Autonomous banking will soon move from differentiator to baseline capability: HCLTech exec

As artificial intelligence moves from experimentation to enterprise deployment, banks are beginning to rethink not just digital channels but the way financial services are designed and delivered. For over a decade, the industry has focused on digitising processes and improving efficiency. The next phase, however, could be far more transformative—autonomous banking systems capable of sensing, deciding and acting in real time, believes Monu Kurien Mathew, SVP and Head, Business Solutions, BFSI at HCLTech. In an interview with TechCircle, he explains why the shift toward autonomous banking is gathering momentum, how banks can modernise without ripping out legacy cores, and why governance frameworks will be central as AI agents begin making large volumes of operational and credit decisions. Edited excerpts.

Banks have spent over a decade on digitisation. At present, what marks the shift from “digital banking” to truly autonomous banking?

Over the last decade, banks have focused on digitising processes—making existing workflows faster and more efficient. Autonomous banking represents a deeper shift. It reimagines the value chain and operating model itself, moving from human-led systems supported by technology to technology-led systems guided by human judgment.

Today, banks have automation, AI models and cloud platforms, but these capabilities often operate in silos. Autonomy emerges when these elements are connected end-to-end—through intelligent automation, AI agents and decisioning layers built on a composable core—allowing systems to sense, decide and act in real time.

The industry is moving from simply doing things faster to doing things better: enabling self-initiating, outcome-driven journeys rather than reactive processes. This shift is happening now because the underlying technologies—AI, cloud and automation—have finally reached enterprise-grade maturity, allowing them to be orchestrated together for tangible business impact.

In a heavily regulated industry, how do you balance autonomy with explainability and regulatory comfort?

In banking, autonomy and regulation are not in conflict—they are designed to coexist. Autonomy does not remove humans from the equation; it elevates them. As systems become more autonomous, human roles shift from operational execution to governance, policy framing, ethics and exception management. Autonomous, agent-based systems operate within clearly defined guardrails and regulatory frameworks, ensuring autonomy remains controlled and accountable.

Explainability is built into the system through decision engines that operate within contextual and regulatory frameworks, ensuring every action is traceable, auditable and defensible. The goal is not automation for its own sake but intelligent orchestration, where systems can act in real time while oversight and accountability remain firmly institutional and human-led.

Legacy technology remains a challenge for many banks. Can autonomy coexist with decades-old core systems?

Autonomous banking does not require a big-bang core replacement. What it requires is architectural maturity and disciplined execution.
Historically, banks have modernised through approaches such as full migration, hollowing the core, or building greenfield platforms. AI now introduces a fourth lever: an intelligence layer that decouples decisioning from the legacy core while preserving it as the system of record. AI agents can orchestrate processes across existing systems using API wrapping, data virtualisation and automation. This allows banks to innovate without immediately replacing decades of infrastructure.

In practice, what we see is gradual core hollowing. Intelligence moves outward first—modernising capabilities incrementally, reducing dependence on rigid cores over time and creating optionality rather than risk.

Autonomous mortgages are often cited as a breakthrough use case. What changes do they bring beyond faster approvals?

Autonomous mortgages transform one of the most manual and friction-heavy processes in banking. Traditionally, document intake, validation and reconciliation consume the bulk of operational effort. AI agents can now interpret and analyse the variability in real-world documents—from payslips and tax returns to bank statements—rather than simply processing volume.
This turns risk management into a continuous, data-driven process instead of a series of checkpoints. Compliance becomes embedded directly within workflows rather than applied retrospectively.
Operationally, employees move away from data entry toward higher-value roles focused on oversight and exception management. The result is lower cost-to-serve, faster cycle times and more intelligent underwriting.

As AI systems make high-volume credit and operational decisions, where does accountability sit?

As AI takes on high-volume decision-making, the accountability model does not change—it becomes clearer. We operate on a Human + Machine model. Machines handle the predictable 90% at speed and scale, while humans govern the critical 10%—exceptions, ethics, judgment calls and regulatory challenges. AI systems operate under human-defined policies, guardrails and controls. When decisions are challenged by regulators or customers, accountability remains with the institution, just as it always has.
Autonomy does not dilute accountability—it formalises it. Humans govern, the institution remains responsible, and control never leaves the bank.

How are leading banks redefining roles as machines take over predictive and transactional intelligence?

Leading banks are moving people up the value chain. Employees are no longer operators stitching systems together—AI increasingly handles that work. Humans become supervisors, governors and strategic controllers of outcomes. In this Human + Machine model, agents execute at scale, while humans interpret policy, exercise complex judgment and handle exceptions.
Predictive and transactional intelligence sits with machines, while trust, governance and accountability remain with people.
The closest analogy is aviation. Autopilot manages most of the flight, but pilots remain responsible for the journey and the outcome. Banking autonomy works in a similar way.

With AI agents expected to handle most service interactions, how should banks approach governance?

Governance has to be engineered into the system from day one. We treat AI agents as regulated digital operators embedded directly into the architecture. Models undergo rigorous pre-deployment testing for fairness and are continuously monitored in production. Performance is tracked in real time, model drift is detected early, and retraining happens under strict governance.
Decision rights are aligned to risk through a tiered autonomy model. Low-risk interactions can run autonomously, higher-risk decisions require human oversight, and critical decisions remain human-led. Every decision is traceable, ownership is clear, and AI risk is governed at the board level.

Will end-to-end autonomy become table stakes for global banks?

End-to-end autonomy is likely to follow the trajectory of many transformative technologies—what begins as a differentiator eventually becomes a baseline capability. Banks that continue optimising isolated processes will struggle to compete with institutions that orchestrate intelligence across the entire value chain. Early movers will capture structural advantages in cost, speed and risk precision.
Ultimately, the real divide will not be technology adoption but operating model ambition. Institutions willing to redesign workflows and rethink banking journeys will lead the market, while those that simply digitise legacy processes risk falling behind.