“AI is the new electricity.”
— Andrew Ng, co-founder of Coursera and artificial intelligence pioneer
Picture this: your financial advisor, supercharged with AI smarts. That’s exactly what was on the table at the MarketCounsel Summit 2023, where big names in advisor technology gathered to spill the tea on AI’s future role in wealth management. We’re not just talking predictions here; this is AI that will be able to craft savvy content and business strategies like a pro.
Leading the charge was Joel Bruckenstein, President of T3 Technology Tools for Today, who, along with his panel of tech aces, dove into this brave new world. They’re not sitting around waiting for the future – they’re already weaving AI into the fabric of financial planning, from analyzing portfolios to penning proposals. Think of it as having a financial wizard by your side. This article is your ticket to the front row of their groundbreaking discussion, so get ready – the realm of wealth management is gearing up for an exciting shake-up.
The panel included:
- Brian McLaughlin, President at Orion Advisor Technology
- Oleg Tishkevich, CEO at Invent
- Craig Iskowitz, CEO at Ezra Group
Generative AI Basics
“Generative AI is very different from the types of AI we used to talk about a few years ago,” McLaughlin began. The old AI was mostly predictive, guessing at the next big book, movie, or trend.
But when ChatGPT entered the scene in November 2022, it opened the door for AI to create more than just insights, but new content. Its model was trained on billions of data points encompassing other people’s experiences, other datasets, and information from all across the internet. The generative model can summarize videos, text, or audio, it can write articles, build spreadsheets and even code.
McLaughlin noted that Orion has incorporated generative AI into their tools to assist with tasks such as portfolio comparisons, where it can draw data straight from the market and determine which set of stocks is better, and even use the historical data to write proposals.
While AI offers vast powers, it still requires practice to incorporate into your workflow. “You have to get into the habit of using it,” Bruckenstein advised. “I forget the tools are there, but whenever I use them, they’re helpful,” he said
In fact, Bruckenstein admitted that he turned to ChatGPT to help come up with questions to ask the panelists. Tishkevich emphasized that users still need to get better at asking AI the right questions and leveraging it properly, and even then cannot go so far as to create behavioral change.
As a consultant, Iskowitz encourages all firms to consider incorporating AI in some form into their tech stack, and often tells clients to play around with a free version like ChatGPT or Bard first to get a feel for them. “From data we’ve seen across industries, generative AI creates a 20-30% increase in efficiency for white-collar jobs,” he reported.
These are a Few of My Favorite Tools
Everyone has been trying new AI tools, and everyone has a different favorite. With so many applications popping onto the market, it’s really just a matter of finding what works for you.
In fact, you might not even have to go looking for any new apps at all to take advantage of the latest AI. Ezra Group has identified the trend of fintech apps launching their own AI-based features through the Kitces-Ezra Group Advisor Tech Map. There’s a good chance that software you’re using already has generative AI functions built into them.
Iskowitz highlighted a tool called Pulse360, which provides client meeting automation and advisor compliance. They were ahead of the curve in launching a generative AI-based tool, called AI Writer, before ChatGPT hit the market. Their tool allows advisors to generate meeting summaries, questions for clients, and meeting agendas within their existing workflow. “Every single part of client interaction can be automated in some way– with human oversight– by these generative AI tools,” Iskowitz explained.
When it comes to content generation, Mclaughlin recommended treating an AI’s version like a first rough draft. It’s not what you submit, but hopefully, you only need to change a little bit and it’ll be good to go.
Tishkevich pointed to tools that can detect the emotions of the person you’re speaking to and provide you with instant feedback with how they’re reacting. There is even an AI trading app that can parse the tone of information presented, and include the way that people are talking about a stock online into its overall calculations.
Bruckenstein mentioned that Microsoft which has embedded a number of AI tools into their Office suite, including one which can generate PowerPoint slides, design and content included.
McLaughlin’s favorite tool is Notion, a notes tool which has AI functionality. It can convert meeting notes into a list of actionable steps, create a project, add events to the calendar, cutting down steps throughout a user’s workflow. He also highlighted an app called Superhuman which generates a summary of received emails so you don’t have to read through them.
Beyond individual tools and small efficiencies, Iskowitz sees next-best-actions as the place AI can really make an impact. Morgan Stanley was one of the first firms to adopt a next-best-actions system for their 16,000 financial advisors back in 2018. It is both a platform for personalized communication and engagement with clients, and an AI-based recommendation engine for investment and wealth management ideas that FAs can present to their clients. The firm’s data has shown that advisors using the tool grow their books quicker and have more engaged clients.
Similar examples of this type of technology are already permeating around the edges, through tech like Catchlight, an AI-powered prospecting tool from Fidelity Labs. Catchlight organizes an advisor’s prospect list as a pipeline based on a ‘financial complexity score.’ Many new tools create a feedback loop on their recommendations, following up with advisors to rate and comment on the directions they were given so the algorithm can improve.
This will make advisors exponentially more efficient so they can better grow their business and keep their clients happy, but it will likely have the unfortunate side effect of turning financial advisors into “highly paid Uber drivers as they follow exactly what the app is telling them to do,” Iskowitz warned.
