who possess it. Banks could establish a policy for documenting the reasoning behind consequential AI-related decisions.

Over time, these records could form an institutional case library. New employees could learn from situations faced by the bank and existing ones could use them to test their judgement. Future reviews could examine whether the assumptions made at the time remained valid.

This would also change the meaning of succession planning. Succession should not simply answer, “Who will replace this person?” It should also ask, “What knowledge and judgement does this person carry, and how will the organisation preserve and transfer it?”

What Banks Can Do

Banks can respond to competence lag in four practical ways. They can define the AI competence required for each decision-making and oversight role, assess that competence through realistic scenarios rather than course completion alone, document the reasoning behind consequential AI-supported decisions, and build succession plans for both experienced banking judgement and specialised technical knowledge.

These measures should form part of workforce strategy rather than remain isolated training initiatives. Recruitment, promotion, professional development and succession decisions may increasingly require an additional question: how capable is this person of learning what the next version of the job will require?

A future-ready banker will need financial knowledge, regulatory understanding and business judgement, along with sufficient technology literacy to work effectively with specialists and question AI-supported decisions when necessary.

The objective is not to discard experience, but to ensure that it continues to work effectively alongside new knowledge.

Banking has always depended on trust. In an AI-enabled banking system, that trust will increasingly depend on whether customers, boards and regulators can be confident that technology is being used by people who understand both its capabilities and its limitations.

The debate about AI in banking should therefore not be reduced to how many jobs machines might replace. A more immediate question is whether those who retain decision-making authority will continue developing fast enough to exercise it responsibly.

When the machine gives the answer, does the person with the authority to accept it know enough to question it?