Is Vertical GenAI in Banking solving $600 Billion question?

In recent months, the debate around the future of technology has intensified, centering on whether we're on the brink of a groundbreaking new era or merely witnessing the rise of another tech bubble. At the heart of this discussion is what David Cahn of Sequoia aptly dubs the "$600 billion question"—the substantial annual revenue gap that needs to be bridged to justify current levels of investment in AI. With Wall Street’s initial excitement over AI cooling and concerns about the return on these massive investments mounting, it’s crucial to explore whether generative AI (GenAI) is genuinely transforming industries or if it’s just another flash in the pan.

?GenAI: Beyond the Chatbot Revolution

Generative AI's capabilities extend well beyond the realm of RAG-based chatbots, and we’re witnessing the emergence of vertical AI companies that are harnessing these technologies to tackle real business challenges effectively. These companies are integrating Agentic workflows, chatbots, and machine learning to address complex problems in the banking sector at scale.

Large Language Models (LLMs) have become increasingly adept at data gathering and entry, proving particularly useful in loan origination and merchant risk screening processes. For example, Casca is an agent in the loan origination process by collecting customer KYC (Know Your Customer) information, handling emails, and inputting this data into back-end systems. Its adoption of GenAI on the sales side has shown a 300% increase in conversion rates, which is massive.??

Source: Casca.ai

The interesting thing to note is that the 300% increase in conversion rates is coming at the cost of salespeople, and hence the hypothesis that GenAI will provide the next level of automation is indeed true. It has improved sales conversion dramatically, which so far was only possible by deploying more people on CRM systems. So, indeed, GenAI is going to compete with average salespeople, marketers, etc. The chart below seems to be a more accurate depiction of how GenAI is making an impact.

Source:https://www.dhirubhai.net/pulse/ai-hype-600b-question-46t-opportunity-ashu-garg-cacqc/

Similarly, tools like Coris and Baselayer assist in screening for merchant risk by aggregating website data, NAICS codes, and KYB (Know Your Business) information.??

Source: Coris.ai

LLMs are especially well-suited for tasks that involve handling complex, unstructured data. They can summarize, reformat, and adapt this data to fit the requirements of various legacy systems or forms. Additionally, generative AI is revolutionizing customer communication. For instance, Klarna's customer support AI can handle the same volume of queries as 700 human agents, reducing the average query completion time from 11 minutes to 2 and improving customer satisfaction. Impressively, it can manage inquiries in over 35 languages.

While early chatbot implementations were often inadequate, recent advancements have led to substantial improvements in customer experience, particularly when these tools are well-packaged and focused on specific use cases. Generative AI is now instrumental in making critical decisions across financial services. Companies like Axle deploy AI agents to review sanctions alerts, generate suspicious activity reports (SARs), and manage transaction alerts and onboarding documentation, effectively handling much of the initial investigation and manual workload.??

Source: axleruns.com

These AI capabilities, once considered cutting-edge, are quickly becoming standard across the industry. Whether developed by in-house teams or integrated into existing risk platforms, their adoption is accelerating. Furthermore, transformers—another advanced AI technology—are being leveraged for data analysis. A notable example is Nubank's acquisition of Hyperplane, which uses foundation models trained on large internal datasets for underwriting, collections, and marketing purposes.

In summary, the integration of LLMs and generative AI in financial services is not just a trend but a transformative shift, improving efficiency, accuracy, and customer experience across a range of applications. Naturally, these are very specific use cases, and there is no way it's going to fill the $600 billion gap anytime soon, but this example shows that the hype around GenAI is real. Sooner or later, all companies will be using GenAI in a meaningful way.