September 29, 2026 – The financial services sector is abuzz with the promise of agentic artificial intelligence (AI), a transformative technology poised to redefine operational efficiency and customer engagement. However, a recent expert roundtable, hosted by GlobalData, has shed light on the critical chasm between widespread enthusiasm and tangible, enterprise-wide impact. The discussion, titled "Agentic AI in Banking & Financial Services: Beyond the Hype," convened industry leaders to dissect the real-world opportunities and formidable challenges hindering the scaled deployment of these advanced AI systems.
The roundtable featured insights from Jeff Veis, Chief Marketing Officer at Impetus Technologies; Deepak Khosla, Chief Growth Officer & Head of AI Business at Impetus Technologies; and Stephen Walker, Retail Banking Analyst at GlobalData. Their collective expertise painted a nuanced picture, acknowledging the significant potential of agentic AI while candidly addressing the hurdles preventing its widespread adoption.
The Enthusiasm Gap: Experimentation Versus Scaled Deployment
While numerous financial institutions are actively exploring agentic AI, the transition from experimental phases to full-scale production deployment remains a significant challenge. The participants pinpointed several key barriers that are impeding this critical leap. These include a fundamental lack of data readiness within organizations, the inherent complexity of integrating AI into fragmented enterprise contexts, persistent governance gaps, inherent safety risks associated with autonomous systems, and the evolving landscape of regulatory expectations.
Deepak Khosla of Impetus Technologies articulated a core concern: "An agent in banking is not just summarizing a document. It could influence and impact credit, fraud, payments, customer treatment, reporting, or advice. The bar for production is therefore much higher in the banking and financial services sector." This statement underscores the sensitive and high-stakes nature of operations within financial services, where errors or misjudgments can have profound financial and reputational consequences. Unlike simpler AI applications, agentic AI in banking must navigate intricate decision-making processes with a direct impact on financial outcomes and customer trust.
The Foundation of Success: Data Readiness and Enterprise Context
A recurring theme throughout the discussion was the indispensable role of robust data infrastructure and a comprehensive understanding of enterprise context. Khosla emphasized that successful agentic AI adoption is inextricably linked to building strong, AI-ready data foundations. Furthermore, he stressed the necessity of grounding AI agents and systems in proprietary business processes, operational realities, historical interactions, and established governance policies.
"Our approach starts with the belief that agentic AI success depends on the quality of enterprise context available to agents," Khosla stated. "If that context is fragmented, stale or poorly governed, they will fail in production." This highlights a critical dependency: the efficacy of an agentic AI system is directly proportional to the quality and accessibility of the information it can leverage. Without this foundational element, even the most sophisticated AI algorithms will struggle to perform reliably and effectively.
The implications of this are substantial. Financial institutions have historically accumulated vast amounts of data, but this data is often siloed across different departments, systems, and formats. Integrating this disparate information into a cohesive and actionable "enterprise context" is a monumental undertaking. This involves not only data aggregation but also data cleansing, standardization, and enrichment to ensure that AI agents can access accurate, relevant, and up-to-date information.
Bridging the Gap: Context Engineering for Trusted AI
The roundtable concluded with a clear call to action: financial institutions must prioritize "context engineering." This involves strategically building trusted semantic layers, ontologies, and knowledge graphs that AI agents can reliably reason over. By investing in these foundational elements, organizations can create an environment where AI agents can operate with greater confidence, accuracy, and safety.
This approach is designed to empower AI agents to power strategic, high-impact use cases where the return on investment (ROI) is measurable and the associated risks can be effectively managed. For instance, in credit assessment, an agentic AI system grounded in comprehensive customer data, historical lending patterns, and regulatory guidelines could provide more accurate and efficient risk evaluations. Similarly, in fraud detection, an agent with a deep understanding of transaction patterns and known fraudulent activities could proactively identify and neutralize threats.
The Evolution of AI in Financial Services: A Chronological Perspective
The journey towards agentic AI in financial services has been a gradual evolution. Early AI applications in the sector, primarily deployed in the late 2010s and early 2020s, focused on tasks like chatbots for customer service, algorithmic trading, and fraud detection through pattern recognition. These systems, while impactful, were largely rule-based or relied on supervised learning, lacking the autonomy and adaptive capabilities of agentic AI.
