The financial services landscape is currently undergoing its most significant structural shift since the advent of the smartphone, transitioning from a mobile-first philosophy to an "agentic-first" paradigm. At the FinovateEurope 2026 conference in London, industry leaders gathered to dissect the implications of this transition, with a particular focus on how artificial intelligence is moving beyond simple chatbots to become autonomous agents capable of managing complex, high-stakes financial workflows. Central to these discussions was Dimitri Masin, CEO and co-founder of Gradient Labs, who provided a critical assessment of the current state of banking operations and the technological requirements for the next decade of growth. As financial institutions grapple with the limitations of traditional automation, the emergence of agentic banking promises to resolve the long-standing tension between digital efficiency and the massive human overhead still required to manage back-office complexities.
The Shift from Mobile-First to Agent-First Banking
For the past fifteen years, the primary focus of fintech innovation has been the refinement of the user interface (UI) and user experience (UX). The rise of "challenger banks" in the 2010s demonstrated that superior app design and seamless onboarding could disrupt century-old incumbents. However, Masin argues that this "mobile-first" revolution only addressed the surface level of the banking experience. While the customer-facing front end became sleek and intuitive, the internal machinery of most banks remained tethered to manual processes and legacy systems.
The 2026 conference highlights a growing consensus that the "glass screen" era of banking is reaching a plateau. The next phase, agentic banking, suggests a future where the interface may become secondary or even invisible. In this model, AI agents do not merely provide information or answer FAQs; they possess the agency to execute tasks, make reasoned judgments, and navigate the "messy" middle-office processes that have traditionally required human intervention. This transition represents a move from passive tools to active digital employees, fundamentally altering the cost structure and operational capacity of modern financial institutions.
The Persistent Challenge of Customer Operations
Despite the success of the first wave of fintech, a significant operational bottleneck remains. Masin, drawing on his extensive background at Monzo—a pioneer in the UK challenger bank space—noted that even the most advanced digital banks still rely on "gigantic human organizations" to maintain their accounts. These organizations are often tasked with handling exceptions, investigating fraud alerts, managing complex disputes, and ensuring regulatory compliance—tasks that simple, rule-based automation has historically failed to master.
According to industry data presented at the summit, manual customer operations can account for as much as 30% to 50% of a bank’s total operating expenses. These costs are not just financial; they manifest as delays in service, inconsistent decision-making, and "bad customer experiences" when human staff are overwhelmed by volume. Masin contends that the "second half of the problem"—the back-end operational complexity—remains unsolved for most of the industry. Traditional automation, often referred to as Robotic Process Automation (RPA), is rigid and breaks when faced with the nuance of human language or the ambiguity of financial regulations.
Defining the Agentic Transformation
The core of the agentic transformation lies in the ability of AI to handle "judgment and nuance." Unlike the previous generation of AI, which was largely predictive or generative in a vacuum, agentic AI is characterized by its ability to use tools, access databases, and follow multi-step reasoning paths to achieve a specific goal.
At FinovateEurope 2026, Gradient Labs demonstrated how these agents could be embedded directly into a bank’s core systems. Founded in 2023, the London-based firm has focused on creating "autonomous workers" for financial services. Masin explained that the leap from "traditional automation" to "agentic automation" is the difference between a system that can only follow a fixed flowchart and a system that can understand a customer’s unique intent, cross-reference it with internal policy, and execute the necessary corrective actions. This includes workflows such as anti-money laundering (AML) checks, "Know Your Customer" (KYC) remediations, and complex billing inquiries—areas where a single mistake can lead to significant regulatory fines.
Supporting Data: The Economic Imperative of AI Integration
The urgency behind this shift is underscored by recent economic data. A 2025 report from McKinsey & Company estimated that generative and agentic AI could add between $200 billion and $400 billion in value to the global banking sector annually. This value is primarily driven by productivity gains in front-line and back-office operations.
