ChatGPT and Other Large Language Models (Part 2 of 2): Wealth Management Edition
What Comes Next?
Abstract
This report is the second in our two-part series on ChatGPT and other Large Language Models. The first report "ChatGPT: What Non Techies Need to Know" unravels the AI technology that enables tools such as ChatGPT for those who do not have a background in IT. "ChatGPT and Other Large Language Models (Part 2 of 2): Wealth Management Edition" takes a deeper look at the current and future state technologies, wealth management use cases and obstacles to deployment.
Large language models are clearly breakthrough technologies that will transform the wealth management industry. However, "transformational" use cases are still evolving. Many wealth managers have articulated strategies for using the technology, but few firms have developed "breakthrough" applications. Most AI applications using LLMs have focused on helping advisors write emails, compose social posts, better manage meetings or search for documents. Perhaps useful, but certainly not transformational.
The challenge for wealth managers will be to position these tools within the advisor's workflow in a way that creates greater efficiency and deepens the client engagement. To do this, wealth managers will need to rationalize workflows, integrate regulatory and compliance requirements, create use case guardrails and ensure the accuracy and transparency of outputs created by LLMs. Advancements are being made. But true success will be realized only if the LLM tool solves a material business problem and creates a compelling benefit for advisors. Otherwise, it will be another nifty, unused app that sits in the advisor's tool kit.
Key Findings:
•ChatGPT and other Large Language Models (LLMs) represent augmented intelligence tools, which allow us to combine artificial intelligence (AI) with human intelligence to enhance and amplify human abilities, e.g., at a base level, generate content and get answers quickly, and at a higher level, improve our decision-making, problem-solving, and overall cognitive abilities. |
|
• Tools that deliver augmented intelligence have the potential to be transformational—and to add value to human tasks. • The vision of a "co-pilot" for every profession is no longer a moonshot. In wealth management, one firm envisions an on-demand CIO for advisors via GPT technology. • LLMs will prompt organizations to reassess employee talent requirements that can potentially be augmented with AI. • Depending on cost efficiency, the playing field could level with any size enterprise accessing LLMs via an API. |
|
• The risk of doing nothing could be substantial as early adopters aggressively pursue new use cases to augment businesses. • Organizations may be less operationally efficient than peers using LLMs, which may lead to long-term challenges and diminished profitability. |
• For any industry, adoption of augmented intelligence tools requires mapping unchartered waters and extensive collaboration across enterprise stakeholders and regulators. |
|
• The abundance of press surrounding GPT-4 and other LLMs might lead one to believe that this is the optimal route to employing AI within wealth management firms. And early adopters such as Broadridge, FMG, Morgan Stanley, and Orion have made strong cases. However, Celent believes that given the inherent risks associated with LLMs each firm must take a step back and weigh the options. For many wealth managers a “wait-and-see” approach may be best. • In addition, there are other AI solutions in the market that may be more suitable investments. |