The business imperative
The transformative potential of artificial intelligence in banking continues to present both immense opportunity and significant challenges. According to McKinsey & Company’s projections, successful AI implementation could boost banks’ annual operating profits by between £200bn and £340bn. However, achieving these gains requires a fundamentally different approach to implementation than many institutions currently employ.
Financial institutions have frequently approached AI initiatives with ambitious executive-level driven goals but struggle with practical execution. Despite widespread adoption of integrated data strategies and the appointment of Chief Data Officers, many banks still grapple with fundamental data challenges. This is largely due to the vast amounts of data spread across disparate systems. Firms often struggle to consolidate information, lacking a Single Source of Truth (SSOT). Instead, they are faced with multiple datasets on the same subject, without a clear, unified source to guide their decisions.
Go deeper with GlobalData
A measured approach to transformation
Modernising banking systems demands a careful and methodical approach, much like steering an oil tanker. When building or developing data strategies and architectures, sudden changes can disrupt the entire operation, risking instability and failure. This reality calls for a phased, bite-sized approach to transformation, where institutions systematically rebuild their architecture one component at a time. This method allows clients to manage the process in a more controlled and effective manner, ensuring more reliable progress and better outcomes.
Innovation with control
An example of a practical application of AI is in loan and mortgage approvals which highlights the need for careful implementation. This remains a contentious area, as the models used to assess or approve financing must still undergo rigorous validation to comply with regulatory standards. This aligns with the EU AI Act’s classification of such systems as “high-risk”, mandating strict regulatory requirements, including robust risk management systems, data governance protocols and comprehensive technical documentation.
Enhancing efficiency
AI offers use cases across a vast scope, from testing frameworks to reconciling and cleaning data, and even generating code to help build and scale data platforms. These capabilities empower institutions to reduce operating costs by using modern architectures that provide greater flexibility and variability in their cost base.
From a commercial perspective, the transformation of customer service showcases the practical benefits of AI. AI-enabled contact centres highlight how technology can enhance customer interactions and improve productivity, while ensuring security. For example, such centres enable customer service representatives to identify and resolve customer inquiries with greater efficiency. Moreover, AI empowers agents by providing better access to relevant information, enabling them to address customer needs quicker.
Data custodianship
Better data quality, management, and governance lay the foundations for more effective AI implementation. Open banking supports this by allowing banks to securely share customer data with authorised third parties through APIs, fostering competition and innovation in financial services. By enforcing standardised formats, stronger security protocols, and real-time authentication, banks can improve data management and collaborate with fintechs to deliver smarter, more personalised financial solutions. This aligns with the Basel Committee on Banking Supervision’s (BCBS) emphasis on model outcome transparency, robust governance structures and institutional resilience.
Cultural change
Beyond the banking community’s dedication to continued operational governance, firms need to also lead through action. Successful AI implementation requires a cultural shift. One of the biggest challenges in any transformation is managing the people involved and navigating the complexities of change. Institutions need to address this with comprehensive training and governance frameworks. The EU AI Act emphasises this need, requiring human oversight and transparency in AI systems. The likelihood of success occurs when the business takes ownership and is actively engaged from the beginning. AI governance boards have become essential oversight mechanisms, tasked with evaluating AI use cases, understanding their implications, and determining which AI systems should be prioritised for investment.
Building customer trust
Customer trust is fundamental to the success of AI implementation and it requires a careful balance between innovation and security. Recent advancements highlight how effective security measures can significantly boost customer confidence. A seamless and efficient experience not only meets customer needs but also encourages loyalty, as customers are more likely to return to institutions that resolve their issues quickly and effectively. Banks, in turn, benefit from this customer-centric approach. By addressing problems promptly, they will minimise costs and avoid the dissatisfaction of a long-drawn-out resolution. By combining advanced AI capabilities with robust security processes, institutions can foster trust and improve efficiency ensuring higher customer satisfaction and brand loyalty.
The future of banking technology hinges on the development of resilient AI systems capable of adapting to ever-changing conditions. As financial institutions continue to expand their AI expertise, the challenge lies in striking the right balance between robust security controls and the delivery of innovative, customer-focused services. Achieving this balance not only safeguards sensitive data but also drives the continuous improvement of processes and technologies. By focusing on resilience, banks can steadily modernise their operations, phasing out outdated legacy systems, while evolving towards a more agile and future-ready state.
Vikas Krishan is Chief Digital Business Officer at Altimetrik
