By Alexandra Mousavizadeh, CEO and co-founder of Evident Insights
As AI continues to permeate all manner of day-to-day banking functions, from customer service to risk management, the need for banks to measure their AI ROI has become a Group-level priority.
However, quantifying the success of AI initiatives presents unique challenges, especially given the significant upfront costs associated with infrastructure, model development and operationalising. According to McKinsey, for every $1 invested in AI model development, an additional $3 must be spent on change management, highlighting the substantial commitment required.
So how are the big banks thinking about ROI, and what are the lessons for the wider financial services sector?
The banks leading the way on AI adoption are already defining robust measurement frameworks that assess the immediate impact of AI projects and adapt to evolving use cases.
Their goal is to develop centralised, repeatable capabilities that reduce AI project costs, make it easier to scale successful initiatives and encourage the sharing of AI best practices across the organisation.
By standardising their approach to AI project management and measurement, banks can ensure that each initiative is evaluated consistently, making it easier to compare outcomes and identify areas for improvement.
That said, part of the challenge with measuring AI ROI is the diversity of outcomes that different projects can produce. Each AI initiative is unique and can potentially affect multiple aspects of a bank’s operations. A comprehensive measurement framework is essential for capturing the full spectrum of an AI project’s impact. This framework should be regularly reviewed and refined to reflect the changing dynamics of AI use cases, such as shifts in user adoption and the natural deterioration of AI models over time.
Outcome metrics for AI projects can generally be grouped into five categories, encompassing both “hard” financial metrics and “soft” qualitative metrics. Here’s what the banks are focusing on:
1. Income uplift
One of the most direct ways to measure the success of an AI project is by assessing its ability to generate additional revenue. AI can enhance or create new revenue streams, providing banks with the competitive edge they need in a crowded and noisy global marketplace. For example, AI-driven customer segmentation and personalised marketing can lead to increased product uptake, while advanced analytics can identify new investment opportunities or optimise pricing strategies. By tracking income uplift, banks can quantify the financial return of each additional AI investment.
2. Efficiency gains
AI has the potential to significantly improve operational efficiency by automating routine tasks, optimising processes, and reducing the need for human intervention. This can lead to cost savings, faster processing times and higher overall productivity. For instance, AI-powered chatbots can handle a large volume of customer enquiries without the need for additional staff, while machine learning algorithms can streamline loan approval processes by quickly assessing risk and creditworthiness. Measuring efficiency gains allows banks to assess how effectively an AI project is reducing operational costs or improving service delivery.
3. Risk reduction
AI offers powerful tools for identifying, controlling and mitigating various types of risk. AI models can analyse vast amounts of data to detect fraud, predict market fluctuations or assess credit risk, thereby reducing a bank’s exposure to potential losses. By implementing AI-driven risk management strategies, banks can not only protect their assets but also uphold regulatory compliance and build greater resilience against external shocks. Measuring the impact of AI on risk reduction is essential for understanding its value in safeguarding institutional stability.
4. Customer satisfaction
AI can play a significant role in enhancing the customer experience of everyday banking services. AI technologies such as natural language processing and predictive analytics can personalise interactions, anticipate customer needs and provide faster, more accurate responses to queries. By tracking customer satisfaction metrics, such as Net Promoter Score (NPS) or customer retention rates, banks can gauge the effectiveness of their AI initiatives in building stronger relationships with their clients.
5. Staff satisfaction
Attracting and retaining top talent is a constant challenge for banks, and AI projects can contribute to more fulfilling work and a more productive working environment. By automating mundane tasks and providing tools that enhance productivity, AI can empower employees to focus on higher-value activities that require creativity and strategic thinking. Additionally, AI-driven insights can support better decision-making, further enhancing job satisfaction. Measuring staff satisfaction, through employee engagement surveys or turnover rates, can provide insights into how AI is impacting the workforce and whether it is contributing to a positive organisational culture.
Measure, measure, measure
The dynamic nature of AI projects necessitates continuous measurement throughout their lifecycle. It is essential that banks revisit and adjust their measurement frameworks to ensure they remain relevant as AI maturity accelerates. Conducting regular reviews of outcome metrics, coupled with ongoing refinement of AI strategies, can help banks maintain a clear understanding of their AI investments’ value and make informed decisions about future projects.
With AI here to stay, and adoption in some of the world’s biggest financial institutions occurring at an astonishing pace, banks that can establish centralised, repeatable capabilities and a robust measurement framework will be better equipped to navigate the complexities of AI and achieve sustained success in the AI-first industry landscape of tomorrow.
