Banking #chatbots need to support external account aggregation. Here's why. Click To Tweet

Last month, Bank of America unveiled their new AI-powered chatbot Erica at Money20/20. The chatbot, which will be available starting next year, is being touted as an intelligent virtual financial assistant that can help Bank of America customers make smarter financial decisions.

Like other financial chatbots, Erica provides answers about a user’s finances in normal conversational language. Users can ask questions in plain English through voice and text messaging and get fast insights about their finances. The ultimate goal of this service is to provide users simpler, faster, and more complete access to their financial information to help improve their financial well-being.

But does Erica really accomplish that?

In Bank of America’s demonstration at Money20/20, Erica “talked” with a user about her typical spending and offered an insightful recommendation. But what if Erica’s recommendation didn’t factor in expenses on any accounts not held at Bank of America? The resulting insights wouldn’t be very useful and might actually provide bad information that could hurt the user’s long-term financial health. This demonstration underscores a potential problem with chatbots designed to provide helpful recommendations to users: siloed financial data.

According to a recent study, nearly half of Americans bank with more than one institution. Credit cards, student loans, primary checking accounts, and increasingly, third-party apps like Digit and Acorns, all hold clues to consumers’ financial lives, and these accounts are frequently spread out across multiple institutions and platforms. Proprietary banking bots that don’t offer account aggregation are only useful to users who keep their entire financial life within a single institution. This means that for roughly half of Americans, recommendations from chatbots lacking account integration would be worthless.

Financial chatbots are part of the latest wave of innovative digital banking products designed to bolster consumer financial health. According to a recent study by the Center for Financial Services Innovation (CFSI), the nation’s leading authority on consumer financial health, 57% of Americans are struggling financially. Similarly, the Federal Reserve’s 2016 survey of American economic health found that 42% of Americans were unable to pay their bills at least once within the last year, and nearly half of American households could not pay for an unexpected $400 expense without selling assets or going into debt. Clearly, American consumers need effective financial planning services more than ever before.

Financial chatbots offer a potential solution to this problem. By providing instant, on-demand access to users’ financial information, they can help Americans track their spending, build a budget, and set and reach savings goals more easily than ever before. And because they’re built with predictive machine learning algorithms, they can identify patterns in consumer behavior and provide proactive recommendations to guide the user’s financial decision-making process. However, institutions that don’t factor in all aspects of their customers’ financial lives can’t provide the holistic financial perspective needed to improve consumer financial well-being.

In order for banks to provide effective money management services and cultivate financially healthy consumers, a more open exchange of consumer financial data is needed. In a recent report, the CFSI highlighted the need for increased transparency of consumer financial data and outlined the following principles for effective data sharing:

  • Available: Consumers have the ability to view their financial information within the trusted and secure third-party application of their choice.
  • Reliable: Consumer financial data are timely, consistent, accurate and complete.
  • User-permissioned: Consumers provide explicit consent for access to and use of their data. Consumers can easily view, modify and revoke consent for data sharing.
  • Secure: All entities follow applicable laws and industry best practices with regard to data privacy and security.
  • Limited to the application functionality: Only the minimum amount of data required for application functionality are collected, and the data are stored for the minimum amount of time needed.¹

To be truly useful to the average consumer, financial chatbots must comply with these principles and offer full account aggregation that allows users to consolidate their entire financial life in a single location. Institutions that adopt an open data sharing model and provide their customers unrestricted access to their financial data will win increased market share and brand loyalty and align themselves for future profitability.

¹CFSI’s Consumer Data Sharing Principles: A Framework for Industry-Wide Collaboration, October 2016