Journal of Financial Planning: October 2023
Dani Fava is group president of product innovation at Envestnet, the financial services technology company transforming the way financial advice and wellness are delivered. For more information, please visit www.envestnet.com, subscribe to https://envestnet.blog, and follow the company on Twitter (@ENVintel) and LinkedIn (www.linkedin.com/company/envestnet).
JOIN THE DISCUSSION: Discuss this article with fellow FPA Members through FPA's Knowledge Circles.
FEEDBACK: If you have any questions or comments on this article, please contact the editor HERE.
The world of tax planning for high-net-worth clients is a complex labyrinth of changing tax codes, market uncertainties, and adaptive financial strategies. Today, financial planners must be forward thinking—not only mastering the nuances of tax regulations, but also anticipating economic trends and adapting strategies to fluid tax landscapes. Add to that a myriad of client types, including inheritors, entrepreneurs, and high-income individuals. Amidst these complexities, technology is stepping up, with generative AI emerging as a potential game changer in tax planning.
The Current Landscape and Exploring AI Applications
Generative AI, a subset of artificial intelligence, is a technology capable of creating new compositions. It has demonstrated the ability to compose everything from conversational text, creative images, code, and more. Before generative AI, other forms of AI, like machine learning, showed the capacity to make predictions based on existing data. Generative models go a step further, generating new works with the same artifacts as the training set (text, images, code, etc.). These models are preferred to help planners with tasks that require the creation of new ideas, prediction of future scenarios, or simulation of complex processes. The applications of generative AI span multiple industries from art, where it can create unique pieces, to science, where it can simulate molecular structures.
The financial sector is already reaping the benefits of generative AI. It’s used in algorithmic trading, risk management, and fraud detection. In fact, earlier this year, Goldman Sachs announced1 it started experimenting with various use cases like classification and categorization for documents, and JPMorgan announced2 the early development stages of its own ChatGPT-like software to help with selecting the right investment plans for customers.
The Future of Tax Planning
Generative AI’s application in tax planning, particularly for high-net-worth clients, holds immense potential.
This technology can simulate myriad tax scenarios based on varying assumptions to identify optimal tax strategies. It’s like having an army of tax experts running a multitude of scenario analyses in seconds, providing a level of optimization that would have been unimaginable five years ago. In a real-world example, data scientists at Salesforce have developed a system called the AI Economist that uses reinforcement learning to identify optimal tax policies for a simulated economy. The AI-controlled policymaker runs its own reinforcement-learning algorithm to devise the tax rate for all workers in the simulation. It’s easy to imagine how an application like this can be extended to high-net-worth taxpayers.3
Moreover, generative AI has the potential to stay one step ahead of the ever-evolving tax laws and regulations. It can predict how these changes may impact a client’s tax liabilities and suggest adjustments in their tax strategy proactively.
Lastly, by employing predictive analytics, generative AI can forecast market trends. This foresight equips financial planners with crucial insights to guide clients on future investment strategies while keeping tax implications front and center.
Consider the hypothetical case of a global investment firm that may utilize generative AI to optimize tax planning for its high-net-worth clients. Firms like this face the challenge of efficiently navigating the complex tax landscape and making informed decisions—for potentially thousands of clients at once. If global investment firms implement generative AI, they could simulate thousands of tax scenarios and optimized tax strategies using their customers’ existing data, which could save their clients millions of dollars annually. The challenge here, seemingly, would be the communication to these clients of potential strategies to minimize tax impact. However, generative AI is uniquely capable of crafting conversational text in personalized form. In other words, the machine can write a targeted email with instructions to every individual client. Moreover, the AI system would also stay ahead of changing tax laws, allowing the firm to provide proactive tax advice.
Pulling the thread on AI’s potential to create conversational text to instruct clients, it can also create content to educate clients. Tax code is one of the most complex topics and is often considered one of the driest. Imagine inserting AI to craft proactive educational material on the adviser’s behalf. AI will understand their target market, is receptive to prompts that tell it what language to use, and can scour its source material in seconds for tips and tricks.
