Welcome! Why you should subscribe…
Welcome! Subscribe to get full access to our newsletter and feel free to visit us at O'Boyle Wealth Management. Upon subscribing, every new edition of our newsletter goes directly to your inbox. Never miss an update!
Why I started this experiment with Artificial Intelligence and financial planning?
Hi there! My name is Joe O’Boyle. I am a financial planner and principal of O’Boyle Wealth Management in Beverly Hills, California.
I am very interested in Artificial Intelligence and I wanted to see how Open AI’s ChatGPT tool would respond to the financial planning and investing questions that our clients ask on a regular basis.
The initial idea was that I would refine the AI answers/responses and offer you, our readers, unique planning and investment perspectives towards helping you achieve your goals. So that you can tell the difference between the AI response and my response in each article… my response is written in italics.
The ensuing journey and the results have been somewhat different than what I originally expected…
What I’ve found thus far:
Initially, I would ask AI one question via ChatGPT and then refine the AI response. However, the AI response was generally too vague and occasionally had errors. This approach wasn’t going to get it done.
Next, I would ask AI multiple questions about a specific topic in a variety of different ways. This might involve 5 to 10 different prompts about the same question. This would help shape and guide AI’s response, but it also meant I now had 5 to 10 different responses to review. As you might imagine, this resulted in my taking on more of the role of an editor, not a writer. I was synthesizing the AI responses and organizing them into one coherent response for you, our reader.
In the next iteration, after 5 to 10 prompts about a single topic, I would then ask AI to refine or reshape a response to a specific paragraph or idea within AI’s own initial response(s). In many cases, I would offer additional details or explain that the point AI initially made was actually incorrect or only partially complete or could be expanded upon. The Artificial Intelligence ChatGPT tool would then correct and provide a better response (generally) to that specific paragraph or idea. This approach produced a better outcome for you, the reader. The challenge is that, depending on the topic and quality of AI responses, this became time consuming.
Surprisingly, basic math problems with compound interest calculations would sometimes be incorrect. Additionally, I found that quotes by an esteemed investor would sometimes be made up or not entirely accurate. The fact checking that was required added another layer of time and complexity.
My main takeaway in using OpenAI’s ChatGPT tool is that it is an exciting time! The idea behind large language models (LLM) with reinforcement learning from human feedback (RLHF) is that over time the AI tool (such as ChatGPT) will get better and better due to human use. Eventually, the goal is that when you ask a question, the AI tool will give you the answer you’re looking for… and maybe in the not too distant future will then even be able to execute an action on your behalf.
What our readers (investors) want:
“Don’t teach me about financial planning and investing. Just tell me what to do.”
Ultimately, based on my 20 years of experience, investors (our readers) don’t want general information about a topic, they want guidance that’s specific to their situation. There is a big difference.
For example, sharing that a Roth IRA may grow tax-free for retirement is information. While helpful, it’s not necessarily actionable.
However, when your income is above a certain threshold, sharing that you don’t actually qualify to contribute directly to a Roth IRA, but that you may, instead, contribute the maximum of $6,500 for the 2023 tax year to an after-tax Traditional IRA and subsequently convert this full balance to your Roth IRA (as long as you don’t have any other pre-tax IRA assets) is guidance. For what its worth, this strategy is known as a “backdoor Roth conversion” and is a way that our wealthy clients position their funds to potentially grow tax-free for retirement in a Roth IRA. We typically recommend that our clients, if applicable to their specific situation, take advantage of contributing to an after-tax Roth 401(k), a Roth IRA (via a Traditional IRA contribution and backdoor Roth conversion), and a pre-tax Health Savings Account. This strategy offers up a planning trifecta to have your money potentially growing tax-free for your retirement across all 3 tax-advantaged account types.
This is guidance. Whether from AI or from a financial advisor, this is what our readers want.
To find out more about the O’Boyle Wealth Management, the company that provides the insights for this newsletter, visit www.oboylewm.com.
P.S. Just for fun… I asked AI to review this “About Page” which I wrote exclusively (no AI).
Here is the AI response:
The content is well-written and thorough, presenting your motivations, methods, challenges, and findings while working with AI in financial planning. You successfully depict the growth and potential of AI while highlighting its current limitations.
There are a few areas that you may consider expanding upon or adding for a more comprehensive outlook:
Specific examples or case studies: You could include specific examples or anecdotes where AI either triumphed or stumbled significantly in the context of your financial planning and investment advice. This could give the reader a clearer understanding of what exactly these interactions with the AI look like in practice.
Client's perspective and feedback: This entire experiment aims to better serve your clients, and it could be insightful to share some of their reactions or feedback on this new method. Have they found it beneficial or confusing? How have they responded to your use of AI technology?
Future outlook and plan for integration: While you've mentioned the AI's future potential, you might want to clarify more explicitly how you intend to incorporate this technology in your practice in the future. Will it be a supplemental tool, or are you aiming for it to take over certain roles?
Ethical implications: Touching upon the ethical considerations when using AI in financial planning could add another layer of depth to your discussion. This can include data security, responsibility for AI-generated advice, etc.
Evaluation of AI-generated financial advice: It would be interesting to hear your thoughts on the accuracy, reliability, and depth of the financial advice generated by the AI. How does it compare to traditional financial advice?
AI’s scalability and accessibility: One of the benefits of AI is its scalability, being able to handle a large number of users simultaneously, and its potential to democratize financial advice. You could reflect on these aspects.
Remember that adding these elements would probably require you to provide more context and explanation for readers who might be unfamiliar with these topics, so that they can fully understand and appreciate your insights.
