“Some people call this artificial intelligence, but the reality is this technology will enhance humans. So instead we should refer to it as augmented intelligence.”
—Ginni Rometty, former CEO, IBM
The AI revolution is upon us, but the main question is whether AI will augment and empower advisors — or simply replace them? According to Gartner, by the end of 2024, 75% of enterprises will have shifted from piloting to operationalizing AI, driving a 5 times increase in streaming data and analytics infrastructures. The financial services industry is buzzing with AI-driven change, and it’s more than just noise; it’s music to the ears of industry experts. According to a recent Gartner survey, 75% of enterprises are moving from merely piloting to operationalizing AI, setting the stage for a revolution where financial advisors are not replaced, but radically augmented.
How are AI and machine learning transforming wealth management? To find out, Ezra Group recently hosted a webinar titled, ‘Advisor AI: From Data to Dollars’, focusing on how these technologies can enhance investment decision-making, optimizing portfolio strategies, and delivering personalized services to clients.
AI tools can best be considered as a support mechanism for advisors, not as a replacement, according to Ted Denbow, VP at financial planning software provider RightCapita and Dani Fava, Group Head of Product Innovation at Envestnet. This article focuses on their contributions, which covered conservative methods that firms of all categories can leverage to incorporate AI effectively, while minimizing risk.
Treat AI Like an Intern
What role should AI technology take on in your firm? RightCapital’s Denbow argues that AI should be treated like an intern. While the software may one day be handling complex processes, in its current iteration he believes it is best kept to busywork and simple operational efficiencies. “If it could get coffee, we’d love it to do that,” he joked, “but we think its main use will be in compiling information and serving it to teams quickly and efficiently.”
Denbow did consider the possibility that RightCapital might work to build an in-house LLM for their support and health staff, not to replace them but to allow them to provide information to advisors faster as they’re working with customers. He emphasized that they want to empower advisors to make the right decisions, not just give them the answers. In his more conservative perspective, the advisor is the fiduciary and should be the one making the recommendations to the client, not the software.
A Simple Focus
The financial planning features that Denbow sees AI as only being competent to manage the simplest financial planning features, such as cash flow analysis. He focused on the software helping to identify planning opportunities based on any gaps in cash flow data, following straight line projections and clear rules.
AI-powered software could also review client profile data to determine what information, cost points, and educational items the advisor can use to engage the client and start conversations. Denbow believes the biggest opportunity for AI is to enable advisors to have more frequent and effective client interactions.
AI Boundaries
Similarly, Envestnet’s Fava emphasized the role of the advisor as an essential connection point between the technology and the client to make sure the information is appropriately applied. For Fava, that process is all about asking the right questions. “What advisors are really good at is bringing financial benefit to customers in an non-intrusive, approachable manner,” she argued.
As an example, if your AI software is 75% certain that a particular client has held-away assets, it would be inappropriate to accuse them of hiding assets. Instead, Fava recommended that the advisor should initiate a general conversation about held-away assets and better ways that those funds could be used. Oftentimes, there’s no need to directly reference the predictive analysis that led to the conversation.
Fava also described Envestnet’s experimental process for creating capital market assumptions, which they augment with large language models (LLMs). When looking at factors that are multinational, for example, the market risk free rate for Germany, they can produce capital market assumptions which are less mainstream. While this has yet to be implemented, it has the potential to become a valuable product based onf their unique dataset.
What’s Your AI Superpower?
Most large language models (LLMs) that firms are working with are open source, as the process to design your own is labor intensive. Since the software is similar across most firms, the differentiator becomes the unique dataset that each firm puts into it, Fava explained.
She noted that this process can be surprisingly simple, and can be done straight from ChatGPT, going into Microsoft Azure and training the AI on documents you have in your Sharepoint, OneDrive or other cloud repository.
“At the bones of it, if you have a unique dataset that’s not publicly available, that’s your superpower right there,” Fava advised. When trying to plan how your company can effectively harness generative AI, start building your foundation around what your unique dataset allows you to do.
- Enhanced data security and privacy: With open-source LLMs, organizations can deploy the model on their own infrastructure and thus, have more control over their data.
- Cost savings: Open-source LLMs eliminate licensing fees, making them a cost-effective solution for enterprises and startups with tight budgets.
- Reduced vendor dependency: Businesses can reduce reliance on a single vendor, promoting flexibility and preventing vendor lock-in.
- Code transparency: Open-source LLMs offer transparency into their underlying code, allowing organizations to inspect and validate the model’s functionality.
- Language model customization: Tailoring the model to specific industry or domain needs is more manageable with open-source LLMs. Organizations can fine-tune the model to suit their unique requirements.
- Active community support: Open-source projects often have thriving communities of developers and experts. It means quicker issue resolution, access to helpful resources, and a collaborative environment for problem-solving.
- Fosters innovation: Open-source LLMs encourage innovation by enabling organizations to experiment and build upon existing models. Startups, in particular, can leverage these models as a foundation for creative and unique applications.
Low-Barrier Efficiency
An example of an AI application that is easy to implement and can provide tremendous value is content generation, which Fava highly recommends. If an advisor has a monthly newsletter, they can enter the parameters into a generative AI tool and it will generate a list of ideas that your target client segment would be interested in. You can then ask the AI to write a blog post in the style of a popular author, The Morning Brew, for example, and receive completed content. Most advisors report that they struggle to plan and create regular content and how to market themselves, and AI can be a very powerful tool to increase their reach and productivity, she observed.
For a long time, the rule of thumb has remained that an advisor can handle around 150 clients, but the power of AI is quickly expanding that limit. Fava predicts that while change won’t be overnight, we will start to see an increase in the number of clients a single advisor can handle.
“It isn’t about replacing advisors, but augmenting them,” Fava emphasized. AI will allow them to push out more content, spend more time in face to face meetings, and analyze their book better to identify additional revenue opportunities.
Just by using the tools already available on the market, advisors can automate a considerable amount of their workflow to increase their efficiency. Fava also highlighted this as a great opportunity for the consultant market, which can come in and teach firms how to effectively harness AI. In the course of a few months, smart and efficient AI use can completely change advisors’ workflow without a high cost or barrier to entry.
Don’t Rush into AI
At this point, AI is just another capability in your tech stack. Few firms have the finances and resources to invest in private GPT systems or full-scale AI solutions, so the best way to start incorporating the technology is with a conservative, long-term approach. By starting with existing applications that have added AI-powered features, firms have a much greater chance of successfully and safely integrating AI into their everyday workflows.
