From Pilots to Powerhouses: How Agentic AI and Tokenization Are Redefining Wholesale Banking in 2026

Introduction

Wholesale banking has always been a domain defined by scale, speed, and razor-thin margins. But as of March 2026, the sector finds itself at a decisive inflection point. What was once a cautious experimentation phase with emerging technologies has rapidly evolved into a full-scale transformation agenda. At the heart of this shift lie two interlocking forces: the rise of agentic artificial intelligence and the accelerating adoption of tokenization across financial assets and infrastructure.

Leading industry analyses from firms such as Accenture, World Economic Forum, Deloitte, and KPMG point to a clear conclusion: wholesale banks are no longer merely exploring innovation—they are operationalizing it at scale. The implications are profound, not just for efficiency and cost structures, but for how financial institutions fundamentally operate and compete.

The Rise of Agentic AI: From Assistance to Autonomy

Artificial intelligence in banking is not new. For years, institutions have deployed machine learning models to enhance fraud detection, credit scoring, and customer insights. What is new, however, is the transition from passive AI tools to active, autonomous systems—often referred to as “agentic AI.”

Agentic AI systems go beyond analysis. They act. These systems are capable of making decisions, initiating processes, and adapting dynamically without constant human intervention. In wholesale banking, this translates into real-world applications across trade settlement, compliance monitoring, liquidity management, and risk assessment.

The shift is subtle but significant. Instead of human operators managing workflows supported by AI, the model is increasingly inverted: AI agents manage workflows, with humans providing oversight. This inversion has the potential to redefine operational models entirely.

Consider post-trade processing, an area historically plagued by inefficiencies and reconciliation challenges. Agentic AI can autonomously identify discrepancies, initiate corrective actions, and ensure settlement finality in near real-time. Similarly, in compliance, AI agents can continuously monitor transactions, interpret regulatory requirements, and flag anomalies with a level of consistency that manual processes struggle to match.

The strategic implication is clear: banks that successfully deploy agentic AI will achieve not just incremental efficiency gains, but structural cost advantages and improved risk outcomes.

Tokenization: Rebuilding Financial Infrastructure from the Ground Up

Running parallel to the rise of agentic AI is the rapid advancement of tokenization. This involves converting real-world assets—such as bonds, deposits, and securities—into digital tokens that can be issued, traded, and settled on programmable platforms.

Tokenization is not merely a technological upgrade; it represents a rethinking of financial infrastructure. Traditional systems rely on multiple intermediaries, fragmented ledgers, and delayed settlement cycles. Tokenized systems, by contrast, enable atomic settlement, real-time ownership transfer, and embedded compliance through smart contracts.

Recent collaborations, such as those involving Quant and Murex, signal that institutional-grade tokenized infrastructure is moving from concept to production. These platforms are being designed to integrate seamlessly with existing capital markets systems while offering the benefits of distributed ledger technology.

For treasury operations, the implications are particularly compelling. Tokenized deposits and programmable cash enable more efficient liquidity management, reducing the need for intraday buffers and improving capital efficiency. In capital markets, tokenized securities can streamline issuance processes, lower operational costs, and enhance transparency.

Yet, the true power of tokenization emerges when combined with agentic AI. Autonomous agents operating on programmable financial infrastructure can execute transactions, manage portfolios, and enforce compliance rules in real time. This convergence is where the transformative potential becomes exponential.

Regulation: Innovation Under Watchful Eyes

As technology accelerates, regulators are working to ensure that innovation does not come at the expense of stability. Authorities such as the UK’s Financial Conduct Authority have made it clear that resilience, market integrity, and consumer protection remain paramount.

Recent regulatory priorities in wholesale markets emphasize the safe adoption of technologies like AI and distributed ledger systems. The message is not one of resistance, but of controlled evolution. Banks are expected to demonstrate that their use of advanced technologies does not introduce systemic risks or operational vulnerabilities.

