Vendors
日本語
search
  2. //
  3. Reports & insights
  4. //
  5. Three Gen AI Use Cases in Wealth and Asset Management Today: Lessons from AWS

Three Gen AI Use Cases in Wealth and Asset Management Today: Lessons from AWS

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)
10 April 2024
Chris Bendtsen

Where is generative AI being used in wealth and asset management today?

I spoke to leaders at AWS to learn about typical use cases for generative AI in financial services – and in wealth and asset management in particular – that AWS customers are focusing on right now. Three of the most common gen AI use cases among financial services customers at AWS are:

  1. AI assistants
  2. Code development
  3. Call center automation

AI assistants

Many of AWS's financial services projects leveraging Amazon Bedrock (AWS's fully managed gen AI service) involve an employee-facing AI “agent” or assistant. Organizations use agents to accelerate employees in making human decisions by summarizing relevant information and data.

Amazon Bedrock Agents connect to company systems, data sources, and knowledge bases to retrieve and summarize information, enabling humans to do their jobs more quickly and accurately. Agents can apply to insurance agents processing claims, investment analysts producing research, financial advisors giving investment advice or proposing a financial plan, and virtually any task where a human extracts information from a document repository or database.

Asset management firm Bridgewater Associates leverages Anthropic’s foundation model Claude on Amazon Bedrock to develop an “Investment Analyst Assistant” that supports junior analysts. Human analysts can ask the gen AI agent in natural language to create data visualizations for stress testing market hypotheses as part of investment research, saving them the time of having to create the charts themselves.

Code development

Generative AI code development, which spans across business functions and can be a subset of the agents use case above, is starting to see rapid adoption within financial services. Gen AI is being leveraged as a coding accelerant in two ways:

  • Empowering junior developers to code faster
  • Speeding up mundane tasks for experienced developers, which frees up time for more challenging work

AWS has a few gen AI code generation products: Amazon Bedrock Agents, Amazon CodeWhisperer, and Amazon Q Developer (an off-the-shelf AI assistant currently available in preview).

Athene, a retirement services company offering annuities products, piloted Amazon Bedrock Agents to mine legacy code documentation files and perform data mapping. Human analysts were then able to ask the agent questions in natural language to understand the data mappings and logic. In an AWS re:Invent presentation last November, Athene shared that the pilot streamlined what was an 80-hour manual process into just a few minutes.

Call center automation

AWS is also seeing increasing adoption in financial services for gen AI in customer service call centers, specifically with:

  • Transcript summarization
  • Post-call analytics
  • Response recommendations and issue resolution (via Amazon Q in Connect, which is currently available in preview)

Principal Financial partners with AWS for call center post-call analytics to better understand the voice of the customer and customer service agent performance. By leveraging Amazon Transcribe and Amazon Bedrock’s generative AI, Principal Financial generates call summaries, insights, and sentiment analysis. Customer service agents can also take advantage of a gen AI tool powered by Anthropic’s Claude. The human agents can ask questions in natural language about call insights and recommendations for improvement.

Other gen AI lessons from AWS

Here are a few more takeaways from my conversations with AWS on their view of generative AI as it stands today:

  1. Model choice: customers appreciate the choice of foundation models, where the optimal model can vary depending on performance requirements and costs.
  2. Human in the loop: generative AI is meant to accelerate human decision-making, not replace it.
  3. Generative AI is not the answer to every problem. Traditional AI, machine learning, and natural language processing are often the best solutions.
  4. Legacy financial institutions, especially in insurance, are moving much faster than expected in testing and implementing generative AI.