AI in Financial Services
It should come as no surprise that the financial services industry has at their disposal some of the most sophisticated technology in the world. The annual IT budgets of the largest banks are in the billions of dollars, equal in size to the government budgets of many emerging market countries. The software purchasing habits of a large bank can single-handedly move the stock prices of some of the world’s largest technology companies. Add to this that the industry’s paramount concern is moving, securing and transacting with the digital 1’s and 0’s that represent the world’s money supply, and that they have access to billions of financial records and can see trillions of transactions.
The financial services industry has been an early adopter and proponent of Artificial Intelligence, with most large established banks having a defined AI strategy and many having AI development teams running in-house. Financial services is also rich ground for innovative AI startups.
A recurring theme, however, for the larger banks is the lack of available qualified AI developers and data scientists, which may be holding back the speed of development.
The previous wave of startups and consumer facing applications for the financial services sector was when the term ‘Fintech’ was first introduced. These applications and services were largely inspired by smartphone adoption. Smartphones allowed financial services startups and major banks to take advantage of technologies like location awareness, encryption, signature capture, secure login. The emergence of public and private cloud computing platforms secure enough to handle financial data was also a key enabling technology.
Artificial intelligence is leading the next wave of applications and services for the financial services industry. AI technologies are adept at handling unstructured data such as images, video, audio, location and time-series data. Data can either be created by devices, online activity or market activity. AI solutions have been developed for fraud detection, assessing credit risk and identity verification. In the insurance sector, AI solutions have been developed for automating accident claims, identifying health risks and enabling customers to avoid them and insurance fraud detection.
One way banks are using artificial intelligence is to improve customer service and engagement. Many have rolled out chatbots for real time customer service and information purposes, leveraging their customers presence on the major chat platforms. Others have developed full virtual assistants, similar to Apple’s Siri or Amazon’s Alexa — to help customers find products or conduct financial transactions.
Many banks are also using AI to automate their own internal activities, such as filling out forms, filing records and receipts, and assessing risks.
Following are several examples of AI initiatives at the largest banks:
- JP Morgan is using AI to automate credit agreement analysis. It recently unveiled its Cointract Intelligence (COiN) platform to analyze these agreements and highlight key clauses and data points. This work previously required 360,000 hours of human labor.
- Wells Fargo announced the creation of a dedicated Artificial Intelligence Enterprise Solutions team which will pursue payments technologies and improved customer experience for corporate banking customers. They have also published on the use of AI in cyber security, suggesting they may also be pursuing research in this area as well.
- Bank of America released Erica, an AI-powered virtual assistant that they plan to roll out in their mobile app and at their ATM network.
- Citibank has made several investments in AI related companies, including Feedzai which uses AI to identify and fight fraud in online banking; and Clarity Money, which recommends financial products to help customers save money and better manage their finances.
Financial Services is also a popular sector for a large number of AI startups. In many cases, start-ups are aiming to disrupt the traditional businesses of the large banks. In other cases, they are looking to provide advanced new services to the banks to allow them to improve their product offerings.
Several major themes have emerged among AI startups in the financial services sector: fraud detection, advisory services, personal financial management and trading assistance and execution.
In terms of fraud detection, AI has proven itself to be extremely useful. Comparing a customer’s spending behavior against massive amounts of historical data allows for extremely sophisticated and proactive fraud alerts, which are continually learning and being updated.
Advisory services such as robo-advisors can use various data sources to ascertain risk tolerance and recommend appropriate investments and financial products.
Personal financial management has been a particularly promising space for fintech, having seen several successful startups, including Mint and Wallet. These platforms have the ability to accumulate personal financial data to track finances over time and make recommendations, and their ease of use brings sound financial management within reach of many who previously never had the patience or know-how to make a spreadsheet to keep track of their finances, track spending by categories over time or create spending or savings goals.
Following is a selected list of several of the most exciting recent AI startups in the sector:
DreamQuark is a platform for developing and deploying AI applications designed specifically for the banking and insurance industries. Use cases include product recommendation and customer segmentation engines, fraud detection and credit scoring, and compliance functions.
Alpaca provides dashboards for financial market predictions. They use deep learning on high frequency tick data to recognize patterns indicating price change for their market forecasting models. They developed their own high bandwidth scalable database server optimized for financial time-series data called MarketStore, which they have made open source.
DataVisor uses AI to detect fraud and other financial crimes. The company employs unsupervised learning models to find previously unknown fraud patterns, resulting in up to 50% higher performance than competing systems.
Quantexa is another exciting new company in the financial services space. They use sophisticated entity resolution and network building technology to track actors based on large amounts of disparate and sometimes incomplete financial data. They use AI to predict default risks, proactively identify fraud, and profile not only bad actors but good customers and even identifying connections between them.
Neuromation looks forward to assisting in the development of the next phase of AI-driven applications for the financial services industry. For more information, please contact us directly.
By Angus Roven,
Neuromation Investor Relations Analyst
