Our recent discussion on AI, led by Ari Gross, our CEO and Co-Founder, and Bob Noble, our Chief Product and Technology Office, covers a wide range of topics. This educational session provides listeners with a thorough explanation around the origins of AI, its impact on lending, the training process, the potential ROI and cost savings, and what lies ahead for our industry in 2024 and beyond.
Expanding the AI Use Case Landscape in Lending
While credit scoring, loan approval, fraud detection, and collection management are now well-established applications, AI is rapidly unlocking new possibilities in lending. Lenders increasingly use AI for:
- Dynamic Loan Pricing: AI models analyze market data, borrower risk profiles, and competitive rates to set personalized interest rates, optimizing both profitability and borrower appeal.
- Portfolio Risk Monitoring: Machine learning continuously tracks loan portfolios, detecting early warning signals of potential defaults or concentration risks, enabling proactive intervention.
- Hyper-Personalized Loan Products: By mining customer data and behavioral patterns, AI helps lenders design tailored loan offers and repayment plans for specific segments, such as gig workers or first-time borrowers.
- Alternative Data Credit Assessment: For underbanked or thin-file applicants, AI leverages non-traditional data sources, like utility payments or mobile transactions, to assess creditworthiness, expanding access to credit.
- Automated Compliance Monitoring: AI tools can scan lending activities in real time to flag compliance issues, reducing regulatory risk and manual workload.
These emerging use cases demonstrate that AI’s role in lending is far broader than traditional automation, supporting smarter, more inclusive, and adaptive lending strategies.
How AI Is Reshaping Relationship Lending
While AI is revolutionizing many aspects of lending, its impact on relationship-based lending deserves special attention. Traditionally, relationship lending relies on close, long-term interactions between lenders and borrowers, enabling banks to gather information, such as trust, reputation, and nuanced understanding of a client’s business, that goes beyond what’s found in financial statements.
Rather than replacing relationship lending, AI can complement it: relationship managers may use AI-driven insights to better understand clients’ needs, anticipate challenges, and offer more personalized solutions. For example, AI can flag early warning signs in a borrower’s financial behavior, allowing the relationship manager to proactively reach out and support the client.
However, the integration of AI also raises important questions. As lending decisions become more automated, there’s a risk that the personal touch and trust central to relationship lending could diminish. Borrowers may worry that their unique circumstances are overlooked by algorithms, while lenders might find it harder to differentiate through service and long-term partnerships.
The balance between AI and human expertise is especially critical during periods of economic stress. Research suggests that while AI can help standardize and streamline credit assessments, it may not fully replicate the flexibility and protection that relationship lending offers clients in times of crisis. Therefore, leading institutions are exploring hybrid models, leveraging AI for data-driven insights while keeping relationship managers at the core of client interactions.
The current perception of AI in lending is often met with skepticism. People question whether AI actually works, how it will change their organizations, workforce, and customer experience. Concerns often arise around the cost, complexity, and overhead of running an AI solution.
In our on-demand recording, Ari unpacks the origins of AI, highlighting the work of Alan Turing, the term “AI” that was coined by John McCarthy, and the development of expert systems in the 1980s, with rules and early learning, further pushed the boundaries of AI.
What is true artificial intelligence? Today, AI is not about “thinking” but rather about using big data effectively to continuously learn and improve over time. Chat GPT, a Generative AI system, is particularly useful in interacting with its users, facilitating conversations and explaining derived solutions.
The scalability of AI-powered operations is a significant advantage, especially as the market recovers in 2024. Lenders and insurers, aided by AI, can become true business experts, freeing up valuable time previously spent on manual tasks like data correction and analysis. This shift will drastically change the customer experience, allowing for near real-time document collection and data perfection. Intelligent decision-making will be driven by automated income reviews, fraud detection, and scoring the quality of information, empowering lending experts to reach automated loan decisioning without extensive manual labor.
Practical Examples of AI Use Cases in Action
- A lender uses AI to instantly flag a loan application for secondary review after detecting an unusual pattern in the applicant’s transaction history, preventing potential fraud before funds are disbursed.
- AI-driven systems monitor a portfolio of small business loans, alerting risk managers when repayment behaviors suggest early signs of distress, allowing for timely outreach.
- For a new-to-credit applicant, AI incorporates alternative data, such as consistent utility bill payments, to generate a robust credit profile, enabling approval where traditional methods fall short.
