How Fintechs Can Leverage Artificial Intelligence

There seems to be ongoing confusion between machine learning (ML) and artificial intelligence (AI). While often used interchangeably, the reality is that they’re not the same.

In other words, the relationship between ML and AI can be expressed in the following equation:

In my previous article, I covered ML and what to consider before investing in these solutions, so now I will focus on AI. Looking more closely at AI, there are two types. General AI is the representation of general human cognition in software. Applied AI, on the other hand, is the ability of a computer or computer-controlled device to perform a discreet, human task.

Opportunities for fintech lie within applied AI, and its potential is valued at $1 trillion in cost savings by 2030, according to a report by Autonomous Research LLP. Savings will be realized throughout the front, middle and back office. In the front office, expected savings are $490 billion, of which almost half will come from applying AI to security and administrative tasks at retail branches. In the middle office, expected savings are $350 billion, of which more than half will come from applying AI to compliance, KYC/AML and authentication. In the back office, expected savings are $200 billion, of which almost a quarter will come from applying AI to underwriting and collections. But AI is not just about savings. As product development and customer service costs decrease, fintechs will be able to launch new products and enter new markets faster. Simply put, as AI lowers the cost to serve customers, you can serve more customers. More importantly, as a result of ML, new learnings will lead to new data that drives smarter actions to increase revenue.

While the future of AI in fintech is bright, the current state is, well, murky.

While most fintechs are aware of AI and starting to experiment with it, very few have actually brought AI into production. Progress has been made on operational processes such as chatbots for the front office, KYC/AML for the middle office, and risk underwriting for the back office. However, most fintechs have yet to apply AI in higher-level functions such as biometrics, voice assistants and compliance using smart contracts. Therefore, fintechs have a long way to go before they’re able to systematically transform entire business processes.

Two key factors are holding fintechs back from realizing the full potential of AI: consumer sentiment and risk. In general, consumers are wary of AI in finance. A recent survey found that consumers are more comfortable undergoing robo-surgery than getting financial advice from a robo-advisor. And unlike Amazon, Spotify and Netflix, fintechs experience real financial and regulatory costs to making bad AI-driven decisions. While a poor product, song or movie recommendation may at worst lead to an irritated customer, a wrong credit decision leads to lower profitability and potential fines for fintechs and insufficient access to credit for customers.

While overcoming consumer sentiment will require education over time, there are practical steps that fintechs can take now to address risk.

1. Start with operational use cases in the middle and back office. Rather than focusing on digital transformation, focus on the areas where AI can incrementally improve daily decision making. This means starting with operational use cases in the middle and back office, such a KYC, verification and document review. Applying AI to these areas will enable employees to focus on more customer-facing, value-added activities. And since these tasks are typically repeated at high frequency, fintechs will be able to realize cost savings, productivity gains and conversion lift quickly.

2. Embrace redundancy and remediation. This means designing processes that allow for manual override and remediation of AI outcomes. For example, at my company, analytics and operations work together to detect and prevent fraud.