Furthermore, data from the 2026 Finovate sentiment survey indicates that 68% of European banking executives now view "operational autonomy" as a higher priority than "customer acquisition." This reflects a shift in market conditions where capital is more expensive, and the path to profitability requires extreme operational efficiency rather than just rapid user growth. In the UK specifically, the Financial Conduct Authority (FCA) has noted a 15% year-over-year increase in the volume of complex customer disputes, a trend that is becoming unsustainable for banks relying solely on human labor.
The Strategic Dilemma: Build vs. Buy in the Age of AI
A recurring theme in Masin’s discourse is the "build vs. buy" decision that banks face when integrating AI. In the early days of mobile banking, many large incumbents attempted to build their own apps in-house, often with mixed results. The AI era presents an even more complex version of this dilemma.
Masin suggests that while banks must own their data and their customer relationships, the underlying orchestration of AI agents is a specialized discipline. Building a robust agentic layer requires expertise in large language model (LLM) fine-tuning, prompt engineering, and the creation of "guardrails" to ensure safety and compliance. For many banks, attempting to build these systems from scratch can lead to "innovation theater"—projects that look impressive in a lab but fail to scale in a production environment.
Gradient Labs’ approach advocates for a middle ground: embedding specialized, pre-trained AI agents into existing bank architectures. This allows banks to maintain control over their proprietary data while leveraging the rapid advancements in AI agent technology developed by specialized firms. The "buy" component focuses on the intelligence layer, while the "build" component focuses on the integration and the unique business logic of the specific institution.
Regulatory and Industry Responses
The move toward agentic banking is not without its detractors and hurdles. Regulators, particularly in the European Union under the framework of the AI Act, are closely monitoring how autonomous agents make decisions that affect consumer credit, account closures, and fraud investigations. The "black box" nature of some AI models remains a point of contention.
In response to these concerns, industry bodies at FinovateEurope 2026 emphasized the need for "explainable AI." Masin noted that for an agent to be useful in a banking context, it must be able to provide an audit trail of its reasoning. If an AI agent denies a dispute claim, it must be able to cite the specific policy and the evidence it used to reach that conclusion. This transparency is essential for maintaining trust with both regulators and customers.
Representatives from major European institutions, including BNP Paribas and HSBC, have expressed a cautious but optimistic outlook. The general sentiment is that while humans will remain "in the loop" for high-level oversight and the most sensitive cases, the day-to-day management of the millions of transactions that occur in a modern bank must inevitably be handed over to autonomous systems to remain competitive.
Broader Impact and Implications for the Future of Work
The long-term implications of agentic banking extend beyond the balance sheets of financial institutions. There is a profound impact on the labor market within the financial sector. As AI agents take on the "messy processes" previously reserved for humans, the role of the bank employee is evolving. We are seeing a shift from "transactional roles" to "exception-handling roles."
In this new environment, human staff are no longer needed to perform data entry or basic troubleshooting. Instead, they act as supervisors for fleets of AI agents, stepping in only when the AI encounters a scenario that falls outside its programmed parameters or requires high levels of empathy and ethical judgment. This "human-agent collaboration" model is expected to become the standard operating procedure for the global banking industry by 2030.
Furthermore, the debate over the user interface—visual vs. voice—continues to evolve. Masin’s vision of a "screen-free" future suggests that as AI agents become more capable, the need for a complex visual dashboard diminishes. A customer might simply tell their banking agent to "optimize my savings for a mortgage in three years," and the agent will handle the inter-account transfers, investment allocations, and budget adjustments without the user ever needing to open a traditional app.
Conclusion: Preparing for an AI-First Future
The insights shared by Dimitri Masin at FinovateEurope 2026 serve as a roadmap for the next generation of financial services. The industry is moving past the era of digital "wrappers" and into an era of deep operational autonomy. For banks, the challenge is no longer just about having a pretty app; it is about building the underlying intelligence to manage the staggering complexity of modern finance.
Gradient Labs and its peers are signaling that the "second half of the problem"—the human-heavy back office—is finally being addressed. As AI agents move from experimental novelties to core infrastructure, the banks that successfully navigate the transition to agentic operations will be the ones that define the future of the global economy. The era of agentic banking is no longer a distant prediction; it is an active transformation, rewriting the rules of efficiency, customer experience, and institutional growth in real-time.
