AI in wealth management

Artificial Intelligence will increasingly be used in wealth management.

Firms are already able to leverage AI technology to provide content creation, investment recommendations and customized client solutions.

AI is the ability of machines and computer programs to perform tasks that usually require human intelligence, such as perception, reasoning, learning, decision-making and natural language processing. 

What’s captured popular attention lately is generative AI – artificial intelligence that can create new, original content or data, such as images, music, or text, rather than simply processing or analyzing existing data. It uses algorithms and machine learning models to generate new outputs based on patterns and correlations in data.

Let’s understand what programs that create generative AI like ChatGPT are. 

At its core, ChatGPT is a chatbot. You put in a question and some input, and the chatbot produces a response. It’s conversational; it understands the context of the question you asked. This language model generates human-like responses to natural language inputs, such as questions or statements.

ChatGPT is trained on a large corpus of text data using a deep learning algorithm called the Transformer architecture, which allows it to understand the context and meaning of the input text to generate relevant and coherent responses.

Generative Pre-trained Transformer (GPT) is a transformer model that uses attention mechanisms to process sequential data, such as text. This allows the model to focus on the most relevant parts of the input sequence and generate output that considers the context of the entire sequence.

It’s not just college students looking for an easy way to write essays using ChatGPT; many use cases exist. I’ll give you a quick example. I was talking with my 22-year-old son, who recently graduated from college. He has an entrepreneurial spirit and has started his own company, where he is working to create an app.

At first, my son was skeptical about using ChatGPT. Once I showed him how it works, he spent the next two hours working on different use cases for his company. Now he uses ChatGPT to write copy for his website and to help him code. ChatGPT's ability to produce code is not appreciated by many people outside of computer science.  Let’s see if I can change your mindset.

Using ChatGPT for advisor content creation

Using ChatGPT for content creation can help financial advisors save time, maintain consistency, and personalize content – ultimately improving their efficiency and the quality of their services. Content creation is an excellent, classic use of ChatGPT. 

Let’s pretend I’m an advisor. I'm looking closely at the portfolio performance of my clients on AdvisorEngine– specifically, the 60/40 split (60% stocks, 40% bonds) clients. Clients on a 60/40 portfolio model experienced a hit, despite, as you may know, the benefit of the diversification of a 60/40 model. 

I can use ChatGPT to generate an email to send to clients to reassure them. I simply input into ChatGPT, “Write an email to my client explaining how the recent downturn that affected even 60/40 portfolios was unheard of, to remain calm and explain how sticking with the current 60/40 portfolio mixture has been proven to work throughout many years in the market.” 

Here’s ChatGPT’s response:

The first thing you’ll notice, it has pleasantries in the beginning and at the end. While I didn’t input a specific tone in this case. You can always add tone – whether you want to take a professional or friendly approach. You'll also notice that it mimics your question, saying this is not an unheard of event. You'll see that later on in the text as well. This is very common in chatbots and it makes sense. ChatGPT seems to understand that a 60/40 mix is stocks and bonds. Actually, it doesn't really “understand” it; it just knows there is a high correlation between 60/40 and stocks and bonds in the content it has previously digested. 

Now that ChatGPT has created email content to work off, I can customize and send it to my clients. It took me about two minutes to come up with a question and input it into the chatbot. It took ChatGPT about 20 seconds to produce an answer and about one minute to drop the email into AdvisorEngine CRM and disseminate it to the clients using a 60/40 model. That is efficient and a huge time saver. Who wouldn't use something like this? You can create content for a letter or a blog post. Advisors will do this; some are even using it now. 

Using ChatGPT to generate formulas

Microsoft Excel or Google Sheets rank among the most commonly used wealth management tools. Excel and Google Sheets have a vast range of formulas that can help perform calculations and manipulate data. 

If you've never dug deep into these programs, that is okay because now ChatGPT can explain the structure of specific formulas and how to use them. ChatGPT can also help diagnose common problems and formula errors in Excel and Google Sheets. You can ask ChatGPT to explain why you see an error message or how to fix a formatting issue.

ChatGPT can explain complex functions, such as VLOOKUP, SUMIF, COUNTIF, INDEX and MATCH. It’s the perfect resource to write formulas for Excel or Google Sheets, whether simple or intricate. You can ask the chatbot to write more complicated formulas and test their limits. 

The accuracy of ChatGPT's results depends a lot on how clear your prompts are. For ChatGPT prompts to be effective, they need to be very specific. You should tell the chatbot precisely what you want it to do, with as much detail as possible. 

Issues and concerns with AI use

While ChatGPT and other generative AI have benefits and can be useful, there are also several issues and concerns to consider, including the potential for inaccurate information, bias and misinformation. 

Here is an excellent example of a ChatGPT limitation. When “can you form a sentence so that the 4th word begins with a ‘y’?” was imputed, this was ChatGPT’s response:

As you can see, the chatbot provided an incorrect answer to the question. While that might not seem like such a big deal, how about this example – one ChatGPT user asked what the ‘elements of breach of contract under New York law’ were.

He also prompted the chatbot to include case citations. Here was his prompt and ChatGPT’s  reply: 

It seems legit, right? Well, four out of the five citations are hallucinations. Citations one, three, four and five do not exist and citation two is irrelevant. My point in showing you this is that it will require some fact checking and oversight if you put ChatGPT to use in your practice.

Here are some additional issues and concerns for advisors to manage:

Financial advisors should know the legal and regulatory issues involved in using ChatGPT for content creation and take appropriate measures to mitigate potential risks:

• Data privacy: 
  • Ensure the data used to train the ChatGPT language model is compliant with data privacy regulations, such as GDPR and CCPA
  • Data should be collected with proper consent and is stored and processed securely
• Accuracy and reliability:
  • Ensure that the ChatGPT-generated content is accurate and reliable
  • Any inaccurate or misleading information can lead to legal issues and reputational damage.
• Compliance:
  • Double-check that ChatGPT-generated content complies with industry regulations, such as SEC and FINRA rules, and other relevant laws and regulations and your compliance program
• Intellectual property:
  • The use of ChatGPT for content creation can raise intellectual property issues, particularly if the language model is trained on copyrighted material.
  • Advisors must have the necessary rights to use the data used to train the model
• Liability:
  • Advisors may face liability if the ChatGPT-generated content causes harm or damages to clients or other parties
  • Have appropriate insurance coverage to protect against potential liabilities

Where do we go from here? 

Controversy over new art-making technologies is nothing new. Many painters recoiled at the invention of the camera, which they saw as a debasement of human artistry. In the 20th century, digital editing tools and computer-assisted design programs were similarly dismissed by purists for requiring too little skill of their human collaborators.

AI is already transforming the wealth management industry – providing new tools and capabilities that will help advisors enhance the client experience. With the increasing amount of data available, AI can help advisors process and analyze information faster, identify trends and patterns, and generate personalized recommendations that consider each client's unique circumstances.

While there are challenges associated with implementing AI in wealth management, such as regulatory compliance and ethical considerations, the benefits of AI make it a powerful tool for advisors looking to stay competitive and provide the best possible service to their clients. As AI continues to evolve, we expect it to play an increasingly important role in the future of wealth management.