By Rinesh Patel, Global Head of Financial Services, Snowflake
Wall Street AI investments will face increased scrutiny
In 2025, financial institutions will face intense pressure to demonstrate tangible returns on their AI investments. The era of unchecked AI spending will come to a close, as boards and investors demand clear evidence of value. It will no longer be enough to say your organization is merely using AI to win the approval of company leadership. Instead, organizations must actually be driving value from their AI implementations, and leaders will face increased pressure to quantify their AI investments and the wider business impact. As a result, I expect to see a strategic shift towards targeted AI initiatives with measurable outcomes. Rather than throwing spaghetti at the wall to see what sticks, enterprises will have to make compelling business cases to justify proposed AI projects. This new focus on AI ROI will further separate industry leaders from laggards, potentially reshaping the competitive landscape in finance.
General-purpose large language models will be replaced with specialized, industry-specific models
The one-size-fits-all AI model is dead. Financial services companies need highly-specialized, cost-efficient models that are designed specifically for their unique needs and workflows. 2024 made it evident that when it comes to models, bigger is not always better. 2025 will be the year we see smaller, more specialized models go mainstream. Firms will strategically deploy domain-specific models — trained with highly relevant, proprietary data — for tasks ranging from regulatory compliance to risk assessment. This transition will demand a more nuanced approach to AI implementation, with organizations carefully balancing factors like speed, cost, and specificity to best meet their needs. Mastering this new AI ecosystem will become a critical differentiator in the industry.
Next year, AI will be more than just chatbots
Over the past 2 years, we have seen many financial services organizations exploring use cases akin to kids in a candy shop, with chatbots as the flavor of the month. These client-facing use cases – propelled by the prominence and explosive adoption of tools like ChatGPT – have led many to look at generative AI as the technology to transform their customer experience. These use cases tend to grab headlines and capture the full imagination and “coolness” of what AI can potentially achieve. However, there is a growing acceptance that chatbots are just one element of generative AI. And as organizations focus more on AI ROI, we will see a wave of enterprise AI implementations where efficiency and productivity gains are the primary focus. In fact, the most compelling use cases will be internal. One such example is removing the complexities and challenges around data management as teams rely more and more on AI to streamline their end-to-end data lifecycle – from data ingestion to data mining and analysis. Leveraging AI to build a strong data foundation will – over the long run – enable organizations to figure out their commercial strategy and use cases.
Regulators will crack down on the use of AI in finance to establish industry-wide standards
I suspect we’ll see a significant AI-related incident in finance trigger a wave of new regulations, accelerating oversight of AI in the industry. To date, financial institutions have yet to feel the effects of stringent AI regulations, largely because regulators can’t keep up with the fast-moving landscape. Eventually, this will change as regulators have no choice but to act. Financial institutions will need to rapidly develop robust governance frameworks and guardrails for their AI systems. This shift will create an increased focus on fostering responsible AI in fintech, and firms that fail to adapt quickly may face severe penalties, while those who lead in AI compliance will gain a significant competitive edge.
