Fintech Insights

Can AI Save the Sell Side?

April 26, 2018

Can AI Save the Sell Side?

Artificial intelligence (AI) isn’t new to the sell side, but it is largely restricted to helping firms make decisions in reaction to events. Firms that can extend their use of AI to predictive analytics to anticipate market events and customers decisions could set themselves apart in a hyper-competitive market.

Q. Where has AI been adopted so far on the sell side?

A. One example is quantitative trading firms that develop predictive models of financial markets by using a subset of AI to drive trading decisions – taking out the human factor. If we believe that greater than 70 percent of posted liquidity is driven by high frequency trading, then we can identify AI in the market. But that’s reactionary AI in response to an event that’s been created and how it should be handled based on the history. What AI hasn’t done is anticipate client behavior, initiating coverage or positions in proprietary businesses.

On the inside of the market, AI is making real-time decisions about routing for best execution, choosing an execution partner and a market center. That’s a more recent employment of AI, where as little as three years ago, firms used static data to drive execution decisions.

The retail brokerage community has been moving toward deploying AI on the relationship management side with individual investors. For instance, brokerages are employing AI for the robocall processes, electronic marketing and specific collateral management. The firms are targeting clients with merchandise, using the meta data universe of online footprints to drive marketing campaigns and order solicitation.

Q. Do you see any long-term impact of that?

A. The sell side will always have to innovate for the sake of cutting cost. In a short time, the industry has retired many positions that no longer have relevance and reduced the work force by upgrading its technology. Technology, like AI, continues to become more commonplace and accepted as reliable. Introducing this type of automation would have long-term impact on the labor force within the financial services community. We should expect to see a reduction in some areas of IT and sales within a broker-dealer or an institution. The more accepted use of the cloud in financial services has already caused downsizing in hosting resources, AI’s impact will surely trickle in during the next few years as well.

Along those lines, when does the buy side have its “Aha!” moment and determine AI can supply the same services as the sell side, driving many buy-side firms into the technology space, replacing commissions paid the sell side with a robust technology budget.

Q. What if firms can take that leap?

A. The next step is clearly predictive and proactive market interactions. Anticipating the future by marrying in real time the market meta data and client online footprint to determine expected behavior. If you can couple those things together, you could really position yourself to be ready for your client and acquire inventory at premiums to drive revenue.

After all, if you know you have a customer who comes in every day and wants a ham sandwich at 11:00 a.m., then at 10:50 a.m., you make a ham sandwich.

Q. Does that give early adopters an advantage?

A. In short, yes. If I can say with confidence that this client will likely liquidate this asset or augment that holding with a high enough degree of certainty, I’m going to liquidate those shares or secure the asset in anticipation of that decision.

Two things occur if I’m right. First, my revenue will go up because I’m most likely buying or selling at a superior price than what it would be when everyone else figures it out. If you’re in line first and improving your margins, you’ll be able to invest and grow ahead of your peers.

Second, most of the time when you see an analyst making a move, it becomes widely known and the information is useless. You have just a short window to pull that in, make some decisions, develop some strategies and move on. AI will gain traction as it proves it can outperform a trader at these tasks.

Research firms are gaining an advantage in the market, often having their own execution desks that manage client relationships. They have been able to delicately couple the usage of AI, meta data and old-fashioned research to increase business through position initiating and client solicitation. You know they’ve been successful, as the division between research and execution has fallen under regulatory scrutiny in recent years.

The sell side has an interesting, although short, opportunity to remain relevant and maintain its position in the overall markets by quickly embracing better decision-making technologies like AI. If it’s there, if it works and if it’s inexpensive, the buy side will continue to use it. However, the sell side runs the risk of downsizing itself through efficiency, and possibly out of existence. The sell side is reaching its tipping point – and firms that will prosper have made sufficient progress on the automation front and have become technology companies.