Diverse applications of in banking
The real-world applications of AI within financial institutions are diverse and evolving rapidly. Alexandra Mousavizadeh, CEO of Evident, and her team analyze banks by ranking them based on their AI capabilities and observing how these capabilities impact their financial performance over time. Additionally, they examine the entire AI talent stack to understand how banks take use cases from the conceptual phase through to production. This includes studying the evolution of research labs within banks and their shifting roles in driving innovation.
Focusing more on high-value activities
When it comes to assessing impact, very few banks can measure significant revenue uplift attributable to AI. Instead, the real gains are being seen in efficiency improvements and workflow augmentation. Tasks, processes, and decision-making are being optimized to a degree that allows employees to focus more on higher-value activities. Alexandra is also beginning to see task replacements and outright removal of repetitive, mundane functions. While direct revenue generation remains elusive for most institutions, these efficiency gains pave the way for long-term competitiveness and scalability.
"Banks are now seeing their talent pool not only embrace AI but demand an environment that integrates AI strategies. Without such a strategy, many banks struggle with talent retention."
Alexandra Mousavizadeh, CEO, Evident
Better data, governance structures, and talent acquisition
Initially, there were substantial hurdles involving internal trust—how to ensure employees see AI as an enabler that enhances their abilities, rather than as a threat to their jobs. However, banks are now seeing their talent pool not only embrace AI but demand an environment that integrates AI strategies. Without such a strategy, many banks struggle with talent retention. One of the most striking changes has been in the focus on “Training & Development.” Banks are doubling down on educational initiatives, hackathons, and events that both showcase AI’s potential and build excitement internally. A few banks, such as JP Morgan and DBS Bank, have begun to disclose their tangible ROI from AI integrations, which could inspire others to follow.
The obstacles to AI adoption are often tied to the time it takes to bring solutions into production. When evaluating challenges, it’s clear there are three primary barriers: better data, governance structures, and talent acquisition or readiness. These hurdles are shared across financial institutions and directly impact the speed and depth of AI’s scalability. Despite these barriers, many banks are steadily working to refine their approaches, emphasizing a balance between immediate deliverables and long-term integration capabilities.
Banking of the future
Banks are deeply invested in envisioning the “bank of the future,” with many exploring the concept of an AI-first institution. Agentic frameworks and agents are at the heart of this vision. These AI agents are domain-specific, personalized, and designed to automate or enhance specific processes. For example, within corporate markets, agents are already being widely deployed.
"The next wave of innovation comes from agent collaboration—where agents assigned to different processes interact to deliver more advanced solutions, effectively forming an “AI workforce."
Alexandra Mousavizadeh, CEO, Evident
The next wave of innovation comes from agent collaboration—where agents assigned to different processes interact to deliver more advanced solutions, effectively forming an “AI workforce.” However, many banks risk falling behind if they fail to formalize an AI strategy, especially as the gap between early adopters and laggards widens. While there’s an argument for the “first follower” approach, banks that adopt this mindset need to accelerate their efforts rapidly to avoid losing market share.
The underestimated aspects for growth
The banks most aggressively pursuing AI adoption demonstrate a nuanced understanding of what is required to accelerate growth—whether through talent development, data infrastructure investment, or innovative governance models. In the short term, the potential of AI is often overestimated, but in the long term, it is largely underestimated.
One particularly underrated area is the role of quantum computing, which is on the immediate horizon. Evident AI is set to release a Quantum index shortly, highlighting how quantum computing will drive innovation by enabling entirely new capabilities—such as decrypting transactions instantaneously. Quantum computing could alter how we think about security, efficiency, and computational power across the industry.
Flexible, cost-effective, and smaller-scale modeling
There’s a clear gravitation toward more flexible, cost-effective, and smaller-scale open-source AI models. These models offer advantages in terms of adaptability and affordability compared to commercial counterparts. However, regional differences in adoption trends are emerging. European banks, for example, often rely on a decentralized, bottom-up model for AI adoption, mirroring their organizational structures. In contrast, North American banks tend to prefer centralized AI strategies, which have allowed them to advance more rapidly. However, North American institutions are also mindful of the bottlenecks that over-centralization can create and are actively working to strike a balance. Interestingly, European banks are now looking to emulate North American centralization models as they seek to scale AI innovation more effectively.
"Financial institutions could use quantum computing to optimize portfolios in ways that are currently unattainable with classical computing methods."
Alexandra Mousavizadeh, CEO, Evident
Quantum computing in financial portfolio optimization
The most promising quantum-related opportunities lie in portfolio optimization. Financial institutions could use quantum computing to optimize portfolios in ways that are currently unattainable with classical computing methods. This would allow for more precise risk management, enhanced diversification, and better allocation of capital, opening the door to new levels of efficiency and profitability.
Challenges of GPU capacity and governance
The GPU requirements of an agentic framework remain an unknown frontier, raising questions about how much investment is truly necessary to support these initiatives. Energy consumption—a critical aspect of GPU usage—deserves far more attention than it’s currently receiving.
On the governance front, there’s a sense of uncertainty, with institutions feeling their way through regulatory frameworks. Close collaboration between banks and regulators will be essential in developing governance models that balance innovation with compliance. This remains a significant hurdle for widespread AI adoption.