As an experiment, try prompting ChatGPT with the following: “I’m a financial adviser. Most of my clients are wealthy, retired people who are holding low-cost-basis stock that would cause a big gain if they sold. I want to educate them about donating their stock via a donor-advised fund and the tax benefit that they’d see as a result. Can you write me a 300-word newsletter on the topic? Be professional, but not overly complicated.” I’ve done it, and I’ve been pleasantly surprised with the results.
Recognizing AI’s Challenges, Despite Strong Adoption
According to a recent Gartner survey,4 nearly 70 percent of financial services leaders reported that generative AI tools have the potential for benefits rather than risks for their organization. However, this technology is still in its early stages and requires a combination of human oversight and expertise to deliver accurate insights. Still, there’s no denying the level of interest and curiosity to use generative AI as a tool. ChatGPT gained more than 100 million monthly active users in less than three months, according to data from Similarweb,5 making it the fastest-growing application in history.
Despite its promise, generative AI also presents certain challenges from a regulatory lens. Data security and privacy are paramount concerns, especially when using publicly available data models, like OpenAI’s ChatGPT. Luckily, the industry has leaders like Microsoft offering private connections to OpenAI’s large language models (LLMs) as well as companies like Meta offering their LLM as open-source code, which can be hosted in-house.
Other areas of concern include potential discrimination and bias. The underlying data that is used to train LLMs is all human generated. Bias exists in the training data, and therefore bias will exist in the output. Regulators haven’t yet determined how to control and oversee generative AI, which is why most content producers are wisely opting to review generative AI’s output before releasing it into the wild.
Another issue not to be overlooked is ethical considerations, especially when considering tax planning. A machine, even with its training on previous tax scenarios, may not understand things like familial considerations and optics when planning for the transfer of assets. Although AI creates massive scale, it also creates an accelerated need for human oversight in all decision-making and communication.
It’s essential to address these challenges head-on to leverage the benefits of this technology effectively.
Where to Go from Here
Generative AI isn’t a magic wand that will solve all tax planning challenges. However, it’s an incredibly powerful tool that can augment a financial planner’s ability to serve high-net-worth clients more effectively. The adoption of generative AI in tax planning is more than a trend—it’s the future, and it’s here to stay. As financial planners, we must keep abreast of these technological advancements and be ready to incorporate them into our practice for the benefit of our clients.
Beyond tax planning, I encourage you to think about the ways generative AI can help improve client engagement and empower you to deliver a more personalized client experience overall. Think about the ways it can enhance your interactions and adapt conversations to meet your clients where they are. Think about how data captured from these interactions can be summarized and utilized to easily outline priorities and goals, as well as recount prior conversations. As financial professionals embracing this technology, we will have more time to focus on our clients compared to the amount of time spent on business admin.
Generative AI has rapidly evolved and taken over industry conversations. Like past technology adoption trends, it seems those to act first tend to benefit the most. I encourage you to explore how generative AI works and think through how your practice could leverage it. This is a critical first step in being able to offer a competitive edge that will benefit your clients and your business for years to come.
Endnotes
- Bousquette, Isabel. 2023, May 2. “Goldman Sachs CIO Tests Generative AI.” Wall Street Journal. www.wsj.com/articles/goldman-sachs-cio-tests-generative-ai-886b5a4b.
- Son, Hugh. 2023, May 25. “JPMorgan is developing a ChatGPT-like A.I. service that gives investment advice.” CNBC. www.cnbc.com/2023/05/25/jpmorgan-develops-ai-investment-advisor.html.
- Heaven, Will Douglas. 2020, May 5. “An AI can simulate an economy millions of times to create fairer tax policy.” MIT Technology Review. www.technologyreview.com/2020/05/05/1001142/ai-reinforcement-learning-simulate-economy-fairer-tax-policy-income-inequality-recession-pandemic/.
- See www.gartner.com/en/webinar/499907/1167684.
- See www.similarweb.com/blog/insights/ai-news/chatgpt-25-million/.