Have you ever tried reaching out to a bank’s customer service line, only to find yourself waiting on hold for several minutes – sometimes longer – just to speak with a representative? It’s a common and often frustrating experience for both retail and business clients. This lag in responsiveness doesn’t just impact customer satisfaction; it reflects broader operational inefficiencies that many banks are still working to address. As customer expectations continue to rise, especially around speed and convenience, the pressure is mounting for financial institutions to modernize – often by turning to AI-driven solutions.
In this article, we will examine the emergence of AI agents in the banking sector, including the challenges their adoption may pose. We will also explore some expert opinions for addressing them.
The Evolution of AI in the Banking Sector
In a recent podcast, Alexander Stanev, VP Consulting and Financial Services at Sirma, outlines the evolving phases of AI in the banking sector. While banks have been exploring AI for years, the emergence of Generative AI (GenAI) marked a significant shift – particularly in internal knowledge management and accessibility. This phase enabled institutions to test the new AI achievements internally before rolling them out to customers. Customer service did not lag behind – it also advanced with the use of chatbots, for example. However, over the past year, a new phase has begun to unfold with the rise of AI agents – intelligent systems capable of both understanding user intent and executing tasks autonomously.
Often referred to as the new frontier in AI development, AI agents are rapidly advancing in the financial services sector. No surprise, their growth is projected to rise from USD 490.2 million in 2024 to the mind-blowing USD 4,485.5 million by 2030. This results in a striking compound annual growth rate (CAGR) of 45.4% from 2025 to 2030. In Europe alone, the market size of AI agents in financial services is expected to reach USD 1,024.3 million, with a comparable CAGR of 43.7%.
No doubt, this pictures a disruptive development. Yet, it’s essential to understand what underpins this advancement.
The Distinctive Edge of AI Agents
Unlike GenAI, which responds to user prompts, AI agents are designed to be proactive. They are trained to perceive, learn and act with minimal human intervention based on set goals. As IBM notes, AI agents that employ large language models (LLMs) surpass the capabilities of LLM chatbots in their decision-making, problem-solving, and interaction with external environments. This makes them particularly suited to the dynamic context of financial services.
AI agents are increasingly seen as digital coworkers or assistants, each specializing in a specific area of an organization’s operations. As Alexander Stanev notes in the podcast, they function much like a “virtual workforce.” They may be applied seamlessly within various internal processes. Even a straightforward process – such as requesting holiday leave, for instance – can involve several AI agents working in coordination: one checks the remaining paid leave balance, another verifies the employee’s identity, a third initiates the document submission system and requests final approval, etc.
This seamless orchestration of task-specific agents illustrates the growing potential of AI to streamline complex, multi-step processes with minimal human intervention.
Possible Applications of AI Agents in the Banking Sector
Forbes identifies four most typical domains for the application of AI agents in the banking sector that will evolve in the coming years:
- Enhancing efficiency and accuracy, particularly with respect to enhancing operational efficiency in cases such as analyzing credit histories, transaction patterns, etc.;
- Personalization at scale, especially concerning customer interactions where they can offer tailored financial advice, investment recommendations, etc.;
- Fraud detection and prevention by identifying and flagging suspicious activity in real time;
- Future-proofing financial services, particularly within the digital services context.
A best-practice implementation approach, as highlighted by Alexander Stanev, is to introduce these technologies in phases. He further recommends starting with internal use cases – rather than customer-facing ones – supported by human oversight to ensure accurate execution and avoid unwanted risks.
Challenges for Implementation
While AI agents offer significant potential to modernize banking operations, their adoption is not without challenges. Research indicates several critical issues that must be carefully considered when planning to implement AI agents in banks. These include:
- Trust and Transparency. Due to the autonomy of AI decisions, trust becomes of key importance, especially regarding ensuring explainability of AI decisions, customer’s control balance, outcome responsibility, and value alignment.
- Data Rights and Privacy. Data ownership and usage require the establishment of consent models, algorithmic fairness, privacy and data portability.
- Talent and Organization. The workforce may bring additional difficulties, especially with a view to hybrid skills development, integrating AI into organizational structures, establishing a culture that embraces AI-driven change, and attracting specialized talent.
- Regulatory Evolution. Banking regulations need to evolve to address the unique challenges posed by AI, particularly in terms of supervision approaches, responsibility frameworks, cross-border consistency, and embracing innovation.
According to Alexander Stanev, the greatest challenge of AI adoption in the financial services sector is people’s over-expectations related to the general hype surrounding AI. While AI can deliver significant value, it does so primarily for organizations that are well-prepared to use it. In practical terms, this means that organizations should prioritize their data – to achieve maximum results, they need to invest effort in preparing their data to be accurate, precise, and technology-aligned. Ultimately, banks should think data first.
Conclusion
AI agents are set to transform the banking and financial services landscape, opening doors for both organizational development and enhanced customer engagement. Yet, their adoption comes with challenges—from inflated expectations and data readiness issues to concerns over trust, transparency, and talent development.
Partnering with seasoned technology providers like Sirma can help navigate these obstacles. It can also support financial organizations in defining clear automation goals, launching initial or complex projects, and guiding the seamless overall adoption of AI. If you are interested in discussing your specific needs, please do not hesitate to contact our financial services team.
