By all metrics, the financial services industry is embracing AI head-on. Major players like JP Morgan, Morgan Stanley, and Goldman Sachs use the technology to streamline employee workflow, build new modeling applications, and boost efficiency.
While financial services companies spent approximately $35 billion on AI implementation in 2023, investment is predicted to more than double to $97 billion by the year 2027 — the fastest growth rate of any primary industry, according to the International Monetary Fund.
Columbia Business School's Digital Future Initiative recently explored how these new AI technology innovations impact the financial services sector during its inaugural Digital Innovation Conference. Homing in on the theme of AI in Finance, the conference brought together CBS researchers and industry leaders to explore the innovations, challenges, and opportunities AI is unlocking across the industry.
Panelists with expertise in topics ranging from automation and banking to wealth management and fintech shared firsthand insights into technological breakthroughs, workplace adaptability, the dynamics of human capital and machine interaction, regulation, and the outlook for the coming years. Below are the big insights into AI's impact on the financial services industry, and how organizations that correctly embrace it are gaining an edge.
1. AI-augmented workforces are creating more value than ever in fintech
AI is creating more value than ever before in the world of fintech, according to leaders on the conference's "AI's Transformative Impact on Fintech" panel — but human workers are far from being replaced. They stressed that the use of AI is no longer optional, but instead an essential tool for staying competitive, meaning that companies must continuously adapt and improve their AI capabilities to keep pace with evolving market demands. While there is often resistance to adopting AI technologies across organizations, leaders can address this by clearly demonstrating the tangible benefits of AI tools, such as enhanced productivity.
"AI is nothing more than a tool. It's a screwdriver like any other. There's much hype around it because we're all fascinated that a machine can make a sentence up or paint a painting, but we're the inspiration for it," said panelist Sunil Madhu, founder and CEO of Instnt, a SaaS firm that helps businesses to verify, accept, and onboard new online customers.
Additionally, AI allows for streamlining software testing and quality assurance, leading to faster development cycles and better customer satisfaction overall. AI is also utilized to analyze vast amounts of data in real-time to identify and manage fraud risks. This capability allows companies to reshape loss curves and more effectively underwrite risks. Common fraud hurdles, such as identity verification, are also being overcome by AI, allowing companies to distinguish between legitimate users and bad actors. This is particularly important in a landscape where fraudsters quickly adopt new technologies.
Though AI is reshaping job roles in the industry, inciting the need for employees to adapt and learn continuously, the panelists noted that the relationship between technology and human workers is that of augmentation and not replacement.
Francesca Carlesi '05, entrepreneur and CEO of Revolut UK, noted that when it comes to embracing AI as a team, “you need to be careful not to stop at every shiny object that's out there.”
“There is a whole new wave of innovation, and you cannot afford not to be on it. But at the same time, you cannot just drop everything,” Carlesi said.
2. AI is fundamentally reshaping asset management, but humans are still key to gleaning strategic insights
According to industry experts on the "Qualitative vs. Quantitative AI" panel, asset management is increasingly being driven by AI technology. In the past, AI models were siloed and struggled to communicate with one another. That fact, along with lesser processing power, often led models to make errant predictions and miss insights. However, new advancements allow different models to interact and share information, crucial for improving decision-making. Improving data quality and organization will only further increase AI's effectiveness in the asset management space.
"We will see asset management as an AI business in five years. Letting a human meddle in the process would be like a commercial jet pilot turning off the autopilot and GPS, then going by dead reckoning instead, said Aaron Brown, Bloomberg columnist and former chief risk manager at AQR Capital Management, an investment management firm.
The panelists also noted that new AI developments are creating significant opportunities to generate returns in emerging markets and less liquid markets, such as venture capital.
Even with the rise of AI, humans will still play a crucial role, particularly in areas requiring nuanced understanding and supervision. According to the panelists, as AI takes over more routine tasks, future asset managers will need to focus on skills that complement AI capabilities, such as strategic thinking and the ability to interpret AI-generated insights.
3. Be prepared to fail when embracing AI
Business leaders should be prepared to accept failure along their AI journey since many AI projects will not yield immediate results. During a Fireside Chat, Mike Schuster, Head of AI Core at financial services firm Two Sigma, noted that "failing is fine" when learning how to work with AI.
"Exponential growth, by definition, will end at some point," Schuster said.
Schuster added that teamwork is crucial for success with AI due to its complexity and that investing in developing talent to fill institutional knowledge gaps is equally important. He noted that organizations should prioritize training and development in programming, statistics, and creative problem-solving to prepare their workforce for AI-related challenges.
While AI has allowed firms to make great strides in productivity, leaders should remain grounded and skeptical of the hype surrounding AI technologies, according to Schuster. He stressed that many predictions about AI capabilities may materialize differently than expected and the importance of focusing on practical applications rather than speculative trends. Large language models have significant limitations and struggle with reasoning and planning, according to Schuster.
4. The AI Edge in Banking
The banking industry is full of repetitive tasks – document processing, customer inquiries, and data extraction, to name a few. These tasks, while extremely necessary, are often tedious and time-consuming, making them ripe for automation via AI. That’s just one application for the technology in the banking sector, according to leaders in the banking industry who spoke during “The Future of AI in Banking” panel.
They noted that Generative AI can also assist banking professionals in gaining insights by allowing analysts to process thousands of documents in seconds. AI is also being utilized to personalize client experiences with tailored recommendations, which previously required extensive manual, human effort.
Utilizing AI in banking, like in other financial sector industries, is still dependent on high-quality data, according to the panelists. The future of a banking organization’s competitive advantage lies in leveraging proprietary, high-quality data to train their AI systems. While this mainly benefits larger firms, AI lowers the barriers to entry for startups and smaller firms, they noted.
Still, AI’s potential in banking is not without risk. Models often struggle with sudden, unprecedented changes, like financial crises. To that end, the panelists emphasized the need for human oversight and diversified research approaches, as well as the development of smaller, domain-specific models tailored to their needs.
See a photo gallery from the event here.