Many banks and financial firms are investing in AI and seeing positive return from applying AI throughout their operations. AI-based systems are helping to make more informed, safer and profitable decisions. However, with any technology that’s used in a heavily regulated industry there are challenges and pushbacks to adoption.
Bringing AI innovation to financial services
Kumar Srivastava, VP of Product and Strategy of BNY Mellon recently shared with the AI Today podcast insights into AI adoption at the bank. BNY Mellon has a Silicon Valley based Innovation Center that aims to help bring AI innovations to the bank. Kumar has been working in the field of AI for quite a while, starting his career at Microsoft where he was working email spam detection using machine learning and AI for Hotmail. This has given him the ability to appreciate and operate AI on the scale that millions of people need and rely upon every day as AI technology has dramatically expanded its reach across the industry. (Disclosure: I’m a co-host of the AI Today podcast)
In the podcast, Kumar addresses how AI is being developed and used in different ways within the fintech industry. The basic goal of the financial service industry is to help customers make better decisions about their finances and increase the amount of capital they have through investments. The process starts with a specific goal in mind, and users start by looking for an adviser that will help them best meet their financial goal. From here the customer and the advisor meet to decide the best types of investments to be made and then implement that strategy frequently meeting to review and tweak the strategy as needed. Much of this process is still human to human.
The whole industry is based on the idea of making good decisions. If you have information or an edge that you otherwise wouldn’t have, you can make better investment decisions. Firms are always looking for ways to find and implement an advantage to try and make better investment choices than someone else who might not have had that information. To gain an advantage, It comes down to looking for the signals that others might miss and making the best use of those signals. AI can synthesize all of this information and establish patterns. In fact, pattern and anomaly detection is one of the seven patterns of AI. There is a lot of potential in AI to collect all of this information from different sources and aggregate it in a way that customers will be able to make sound investment choices.
Integrating AI technologies to see true value
AI can be used to not only improve the investment strategies described earlier, but also give financial advisors more information to develop an investment strategy for their clients. AI is now assisting in the creation of investment strategies. Some firms are using AI for robo-advising creating digital platforms that provide automated, algorithm-driven financial planning services with little to no human supervision. Additionally, AI is helping create more complete profiles of clients for investment managers so they can better serve each client.
Additionally, AI is being used to detect and proactively manage fraudulent activities. AI technologies are being used to sort transactions into three groups: good or normal activities, bad activities, and mixed or suspicious activities. This ranking system helps to instantly flag certain transactions for human decision makers to take a closer look. These systems help to maintain focus on the more high stakes instances of fraud. AI can look at these results and use them to further improve and limit the instances of risky behavior or identify potential areas worth investigating more deeply.
Challenges to broad AI adoption in Fintech
Just like with any technology, there are some challenges for fintech companies when it comes to adapting AI. The primary problem facing the industry is the lack of skilled talent who can keep up with the fast pace of change and those who can understand both the financial services industry and well as technology. In other words, it comes down to finding people who understand both sides about the technology and the industries that they are working in and anticipating the needs that will arise from both sides. Another issue facing more widespread adoption is the culture around integrating AI in business. Due to either budget constraints, resource limitations, or a risk averse culture, companies want to start small and attacks problems that are really inconsequential. This does a huge disservice because it limits companies from truly building and developing valuable AI solutions and as a result they may get left behind.
There are still some major gaps are in the expectations or applications of AI between senior level management and technology implementers. The biggest issue is the lack of “accountability” that you get when using AI. After all, AI is probabilistic not deterministic so you will never get 100% certainly with machine learning results. Additionally, many algorithms are “black boxes” and don’t provide explainability for the decisions and outputs it provides. For example, when a human employee makes a decision or does something and it is questioned, the person in charge is able to ask them why and then the human employee can give a reason or a rationale behind the decision they made, and this is something that is hard to get out of an AI. Industries like finance that rely on accountability, adopting AI can present challenges.
At the end of the day AI provides a better way of processing information efficiently and accurately in order to help make more sound decisions. Financial firms are always looking for ways to help customers make better decisions about their finances and increase the amount of capital they have through investments. AI is proving to be a valuable tool to help continue to achieve these goals.