If you’ve already read my previous post, you’ll know that Datos Insights held its 2023 Wealth Management and Capital Markets Forum on Tuesday, October 3, in New York City. The Datos team was joined by 160 attendees, including senior representatives from nearly 80 different financial institutions participating in three plenary sessions and giving wealth management track sessions covering some of the topics our clients have identified as being most impactful over the course of 2023.
In the first post, I chose to ignore the elephant in the room—artificial intelligence (AI)—and focus on what I thought was the other overriding topic that permeated throughout the breakouts: personalization. In this post, I’m circling back to the fact that no pre-conference conversation, refreshment hall discussion, or panel session seems to be complete without AI and its impact on wealth management.
Generative AI: Rise of the Machines
The dialogue around AI—particularly Generative AI (GenAI)—didn’t even wait for the conference to open to step front and center. Speaking before the event with a small group of attendees and speakers, there was a very transparent conversation around this topic. While much was discussed, I did walk away with two new perceptions that I think are meaningful as we think about some of the lessons from the conference stage:
- Few want to be on the cutting edge: I didn’t hear a single firm—including several that I would have thought of as already being there—say that they wanted to be on the bleeding edge around GenAI. In fact, what I heard was a lot of not wanting to be the first mover in certain areas—including anything directly client-facing—at this time because the technology is still so new, and the regulatory environment is still so immature.
- The importance of scenario planning: There is at least a theoretical doomsday scenario where AI essentially moves the cheese for the industry. There’s a case to be made for scenario planning—and essentially wargaming—around that to try and ensure that your institution doesn’t find itself on the wrong side of a Black Swan case. Very thought-provoking stuff, indeed.
Gavin Little-Gill opened the conference with a plenary session on Generative AI in Wealth and Capital Markets alongside two very large players in the AI technology and infrastructure space, AWS and Snowflake. While their conversation crossed the borders of wealth and asset management as well as capital markets, the lessons are foundational for all industry silos. Their dialogue can be boiled down to two key lessons:
- Garbage in, garbage out: You need to get your data right before you even think about AI. AI magnifies the consequences of bad data and the importance of establishing a strong foundation of data governance with clearly defined rails and guidelines. Where the data comes from, who has access to it, what happens to derived data, and how you can prove the lineage of the data are all critically important factors.
- No one GenAI solution does it all: There isn’t one GenAI solution to rule them all; different use cases require different fit-for-purpose solutions. Having an infrastructure that’s flexible enough to support more than one Gen AI capability will become critical moving forward. Moreover, the age-old question of build vs. buy appears again. Either option is an excellent choice, according to the panelists. However, going it alone on building requires a significant amount of capability and DevOps; only the largest financial institutions will likely be able to sustain this over the longer term.
AI came up in a number of breakouts that were focused on other topics. In Lisa Asher’s Can Today’s Financial Planning Deliver on the Promise of Financial Wellness, it came up in the context of retirement planning. The panelists were quick to note that AI is not new to retirement planning software solutions; general AI and machine learning have both been incorporated into the last several generations of planning tools. What they did expect to be new and noteworthy with the introduction of GenAI is the speed and ease with which advisors will be able to access key data and incorporate it into their planning process to inform better outcomes.
Similarly, AI was surprisingly not a game-changer yet to the panelists discussing CRM: How to Attain Agility and Efficiency in the Client Journey. This group of panelists was composed primarily of Salesforce users, and, when pressed about why none had mentioned Einstein AI in response to AI and where innovation is coming from, they almost universally were taking a “wait and see” approach to adoption as they looked at balancing the added cost with a clearer view to potential benefits.
The dialogue around AI may have been most impassioned in the Mass Personalization/Customization at Scale: Responding to the Voice of the Customer that I moderated. The panelists clearly took the view that AI—in the form of machine learning—already plays a significant role in the democratization of investing techniques (and their underlying technologies) that were once the domain of only the highest-net-worth investors. An audience question pushed the conversation into the sphere of whether this is all just another industry fad or if there’s steak behind the sizzle. My perspective is that unlike some prior industry hype cycles, such as around blockchain, GenAI isn’t a solution in search of a problem. Indeed, the issue here is that there are already so many obvious problems that GenAI is a solution to. Prioritizing which to go after first—and which to slow walk—will be a key gating step for the industry.
Wally Okby brought the discussion full circle with the closing plenary session: AI: Embrace it? Or Rage Against the Machine? featuring a panel of industry practitioners to book-end the opening panel of technology providers. Wally introduced the concept of the Tree of AI, which draws from capital markets use cases initially but applies equally in many similar ways, such as onboarding, relationship outreach, product development and access, compliance, etc., to wealth management.
According to the panelists, there is a tremendous amount of ideation and proofs of concept currently underway. Still, the actual implementation of GenAI use cases into production environments has been much more limited. Much of this heightened interest is being driven from the highest levels of financial institution management, including CEOs and boards asking CIO/CTOs or heads of digital transformation what their organizations are doing around AI, what they need to do, where investments are going or should go, and, maybe most importantly, how to measure return on investment.
This last point is critical to the realization that GenAI is not a light lift. Not every problem requires this degree of lift to solve (see prioritization above); GenAI shouldn’t be forced in if it’s not necessary.
Finally, this panel was unanimous in its view that every financial institution—even those most reluctant to move quickly—must develop a coherent plan around AI, how to cleanse and prepare its data, how to integrate AI into its existing enterprise tech stack, and where it can deploy low-risk trials and proofs of concept to begin field testing AI within existing business structures and processes.
One of the most important lessons of all around AI—and GenAI in particular—is one that my colleague Wally Okby has been putting front and center in presentations around his research on the topic: Whatever you think GenAI is now, it’s probably already changed and evolved and is now something completely different.
This topic is very much a living dialogue, changing by the day. Several current and upcoming Datos Insights reports feature it, including, most recently, Accelerating Generative AI Into Wealth Management, October 2023.