An AI company just told financial advisers their problem isn’t capability.
It’s that nobody uses the stack they already pay for. Their word, not mine: utilisation.
Anthropic launched Claude for Financial Advisors on 14 September. Nearly twenty partners named in the announcement, most of them with a quote attached: Schwab, BlackRock, Vanguard, Addepar, Envestnet, Orion, Salesforce, Wealthbox. Plenty of names you’d expect.
I read all the quotes. Six of them say a version of the same thing, and it isn’t what I expected a product launch to say.
“AI is most effective for advisors when it works with the data and systems they already use.” That’s Steve Leivent at SS&C Black Diamond.
“Claude for Financial Advisors can move that busy work forward across the systems we already use.” Michael Batnick at Ritholtz Wealth Management.
Peter Nolan, who heads asset and wealth management at Anthropic, put it most plainly in an interview with WealthManagement.com: “Our goal is to drive utilization in the advisor stack today.”
Utilisation. Not capability.
The one party in this story who could credibly have said “here is something your software cannot do” said instead that the problem is you aren’t using the software you already have.
I think they’re right. I also think the reason is less flattering than “people are bad at adopting tools”, so it’s worth going a layer down.
Where the work actually lives
Nothing changes when you add a tool, because capability was never the constraint.
Every system in a practice gets used for the one job it was bought for. The CRM holds client details. The planning tool holds plans. The platform holds the account. Each of them works exactly as sold.
What nobody owns is the space between them, and that space is where a startling amount of the actual work lives. Someone finds the information, checks what’s missing, updates another system and chases the next person. It takes real time. Almost none of it shows up as a process anyone deliberately designed.
Andrew Stavaridis at Envestnet gave me the word for it: “Every hour an advisor spends assembling information is an hour not spent with a client.”
Assembling. Not analysing, not deciding, not advising. Moving things from one system into another by hand, because the two systems were never properly introduced.
The release also cites Kitces research: a typical advisory practice spends only a sixth of its time in client meetings. There’s necessary work in the other five sixths, obviously. What I’d want to know is how much of it is preparation, and how much is assembling information that already exists somewhere in the building.
That’s usually where I start looking.
Before blaming the team
The easy explanation is that people resist change, or that the team isn’t disciplined. I’ve been in enough sales teams to be suspicious of that explanation by default. Before deciding someone is the problem, I want to see what they’re working around.
Here’s what I usually find instead. The tool was bought to solve a symptom someone could name in a meeting. Nobody had written down how the work actually flows, so the tool got configured around that one visible symptom, and everything on either side of it stayed manual. Then the manual parts got absorbed by whoever was closest. Usually one person, usually quietly.
The business carries on, because that person remembers what needs to happen. Give it enough time and everyone starts treating their memory and their follow-up as part of the system.
Nothing broke. Nothing had been built.
Josh Brown of Ritholtz put it bluntly in the release: “I don’t want my CFPs spending hours every week laboring over CRM updates and task assignments.” That’s a very expensive person doing data entry, inside a firm that owns good software.
The part I’ve actually watched
I worked with a financial services business here in South Africa where underwriting was taking 40 to 90 days and investment transfers two to three months. A meaningful number of clients cancelled somewhere in the middle, and every one of those got written off as bad luck.
It wasn’t bad luck, and it wasn’t the software. The process was reactive by design. Everything was handled when it landed instead of anticipated before it landed, so the elapsed time had almost nothing to do with the work and almost everything to do with waiting.
Underwriting came down to around a month. Transfers to four to six weeks. Retention, once tracked, held at 95%. That was across thirty-plus independent advisers who had never worked to a shared process, and who didn’t report to me.
None of it was a tool decision. It was agreeing what happened next, and who owned the gap.
What I’d want to know before buying it
I haven’t used this product, so I’m not going to tell you whether it’s good. Two things I’d want to test, though.
First: what’s left for a person to do?
The release is clear that this writes back into the systems. Administrative actions like CRM updates and draft client communications go to the adviser for approval, and investment recommendations, client communications and compliance determinations stay subject to human review. That’s the right design, and approving a drafted update is genuinely cheaper than writing one.
It still leaves a queue, and someone has to clear it. In every practice I’ve looked at, the queue nobody owns is exactly where work goes to sit.
Second: what happens in month three?
Month one is when it’s new and somebody is still excited about it. By month three I want to know whether the team still opens it, whether the records stay current, and whether one person has quietly gone back to filling the gaps.
Asking a product launch to explain month three is not a normal reaction to a product launch. It’s the only question I ever actually want answered.
What I don’t know yet is whether utilisation is a thing anyone can sell. Every firm I’ve worked with agreed the problem sat between their systems, and then bought another system anyway. A system is something you can put on a purchase order. A decision about who owns the gap is not.
That’s the part I’m still figuring out.
References
- Anthropic, “Claude for Financial Advisors”, 14 September 2026. The partner quotes from Steve Leivent, Michael Batnick, Andrew Stavaridis and Josh Brown, and the approval and review design.
- Davis Janowski, “Anthropic Launches Claude for Financial Advisors”, WealthManagement.com, 14 September 2026. Peter Nolan’s quote.
- Kitces Research, “How Financial Planners Actually Do Financial Planning”. The sixth of a practice’s time spent in client meetings.
- Case 01: financial adviser operations. The underwriting, transfer and retention work.