Vendors
日本語
search
  2. //
  3. Reports & insights
  4. //
  5. Navigating Turbulence: The Promise and Peril of LLMs

Navigating Turbulence: The Promise and Peril of LLMs

Create a vendor selection project
Click to express your interest in this report
Indication of coverage against your requirements
A subscription is required to activate this feature. Contact us for more info.
Celent have reviewed this profile and believe it to be accurate.
We are waiting for the vendor to publish their solution profile. Contact us or request the RFX.
Projects allow you to export Registered Vendor details and survey responses for analysis outside of Marsh CND. Please refer to the Marsh CND User Guide for detailed instructions.
Download Registered Vendor Survey responses as PDF
Contact vendor directly with specific questions (ie. pricing, capacity, etc)
14 May 2023
Jean Sullivan

At our recent in-person event in New York, Navigating Turbulence, I had an opportunity to participate in a panel discussion, “How to Responsibly Harness the Power of Large Language Models.”Much of our discussion centered on uncovering LLMs’ potential use cases for wealth managers. Our conclusion was that LLMs have the potential to disrupt the wealth management industry, but to do so, they will need to overcome significant hurdles and inch closer to generating more trusted outcomes.

Here are some of the more compelling use cases raised during our panel session and their potential impact.

  1. Generating personalized content at scale. Wealth management firms are already using LLMs to help advisors generate personalized social media posts and emails for retail customers. Similarly, LLMs can generate sentiment analyses that can aid in interpreting the tone of a client’s message and shaping an advisor’s response. Still, these use cases are niche; while useful, we saw no industry-wide transformative effect. Additionally, extending the idea of delivering personalized content beyond the above examples to financial advice calls into question the ability of the LLM to deliver compliance approved, accurate responses.
  2. Improving productivity. Large language models are being used to generate and check code, produce documentation, and prioritize internal tasks. These applications make it easier for developers to create complex software programs, which could lead organizations to change their operations as a result. However, alterations like these—to internal processes, to workflows, to the ways code is developed—are inherently disruptive. While using LLMs to improve productivity may seem straightforward, it has deep implications and will require new ways of working.
  3. Leveraging LLMs as platforms. Imagine asking an LLM to find the best hotel in a city you’re about to visit and then using it to book a room. The same idea can be applied to the financial services industry. An individual could ask the LLM to find the best ESG financial advisors within a targeted zip code, then use the platform to generate an email or text. The LLM could also aggregate and send personal background data and information to the advisor. The reverse could also apply for advisors looking for potential leads with specific attributes within a specified zip code. Here, the use cases are limitless but still untested.

LLMs have some serious challenges to overcome—namely, inaccurate output and lack of adherence to compliance requirements—which carry with them significant reputational risk. For LLMs to be truly disruptive and transformative, they must first become a source of truth and trust.