This creates a delicate balancing act. On the one hand, institutions must innovate to remain competitive. On the other hand, they must ensure that new systems are robust, transparent, and auditable. For agentic AI, this raises questions around accountability and explainability. For tokenization, it brings focus to interoperability, legal frameworks, and settlement finality.

The regulatory stance is effectively shaping the pace and direction of transformation. Those institutions that can align innovation with regulatory expectations will be best positioned to lead.

The Operational Reality: Modernization Under Pressure

While the narrative around AI and tokenization is compelling, the reality within many wholesale banks is more complex. Legacy systems continue to dominate core operations, often limiting the speed and scale at which new technologies can be deployed.

Modernization is no longer optional. Aging infrastructure struggles to support the data requirements and processing demands of agentic AI. Similarly, tokenization initiatives require integration with existing systems that were never designed for programmable assets.

Data quality remains another critical challenge. Autonomous systems are only as effective as the data they consume. Inconsistent, incomplete, or siloed data can undermine the performance of AI models and introduce new risks.

Cyber resilience adds yet another layer of complexity. As banks digitize and interconnect their operations, the attack surface expands. Ensuring the security of tokenized assets and AI-driven processes becomes a strategic priority.

All of this is unfolding in a high-volume, low-margin environment. Wholesale banking does not afford the luxury of inefficiency. Every investment must ultimately translate into measurable improvements in cost, speed, or risk management.

Beyond the Headlines: Secondary Forces Shaping the Landscape

While agentic AI and tokenization dominate current discussions, other trends are quietly reshaping the wholesale banking ecosystem. The growth of private credit, for example, is challenging traditional lending models. Non-bank institutions are increasingly capturing market share, forcing banks to rethink their role in financing.

At the same time, margin compression continues to drive efficiency initiatives. Cost discipline is no longer a periodic exercise but a constant imperative. In this context, the appeal of autonomous systems and streamlined infrastructure becomes even more pronounced.

These secondary forces do not diminish the importance of AI and tokenization. Rather, they amplify it. They create the economic conditions that make transformation not just desirable, but necessary.

Conclusion

Wholesale banking in 2026 is defined by convergence. Agentic AI and tokenization are not isolated trends; they are mutually reinforcing forces that together are reshaping the industry’s foundations.

The transition from pilot projects to production-scale deployment marks a critical milestone. Banks are no longer asking whether these technologies will matter but how quickly they can implement them without compromising stability.

Success will depend on more than technology. It will require disciplined execution, robust governance, and a willingness to rethink long-established operating models. Institutions that can navigate this complexity will emerge with stronger, more resilient, and more efficient platforms.

Those who cannot may find themselves outpaced in a market that is rapidly redefining what it means to be competitive.

MY MUSINGS

There is something both exciting and unsettling about the current trajectory of wholesale banking. On one hand, the promise of agentic AI and tokenization is undeniable. The idea of autonomous systems operating on programmable financial rails feels like a natural evolution—perhaps even an inevitable one.

But I cannot help but wonder whether the industry is underestimating the second-order effects.

What happens when decision-making becomes increasingly abstracted from human oversight? Are we building systems we fully understand, or systems we merely trust? And what does accountability look like in a world where an AI agent executes a trade, manages liquidity, or flags a compliance breach?

Tokenization raises equally intriguing questions. If financial assets become fully programmable, do we risk over-engineering markets that are already complex? Or does programmability finally offer the transparency and efficiency the system has long needed?

Then there is the regulatory dimension. Regulators are rightly focused on stability, but can they keep pace with technologies that evolve exponentially rather than incrementally? And if they cannot, who ultimately bears the risk?

Perhaps the most pressing question is this: are banks truly transforming, or are they layering new technologies onto old assumptions?

I would be very interested to hear your thoughts. Are we witnessing a genuine paradigm shift or simply the latest cycle of financial innovation dressed in new language? And more importantly, who stands to gain—and who might be left behind?