Trouble in Paradata
One of the biggest obstacles standing between advisory firms and AI solutions is around data, especially for smaller firms. McLaughlin suggested the use of Amazon’s Bedrock LLM which is a foundational model for AI language learning and can get a firm around 80% of the way to their own private version of ChatGPT. They would need to inject their own data on top of the model and it becomes a siloed, independent LLM.
McLaughlin then walked the audience through a new portfolio comparison tool Orion recently launched., It takes two sets of data based on a large sample of portfolios and provides an analysis with an AI generated narrative around it. The challenge of the tool, Mclaughlin explained, is that it doesn’t factor in the nuances of the advisor-client relationship. “We told them we needed a more narrow view on the LLM and they said, ‘so you want us to make 100,000 of them? And we said yes, we want one for everybody,” he recalled.
There’s obviously some privacy concerns, because Orion is trying to make a data set specific to each individual client. AI learning models and GDPR might not work together yet, but in Mclaughlin’s view everything in technology is solvable, it’s just a matter of time. And with the speed AI technology is advancing at, it might not be very much time at all.
On the opposite end of the scale, perils abound for large companies as well. With a myriad of sectors and advisors, they run the risk of an AI returning certain information that is not pertinent to that particular client or firm, Tishkevich warned. There are no rules about how you can train your system. And while they could have individual advisors block information off, they then will likely end up with a data set that is too small, which means the output would be questionable.
Trials and Regulations
Bruckenstein recounted a recent conversation while visiting Schwab, where they disclosed that they won’t be using AI for financial planning or portfolio management due to regulatory reasons. Any investment that they make in AI for the immediate future will be focused on the client service side, which has fewer compliance issues.
Mclaughlin likewise voiced hesitation at the idea of embedding generative AI into your company just yet, especially for client-facing content generation where regulation is likely to still change. However, he regarded internal next-best-action tools and trend identification to be safer areas.
“There’s always risk,” Iskowitz contended. “The question is how you mitigate it.” Similarly to Mclaughlin, he suggested that the use of private LLMs could result in more personalized and more secure AI usage. In fact, he predicted that within 3-4 years, every company will have their own private ChatGPT that is trained on all internal data.
While it would be difficult to build now, new tools like Amazon Q are resetting what we consider to be AI table stakes. Q is a private LLM that can be trained on different internal data from Salesforce, Google Drive, and other sources to become a powerhouse expert on your company.
As Bruckenstein pointed out, the amount of data required to build a private LLM provides large firms with a clear advantage. But Iskowitz argued that giants like Microsoft, Google, and AWS are sure to provide affordable tools that can open up those capabilities to smaller players.
Moving Up the Value Chain
So how do advisors avoid the trap of technology making their jobs both easier and worse? (Or, fatally, redundant?)
By moving up the value chain.
When robo-advisors came in, the human advisors who couldn’t move beyond building baskets of ETFs went out of business– but the ones that were able to provide other value to their clients took up those users and became much more successful. As Tishkevich put it, the technology “allows advisors to be more strategic and focus on other things.” But that means they have to work towards those bigger-picture ideas, or they’ll likely sink.
“An advisor’s value proposition isn’t portfolio building,” Mclaughlin chided. “It’s understanding their client’s needs, wants, and desires.” The one thing a computer can’t know is the relationship of the client to the advisor, and to their finances.
Bruckenstein lamented that advisors will be expected to provide more services for the same price as AI takes over more basic tasks. In order to scale their businesses, advisors will need to automate wherever possible and reserve human efforts for where they’ll add the most value. “I think the only way we grow ourselves out of this is good technology,” he predicted. “I don’t see any other solution.”
Like Bruckenstein, Mclaughlin described AI as a powerful amplifier for those human skills that makes them scalable. The only way to understand and build those close relationships 500 times over is through AI solutions.
In fact, Iskowitz doesn’t foresee any particular jobs being eliminated just yet. “We would never advise any firms to fire anyone to be replaced by a new technology,” he cautioned. “That’s just not good business practice.” But bringing the technology in does free up employees to spend more time on higher level tasks, thinking about strategy and business growth rather than rote tasks.
The area that will be seeing the most immediate change is customer service, Iskowitz predicted. A general AI-based knowledge system for customer service has been shown to dramatically improve response time and is particularly potent for internal support. Internal help desks have seen 80% drops in level one requests when a chatbot is used to answer advisor questions, Iskowitz added.
Ideally companies wouldn’t be removing those employees, he advised, but rather moving them up the value chain to do more impactful work. Tishkevich expanded on this, noting that by hiring more financial professionals instead of focusing on operations will enable firms to service more clients. Financial advice will hopefully become cheaper to provide, and therefore more accessible to a wider range of Americans.
In The Claws of Innovation
Adopting generative AI will force wealth management firms to not only rethink their tech stack, but their wider infrastructure– not to mention their relationship to clients and their place in the market. Spending time researching and experimenting with new Ai tools can give you an edge, shaving precious time off your advisors’ workflows; while adopting next-best-action tools can keep them driving the most efficient track.
Carefully calculating risks and finding the best implementation for your firm is always worth the effort to avoid falling into any regulatory or technological traps. And more than anything else, you need to constantly reassess the ways that you can both harness AI and offer value beyond it (unless of course, dear reader, you yourself are a bot. In that case, carry on.).