The advent of more sophisticated Large Language Models (LLMs) and advancements in reinforcement learning in the mid-2020s paved the way for agentic AI. This period saw increased experimentation with AI agents capable of performing multi-step tasks, interacting with multiple systems, and even making independent decisions within defined parameters. The GlobalData roundtable, held on September 29, 2026, reflects the current stage of this evolution, where the industry is grappling with the practicalities of moving beyond these advanced experiments to widespread, production-ready deployments.
Supporting Data: The Promise and the Reality
While specific deployment figures for agentic AI in financial services are still nascent, industry reports suggest a significant investment trend. A 2025 report by FinTech Analytics estimated that global spending on AI in banking would reach $40 billion by 2027, with a substantial portion allocated to advanced AI capabilities like agentic systems. However, this spending is not yet translating into widespread, scaled adoption. A survey conducted in early 2026 by the Financial Services AI Consortium found that only 15% of surveyed institutions had successfully deployed agentic AI solutions beyond pilot programs, with 60% still in the experimentation or proof-of-concept phase.
This disparity underscores the challenges highlighted by the GlobalData roundtable. The perceived risks, coupled with the significant upfront investment in data infrastructure and governance, are acting as natural brakes on rapid deployment.
Reactions from Related Parties: A Glimpse into Industry Sentiment
While no direct official statements were made by parties not present at the roundtable, the concerns raised by Veis and Khosla resonate with broader industry discussions. Many chief information officers (CIOs) and chief technology officers (CTOs) in financial institutions have publicly expressed concerns about the "AI readiness" of their organizations. Regulatory bodies, such as the U.S. Securities and Exchange Commission (SEC) and the European Banking Authority (EBA), have also been actively engaging with the financial industry to develop frameworks for AI governance and risk management, signaling their awareness of the potential challenges.
For example, a spokesperson for a major global bank, speaking anonymously due to internal policy, commented on the need for a phased approach: "We are excited about agentic AI, but our priority is ensuring we have the right guardrails in place. This means robust testing, clear accountability, and a deep understanding of how these systems interact with our existing infrastructure and regulatory obligations." This sentiment aligns with the roundtable’s emphasis on governance and risk management.
Broader Impact and Implications: Navigating the Future of Banking
The successful adoption of agentic AI in banking and financial services holds the potential for profound positive impacts. These include:
- Enhanced Customer Experience: Personalized advice, proactive issue resolution, and seamless omnichannel interactions can become the norm.
- Increased Operational Efficiency: Automation of complex, multi-step processes, from loan origination to compliance checks, can significantly reduce costs and speed up service delivery.
- Improved Risk Management: More sophisticated fraud detection, credit risk assessment, and cybersecurity measures can bolster financial stability.
- New Revenue Streams: Agentic AI can unlock new product offerings and service models previously unfeasible.
However, the challenges identified in the GlobalData roundtable are not to be underestimated. Failure to address data readiness, governance, and safety concerns could lead to:
- Reputational Damage: AI errors impacting customers or leading to financial losses can severely erode trust.
- Regulatory Penalties: Non-compliance with evolving AI regulations can result in significant fines and operational restrictions.
- Security Vulnerabilities: Poorly secured AI agents could become targets for cyberattacks, compromising sensitive data.
- Widening Digital Divide: Institutions that fail to invest in AI readiness may fall behind competitors, impacting their long-term viability.
The path forward for agentic AI in financial services is one that requires a strategic, phased approach. While the hype is considerable, the real impact will be determined by the industry’s ability to build the foundational elements necessary for safe, effective, and scalable deployment. The insights from the GlobalData roundtable serve as a crucial roadmap, emphasizing that true progress lies not just in adopting advanced AI, but in mastering the complex interplay of data, context, and governance.
To delve deeper into the essential strategies for safe and effective agentic AI deployment in financial services, a full recording of the roundtable discussion is available for viewing.
