In the summer of 2017, Morgan Stanley Wealth Management (MSWM) announced it had developed an NBA (next best action) system that uses machine learning and data analysis to help its financial advisors (FAs). In contrast to mass-market retail banking, wealth management is a high-touch business that predominantly serves high-net-worth individuals, and as a result has traditionally been slower to embrace digitalization than the rest of the financial sector. The idea of utilizing leading-edge technologies such as AI to upgrade FAs’ capabilities in the wealth management business is an innovation that has garnered considerable attention.
In 2017, MSWM launched its NBA system on a pilot basis at selected branches. The pilot implementation proved successful, prompting MSWM to roll out the system across offices in May 2018. The NBA system supports FAs through such means as preparing lists of advice recommendations on a client-by-client basis and responding to market events by instantly alerting FAs of which clients may be materially affected by the event, based on analysis of not only clients’ attributes and asset holdings but also records of FAs’ previous email correspondence and telephone conversations with clients. The NBA system can also appropriately select research reports that meet a client’s needs from a library of macro and company reports prepared by Morgan Stanley analysts.
Despite this automation, human FAs still play the central role in MSWM’s services. The NBA system augments FAs so they can provide services with more added value. While the notion of machines backing up FAs as they do their job may conjure images of robo-advisors, another way to look at NBA systems is as a knowledge base for FAs, analogous to a search engine that FAs can consult. In any case, given that FAs’ services constitute the core added-value from clients’ standpoint, NBA systems can be thought of as a knowledge platform whereby financial institutions create added value through FAs.
The head of MSWM’s NBA initiative Chief Data and Analytics Officer Jeff McMillan, has spoken about MSWM’s NBA system on various occasions since 2017. One thing he has said that impressed me is that MSDW intends to continue its NBA project indefinitely. Its NBA project began in 2012 with mundane preparations such as drafting of internal documentation and voice-to-text transcription of telephone conversations. An early machine learning implementation yielded encouraging but still rudimentary results according to Mr. McMillan. To improve machine learning’s accuracy, the NBA project team had to completely rewrite the book on wealth management through a tedious process tantamount to producing an encyclopedia. While NBA in the wealth management business has become a hot topic since MSWM unveiled its system, Mr. McMillan emphasizes that, unlike conventional system development projects, MSWM’s NBA project is an ongoing endeavor, not a fixed-term project.
The story of MSWM’s NBA project resonates deeply with me because it exemplifies the difficulty of increasing organizational knowledge. The importance of knowledge management (KM) as a means of upgrading organizational knowledge capabilities was first advocated 20 years ago. Back then, there was a lot of buzz around document-sharing systems as a means of enabling companies to better propose solutions to customers, but the KM boom fizzled out without much to show for it. Why did the KM initiatives of the late 1990s fail to bear fruit? In hindsight, I imagine the reason may be that they were mostly mere system installation projects of the type to which Mr. McMillan alluded.
Whether NBA has staying power remains to be seen, but MSWM’s example suggests that NBA is not something that anyone can undertake on impulse. In that sense, NBA will more likely prove to be a short-lived fad. The resolve to continue plodding ahead even after the limelight has faded is the key determinant of NBA initiatives’ success or failure.