A lot of time and effort has been invested in figuring out the most effective way to leverage generative large language models to support consultants and their firms.
And for good reason. Industry analysis after industry provides tools like ChatGPT to reshape the way we work and redefine the term “modern consultant”.
But is there any value in putting control of the bots in the hands of the end client, rather than confining these tools to just another part of the consultant’s toolbox?
Researchers at the University of Chicago Booth School of Business, including doctoral student Alex G. Kim and Chicago Booth professors Maximilian Muhn and Valeri Nikolaev, are investigating the feasibility of offering services to financial services clients.
A paper published earlier this year titled “TheBloated Disclosure: Can ChatGPT Help Investors Process Information?three researchers set out to determine the usefulness of ChatGPT in helping investors analyze and understand unstructured text data.
Specifically, the team used the stock market as a laboratory, using GPT language models to summarize information disclosed by companies in annual reports and conference calls.
during an interview financial planThe goal, says Kim, is to see if AI can reduce the burden of disclosure and create targeted summaries for investors interested in specific topics, such as ESG activity or corporate risk exposure.
Starting with a randomly selected sample of about 20% of all MD&A and conference call transcripts from 2009 to 2020, Kim and others instruct GPT-3.5 Turbo to generate comprehensive summaries of each document without citing other documents or external sources.
Questions were raised as to whether the shortened abstract still provided complete information. But chatbots have shown a strong ability to preserve — and in some cases enhance — the value of information provided, Kim said, pointing to bloat in corporate disclosures.
The tools can also provide support when issues such as terminology, inexperience or complexity create obstacles that investors feel cannot be cleared, Kim said.
It has brought down the wall of intimidation between customers and the financial services community. Once removed, it may open a path for more people to work with advisors.
“There has been a lot of research and warnings about the readability or informativeness of these disclosures, especially for retail investors. When investors are cognitively limited, they can’t fully process everything, and they tend to drop some of the information. Those parts. Disclosures are made on a case-by-case basis,” Kim said. “It would be cool if we could somehow distill this information and make it more relevant, easier to read, more accessible to retail investors. Before, we didn’t have the technology to do that. In theory, It’s possible. But now, we can actually do it.”
Created by AI and research company OpenAI and released to the public in November 2022, ChatGPT is an AI chatbot based on the GPT (Generative Pre-Trained Transformer) language model. Using deep learning techniques, the tool is able to generate conservative, human-like responses to text input.
Research from the Chicago Booth School of Business found that MD&A and conference call transcripts generated by the technology are typically less than one-fifth the original length. Sentiment is also amplified as the document wraps up, with positive documents generating more positive summaries and vice versa.
The researchers say this may be due to companies “hedging” their views with precautionary statements that don’t contain much real information or are mostly boilerplate. They also found that inflation tends to be higher when a company reports losses, has negative sentiment, and experiences a negative stock market reaction.
The patterns are consistent with companies using irrelevant language in their disclosures to obfuscate negative information, according to a research brief.
When making the case for this kind of customer-facing AI, especially amidst the current AI boom, Kim said there must always be a level of caution every time ChatGPT is implemented.
“It’s a machine. It’s not magic. Generally speaking, all the machine does is complete sentences in the most natural way possible. That’s all. When you ask the machine questions that involve decision-making, sometimes there are problems. “Illusion. It just gives you the wrong information,” King said. “If you try to correct it, it will say. ‘Sorry’ and then it will make up another message.
“We’re trying to minimize the possibility that it’s an illusion or creating fact. Then the decision is on the side of the investor.”
Ken Lotocki, Chief Product Officer Conquest Planning, a Canadian financial planning technology startup, believes it would be beneficial to provide firms, advisors, and clients with direct access to tools powered by generative artificial intelligence and ChatGPT. But again, along with support comes caveats.
Founded in 2018, the Winnipeg, Manitoba-based company relies on its proprietary, AI-based Strategic Advisory Manager to streamline the advisory process by eliminating trial and error when developing financial plans.
With safeguards against providing misleading information, the tools can provide contextual advice specific to each user’s financial status and literacy level, Lotocki said.
“For firms and advisors alike, it is on record that advisors unfortunately spend too much time on various financial software and too little time supporting clients. For example, we found research showing that developing a financial plan can take 17 Hours and more,” Lotocki said in an email to Financial Planning. “Continuing with the example of planning software, generative AI/ChatGPT powered tools can be leveraged within existing applications to enhance user experience, providing simplified access to existing and new features to assist advisors with UI navigation, unique situation analysis and content generation such as report summaries.”
He added that Conquest has an ethos of democratizing financial advice, and they believe generative AI can help make that happen.
“When it comes to end clients, it is well documented that we have a financial literacy problem all over the world. This is why some clients avoid seeking financial advice from professionals more frequently,” Lotocki said. “Clients can be intimidated by things they don’t know and can be embarrassed by their situation. Using generative AI/ChatGPT powered tools can alleviate this where customers can ask questions about their situation at home, Asking questions about products and concepts they may not understand better prepares customers for interactions with expert human advisors and companies.”
Brian McLaughlin, President of Technology, Orion Advisorsaid his team’s approach is to embed assistive technology like ChatGPT into our existing tool suite to make advisors’ lives easier.
“We believe in bringing artificial intelligence to advisors so they can streamline operations and communications so they can focus on what’s most important at hand — their clients,” said McLaughlin, noting that Orion’s third-party integration allows advisors to review all generated Information. Content comes first.
“The reality is that AI can make mistakes,” he said.
McLaughlin also expressed support for AI for tasks such as comparing and contrasting portfolios, refining marketing content, and responding to requests for proposals.
“We encourage consultants to challenge their internal teams to think about how AI can be leveraged,” McLaughlin said.
For Eduardo Fontes, Senior Vice President of Data Science, Wealth.comcontrol is one of the most important things to remember when extending access to these tools to anyone in a business or financial services environment.
Fontes, who has a master’s degree in data science from the Massachusetts Institute of Technology, said proper constraints not only make AI-driven interactions safe from a legal perspective, but hyperspecificity could allow for deeper responses related to topics of interest.
“Because when you look at these large language models today and ask a general question, you probably don’t want your brain to be thinking so generally and broadly about very specific documents. We have the ability to control that. That’s no. 1,” he said explain. “In my opinion, the organizations that are going to really advance this technology, and I think that should explore this technology, will know how to control the ins and outs. Because the ins and outs do matter. Especially in terms of output.”
Ordinary American investors are already paying attention to the technology. According to a recent consumer survey by Magnifi, TIFIN’s AI investing marketplace, 66% of Americans believe they understand how AI works thanks to ChatGPT.
Overall, the survey found that 60 percent of Americans are optimistic about AI. Of those, 24% said they were glad it made their lives easier and better, and 36% said it seemed to have the potential to improve their lives.
The survey also found that 41 percent of respondents would make investment decisions guided by artificial intelligence. Millennials (57%) and Gen Z (52%) are the most eager to use ChatGPT for investment advice.
But a third (34%) of respondents said they did not understand AI and were unsure (24%) or distrusted it (13%).