Beyond automation: why the future of transaction banking depends on hiding complexity, not adding it

As AI reshapes the payments landscape, banks face a more fundamental challenge than adopting new technology. Success will depend on rethinking operating models, simplifying complexity and redefining the role they play in an increasingly agent-driven economy.

 

Recently, artificial intelligence has become the defining conversation in financial services. Yet amid the excitement surrounding generative AI, autonomous agents and ever-more sophisticated automation, one message is becoming increasingly clear: technology itself is no longer the biggest challenge.

 

For transaction banks, the real question is how to navigate a landscape that is becoming exponentially more complex. New payment schemes, evolving regulation, growing customer expectations and legacy infrastructures are converging at precisely the moment AI promises to transform the industry. The temptation is to see artificial intelligence as the solution to this complexity. In reality, it is only one part of a much broader transformation landscape.

 

At a recent roundtable discussion sponsored by Coconet, a striking consensus emerged between banking executives, technology leaders and payments specialists. The organisations that succeed over the next decade will not necessarily be those deploying the most advanced AI models. They will be those that become best at concealing complexity from customers while managing it more effectively behind the scenes.

 

Complexity has become the defining challenge

 

Transaction banking has always been a balancing act, but that balance is becoming increasingly difficult to maintain. Banks are expected to support an expanding array of domestic and cross-border payment schemes, comply with multiple regulatory frameworks, integrate with fintech ecosystems and deliver consumer-grade digital experiences, often while operating on technology platforms designed for a different era.

 

Customers, meanwhile, are largely indifferent to the complexity involved. They are not asking for artificial intelligence; they are asking for faster onboarding, more intuitive services, greater transparency and frictionless payments. They expect financial services to feel as seamless as every other digital experience in their lives.

 

This disconnect creates an important strategic challenge. Banks risk investing heavily in AI initiatives simply because the technology has become a board-level priority, rather than because it solves a genuine customer problem. As several participants observed, clients rarely request AI explicitly. They request outcomes.

 

The implication is significant. AI should not become another layer of complexity added to already fragmented operating models. Instead, it should enable organisations to simplify both internal processes and customer interactions.

 

Strong foundations matter more than sophisticated models

 

While the conversation inevitably turned towards the potential of agentic AI, participants repeatedly returned to a less glamorous subject: foundations.

 

For many organisations, AI is exposing weaknesses that have existed for years. Disconnected workflows, legacy tools, inconsistent data models and fragmented ownership structures become major obstacles when organisations attempt to scale intelligent automation. Rather than accelerating transformation, these limitations often force businesses back to first principles.

 

This is proving to be one of AI’s unexpected benefits. Preparing systems for intelligent automation requires organisations to revisit data quality, modernise architectures and integrate processes that have evolved independently over many years. In effect, AI is accelerating long-overdue operational change rather than replacing it.

 

That principle extends beyond technology. Operating models themselves require rethinking. Product, technology, strategy and operations can no longer evolve in isolation if organisations hope to respond quickly to shifting customer expectations and technological change.

 

However, breaking down organisational silos does not simply mean creating larger teams or new reporting structures. Effective collaboration depends on aligning incentives, accountability and decision-making while preserving clear ownership. Centralisation may improve visibility, but excessive centralisation can introduce new bottlenecks. The challenge is finding the right balance between enterprise-wide co-ordination and domain-level autonomy.

 

Trust remains the industry’s defining currency

 

Few issues generated more debate than trust. Although AI capabilities continue to improve rapidly, participants acknowledged that confidence in autonomous decision-making remains uneven across financial institutions. Human oversight continues to play an essential role, particularly in high-value, high-risk environments where regulatory accountability cannot simply be delegated to algorithms.

 

Interestingly however, the discussion suggested that the “human in the loop” standard should not be viewed as a permanent destination but as a transitional operating model. Today’s challenge is less about replacing human judgement, and more about redefining it. As AI assumes responsibility for increasingly routine analysis, human expertise shifts towards validation, exception handling and strategic oversight. Much as editors became more rather than less important during the digital publishing revolution, financial professionals may increasingly focus on reviewing and interpreting AI-generated recommendations instead of producing every decision themselves.

 

This transition creates its own challenges. If AI performs more of the routine work, how do organisations develop the next generation of experts? Experience has traditionally been accumulated through repetition. As automation removes many of those learning opportunities, banks will need new approaches to capability development and knowledge transfer.

 

From digital channels to digital agents

 

Perhaps the most thought-provoking theme concerned the future relationship between banks and their customers. For years, financial institutions have focused on improving digital channels. Increasingly, however, the conversation is shifting towards digital agents.

 

As personal AI assistants become capable of planning travel, managing finances and making purchasing decisions, the interaction model itself begins to change. Rather than customers logging into banking applications, autonomous agents may increasingly communicate directly with banking systems, executing payments, gathering information and managing financial workflows on behalf of individuals and businesses.

 

For corporate banking, the implications could be profound

 

Some organisations will continue to rely on banks as trusted advisors, integrating AI capabilities into broader financial relationships. Others, particularly sophisticated multinational corporates, may build their own AI ecosystems and expect banks to function primarily as secure execution partners within them. Neither outcome eliminates the role of the bank. Instead, it changes where value is created.

 

Future competitive advantage may depend less on owning the customer interface and more on providing trusted infrastructure, intelligent orchestration and secure connectivity within increasingly automated ecosystems.

 

Governing AI will require AI

 

Greater autonomy inevitably introduces greater risk. As intelligent agents begin interacting with one another at machine speed, traditional governance models become increasingly difficult to sustain. Fraud detection, cyber-security and operational resilience all become more complex when both legitimate and malicious activity can be generated autonomously.

 

This points towards another emerging reality: AI will increasingly be required to supervise AI. Real-time monitoring, anomaly detection and preventative controls will need to operate at a speed beyond human capability. Yet governance remains fundamentally a human responsibility. Organisations will therefore need clear accountability frameworks defining what autonomous systems may do, who authorises them and where responsibility ultimately resides. Technology alone cannot answer those questions.

 

The winners will simplify, not complicate

 

Financial services has entered another period of profound technological change, but unlike previous waves of innovation, AI is not simply introducing new capabilities. It is forcing institutions to reconsider how they organise themselves, how they interact with customers and where they create value.

 

The banks that thrive will not be those that deploy the greatest number of AI tools or build the most sophisticated autonomous agents. They will be those that recognise a more fundamental truth: customers care little about the complexity that exists behind the scenes. Their expectation is simple. Payments should be effortless. Financial services should feel intuitive. Trust should be implicit. The institutions that succeed will therefore be those that use AI not to make banking appear more technologically advanced, but to make complexity disappear altogether.

 

---

To learn more, please visit: www.coconet.de