Recently, I co-hosted Aite-Novarica Group’s Broker-Dealer Technology Research Group meeting, along with my colleagues Nancy Casbarro, Mitch Wein, John Keddy, and Jack Krantz. Joined by an array of carrier-owned broker-dealers, this session gave us the opportunity to dive into emergent issues that showcase the complexity of the regulatory environment faced in this space, as well as the dramatic shifts in demographics reflected in producers and customers alike. Two key areas of focus emerged out of the discussion: generative artificial intelligence (AI) and data.
Challenges and Opportunities With Generative AI
One of the key areas of interest for the group was the challenges and opportunities associated with the deployment of generative AI capabilities to support operations. Like many other lines of business, there’s an almost palpable excitement around how tools like ChatGPT and Bard could help with customer service functions, with the transition of knowledge across an organization, and with improving the quality of advice and the speed with which it is delivered.
While there’s no doubt that these tools offer interesting possibilities, there’s an almost Dickensian potential here. It is the best of times and the worst of times—all at the same time. The tooling can provide some remarkable capabilities which can be used for “good”, perhaps as a force multiplier for many operational functions in front, middle, and back offices.
This may be particularly important in a world of low unemployment rates and in the face of a rapidly aging population of employees across multiple key functions. Registered reps in the field, compliance teams in the Office of Supervisory Jurisdiction (OSJ)/middle office, and operations teams everywhere could benefit.
And yet, there’s a potentially darker side to generative AI. Like many emerging technologies (e.g., robotic process automation (RPA), low-code/no-code solutions) that have great promise, ChatGPT and similar tools need to be deployed with eyes wide open. For one thing, governance frameworks are critical. Identifying what tools can be used, for what purposes, and with what approval processes will be central to avoiding embarrassing and expensive errors. Testing things in production can never be “a thing”, a sentiment broadly agreed with by the group.
Optimizing Data Flows
Moving on from new technical capabilities led us back into a conversation as old as the creation of broker-dealers: challenges with data. While wirehouses and independent broker-dealers (IBD) have business models which allow them to control the flow of data reasonably effectively, insurer-owned broker-dealers face a set of unique problems driven by the wide array of data sources they need to support.
The mix of proprietary and non-proprietary business supported, as well as the combination of both brokerage platform and direct (i.e., “check and app”) business which comprise their portfolio, are truly unique. Most regulators come from wirehouse/IBD backgrounds, which suggests that they have limited experience with the realities faced by carrier-owned broker-dealers. This makes the data challenges these broker-dealers face even more daunting. All of the participants highlighted challenges associated with their turnkey asset management program (TAMP) relationships, as well as issues related to the lack of real standards to support the flow of information from sources such as the DTCC and NSCC. This can be compounded by special challenges associated with proprietary data feeds which predate any need to consider how distributors would consume data from product manufacturers.
There isn’t a magical, commercial solution that can solve these issues. The best success stories come from firms that have taken the time to make investments in creating their own “Rosetta Stone” to allow for the normalization and standardization of data that can then be consumed locally as actionable information.
The consumption of that curated data becomes even more important now as broker-dealers look to support changes in both consumer and producer demographics as well as address an increasingly varied array of tools to support field operations. While the tools available in the market today are impressive, an inability to support them with data that is highly credible will only serve to minimize the value of the investments and create significant frustrations between the front and middle offices, which are foundational to how broker-dealers operate.
Our Broker-Dealer Research Group continuously explores these and other critical issues, such as the onboarding of new registered reps in both broker-dealers and RIAs. The conversation will continue at our next scheduled meeting on August 17th. For more details on these or other issues, please reach out to me at RMcIsaac@Aite-Novarica.com.