Navigating compliance can be a challenging journey for a bank’s front office. Relationship managers often voice frustrations about the escalating demand from compliance tasks such as Know Your Customer (KYC), transaction monitoring or anti-fraud measures. These tasks not only heighten workloads but can strain relationships with clients. Fortunately, it doesn’t have to be this way. By harnessing artificial intelligence (AI), leading banks have shown that compliance and control protocols can be seamlessly integrated into front office’s daily operations and products and a harmonious and productive partnership between the front office and the compliance function is indeed achievable.
Front office’s struggle
The banking industry is frequently targeted by illicit activities, including money laundering and fraud. Criminals are constantly seeking to exploit vulnerabilities. Over the last decade, regulatory bodies worldwide have imposed increasing compliance standards, prompting banks to bolster their defenses against financial crimes.
For banks undergoing compliance transformation, this has posed substantial challenges. Legacy organizational structures and processes often struggle to integrate new controls, creating friction between 1st and 2nd Line of Defense. Outdated data and IT infrastructure hinder the digitization of new processes, leading to burdensome manual tasks like information gathering and contextualization. Moreover, a hesitant adoption of risk culture leaves front office staff uncertain about the practical implications of stricter compliance standards for their day-to-day operations. This conflict can make compliance seem like a barrier to business and create ongoing tension.
AI as the peace maker
For several years, banks have been utilizing AI in compliance risk management, with well-established use cases primarily in predictive AI, particularly for monitoring and screening. Examples include transaction monitoring scenario calibration to reduce alert volumes, machine learning models to automatically close false positive alerts, and entity resolution and network analysis to support investigation of complex cases. While these predictive AI applications made great contributions to enhancing the effectiveness and efficiency of compliance processes, their impact on front office is mostly indirect, as they are largely focused on back-office activities and traditionally require strong analytical and technical skills, which are not commonly found among front office relationship managers.
The emergence of generative AI (GenAI) and, more recently, agentic AI has proven to be transformative. Leading banks have recognized their potential and are actively integrating these technologies into front office compliance processes, as illustrated in Figure 1.
Selected successful examples of GenAI and agentic AI use cases for front office compliance tasks include:
- Data extraction and analysis: Instead of requesting clients to provide information and documents as part of the KYC or periodic review process, relationship managers can now utilize data extraction tools to automatically gather documents from public and/or third-party sources, retrieving relevant data fields. This approach reduces potential friction and back-and-forth during the data collection process and ensures that clients’ KYC files are consistently up to date.
- Generative business intelligence: Relationship managers can use dynamic reporting dashboards to derive insights by directly querying raw data on clients’ transactions and behavior. This capability eliminates the need for time-consuming data analysis and reliance on metrics pre-defined by the compliance function, allowing front office staff to quickly obtain answers to specific questions of interest.
- Narrative generation and documentation: Documenting commercial decisions and customer information, such as ownership structures and sources of wealth, can be cumbersome and repetitive. GenAI tools are ideally suited to assist front-office staff in creating these narratives during client onboarding, requiring only minimal human validation for accuracy and completeness.
- E2E case handling: Agentic AI is transforming complex KYC processes through end-to-end automation, where multiple specialized AI agents are utilized for individual process steps. This includes agents such as the document processing agent, screening agent, validation agent, and orchestration agent, each handling specific tasks within the KYC process. While the overall system is supervised by humans, the deployment of agentic AI significantly reduces the requirement for human involvement in these tasks by over 90%.
Banks are at varying stages in their journey of implementing GenAI and agentic AI solutions for front office compliance tasks. Many institutions have already embarked on deploying these technologies, experiencing substantial benefits such as streamlined operations and enhanced efficiency, enabling relationship managers to prioritize their focus on sales activities. These efforts are supported by strategies like compliance by design, operational excellence, and fostering a strong risk culture, with AI serving as the central element driving the transformation of compliance from a perceived hindrance into a vital, strategic asset for the front office.
Effective implementation
To effectively harness GenAI and agentic AI, banks require more than just suitable technical solutions. They must reinvent the target operating model (TOM) for compliance risk management that supports tailored capabilities. Key components include:
- Robust data foundation: Establishing clear data lineage and ensuring ample data availability and quality to generate accurate insights for informed decision-making.
- AI governance incl. risk management framework: Enhancing existing model risk management frameworks to encompass GenAI and agentic AI models, including dedicated validation methodologies and AI risk governance structures.
- Skills and awareness enhancement: While traditional analytical and coding skills are less critical for applying GenAI and agentic AI, front office staff must develop new skills like prompt engineering and bolster awareness of AI-related risks.
There are broadly two paths banks can take on their AI-driven compliance transformation journeys. The first one is strategy-based and focused on the target state design. On this path, banks typically start with a review of the TOM and definition of their ambitions. Based on these, they can start building central capabilities in areas such as data management or AI risk governance to lay the foundation for widespread AI application deployment.
Alternatively, banks can begin with specific use cases that promise significant improvements in efficiency and effectiveness. For front office compliance tasks, typical high-impact use cases may include KYC and client onboarding, where GenAI and agentic AI can significantly accelerate the due diligence process. This approach appeals to banks focusing on fast AI adoption and achieving rapid gains. However, the challenge is to ensure that the transition does not get bogged down in narrow use cases. To avoid that possibility, decision makers need to quickly expand from individual use cases to the more strategic TOM.
Summary
Building a strong relationship between the front office and compliance teams is both achievable and crucial for banks facing complex regulatory challenges. With advances in AI, especially GenAI and agentic AI, banks can smoothly embed compliance into their operating models and revolutionize the way front-office processes are managed. This approach leads to greater security and control, while also boosting teamwork and efficiency—setting the foundation for lasting success.