I spent most of last week at a conference in Denver so, if you’ll indulge me, I’ll share some of the details with you.
This conference was the Insider’s Forum that has for years been hosted by a long-time thought leader in the planning industry. One of the draws for this conference is how different it is from others I’ve attended. More typical conferences seem full of actively managed mutual fund providers and insurance companies touting their product enhancements and calling it continuing education. The Insider’s Forum has none of that. Instead, it’s a conference built for planning and investment management firms who are fiduciaries for their clients, and not salespeople schlepping investment products. The continuing ed is meaningful and geared toward firm owners, advisors and planners, and the operations folks who support them.
On purpose, the conference doesn’t have mutual funds and insurance companies as sponsors. The exhibit hall is mostly industry-specific tech companies, peppered with custodians and other service providers. It seemed like at least half of those were AI-based, which was a large increase from last year’s conference where there were maybe two or three that emphasized AI notetaking (of emails, phone calls, Zoom meetings, and so forth).
AI notetakers are now considered old school as transcription technology seems to be a standard offering. This year there was lots of talk about new AI agents and pulling in client data from across most (or at least from as many parts as possible) of our firm’s “tech stack” and allowing the agents to read and synthesize the information into a variety of service-related tasks. Like drafting meeting notes and emails to clients, auto-creating tasks, even marketing. I was surprised to learn, for example, that a number of these providers could link to my encrypted document management tool and be able to read all (probably) of the documents I have on file and answer questions about them. Some of these platforms are purpose built for my industry, while at least one company was selling the ability to create custom agents. I first heard the term “silicon staffer” at last year’s conference and we seem much closer to that this year.
That’s exciting in a general sense, but it was interesting to see how the race to get all this done is shaping up in my industry as a proxy for what’s happening across the broader economy. For example, the Holy Grail seems to be connecting an AI-based service provider to literally every software package we use when working with clients. This could be a dozen or more systems. Then we’d access one system to access all. That’s hard to do in practice because of legacy computer systems at many companies not being up to the task, plus individual vendor policies regarding data security and privacy. And those are just a few of the issues. What we seem to have are a bunch of new and interesting companies competing with each other to offer similar services while facing structural limitations that limit how much efficiency we can actually gain from AI. There’s still a ton of capability, of course, just not as much as one would expect with all the fancy marketing, at least at this point.
As I’ve mentioned a number of times in recent months, I’m taking a wait-and-see approach to letting third-party AI platforms access your data. This is getting challenging because big custodians are also getting into the game.
Schwab announced just prior to the conference that they’re linking up with Anthropic to allow people like me to use Claude when working with your data in the Schwab system. Schwab hasn’t released a lot of details yet, but interviews in the trade press suggest this requires an opt-in on my part. Also, Schwab won’t allow Anthropic to store client information or use it to train models, it’s in a read-only format with masked account numbers, and data security would conform to existing regulations, which are pretty strict.
[Update... while Schwab hasn't released specifics, I watched Anthropic's webinar from a few days ago detailing Claude for Financial Advisors, during which they discussed the Schwab integration among other things. How this works is I would need an enterprise subscription to Claude and that links to Schwab, so it's not automatic. The enterprise version comes with various controls like I mentioned above. It also includes other compliance-related tools that aren't available with a basic paid subscription to Claude. It's early days for this and I'm researching options, but I'll likely wait and watch on this well... more to come in future posts, I'm sure.]
And it’s not just Schwab. At about the same time, Vanguard announced its acquisition of a smaller custodian, Altruist. Among other things, Altruist is known for having the industry’s first internal AI agent, a highly regarded tool called Hazel. Presumably that was part of the reason Vanguard bought company, but that’s only a guess.
So, firms across my industry are absolutely “on the AI train”, as I heard several people say at the conference.
There were numerous mentions at the conference about the industry establishing guardrails around how we let AI models use client data. But there seemed to be a lack of clarity about how guardrails would actually work and who holds the ultimate responsibility. To me, it seems like a lot of companies are rushing into getting AI into their systems and workflows but the number of questions around all this will grow along with the technology.
The “AI Doomsday” question came up in various forms. While those questions weren’t downplayed, they’re certainly hard to answer probably because they’re impossible to answer. It seems to be taken for granted that AI technology will continue to be revolutionary, and that it’s only a matter of time before AI is taken mainstream, regardless of the risks.
All that may be so but, again, I’m proceeding with caution and am learning as much as I can in the meantime. I like technology at least as much as the average person and I don’t fear it. Still, all this is moving so quickly that caution is important, especially when it comes to my responsibility for your information. I never want to let the allure of efficiency put that at risk.
Have questions? Ask us. We can help.
