The biggest AI launch wealth management has seen in years happened Monday at a festival on a beach, which means it was witnessed largely by advisors in shorts.
Anthropic introduced Claude for Financial Advisors at Future Proof in Huntington Beach, with Schwab, BlackRock, Vanguard and a dozen other heavyweights signed on as partners, which is the wealth management equivalent of getting the whole family to agree on a restaurant.
The industry spent the rest of the week arguing about what it means, and on Friday one of the sharpest people in wealthtech published the response worth reading, imo.
This week's three stories are one story told three ways. What launched, what the skeptics say, and the move that protects you either way.
Anthropic Launched Claude for Financial Advisors With Half the Industry on Stage.
On Monday at Future Proof, Anthropic rolled out Claude for Financial Advisors, a plugin that bundles two things. Connectors link Claude to the platforms advisors already use, including Charles Schwab, BlackRock, Vanguard, Addepar, Envestnet, iCapital, Orion, SS&C Black Diamond, Wealthbox, Wealth.com and Zocks.
Skills then put that connected data to work on specific jobs, with eight at launch covering pre-meeting prep, post-meeting notes and follow-up, prospect intake, portfolio rebalance review, estate and tax briefing, alternative investments briefing, compliance and AI policy review, and advisor onboarding, which covers most of an advisor's week. Anthropic built the launch around a Kitces figure you've probably lived, that a typical practice spends only about a sixth of its time actually meeting with clients.
The announcement carried quotes from 18 industry executives, which is more blurbs than a paperback, and Dynasty relaunched its Dynasty AI platform on Claude models the same day. On stage, Dynasty CEO Shirl Penney asked a beach full of advisors how many of them got into the business for the back-office paperwork, then called it the biggest moment he has seen in his career in the independent space.
Why you should care: The biggest names in AI and the biggest names in wealth management just agreed, in public, that the advisor's admin hours are the problem worth solving. That validates the direction plenty of you have been building toward for a year or more.
Now the practical detail before you go looking for the install button. The advisor plugin is available on Claude Enterprise plans, which come with audit logs for recordkeeping, so this launch arrives first at large firms and platforms, the way most nice things do.
For a solo or small RIA, the launch is a preview of the pattern rather than a product you'll turn on this week. Connected data plus repeatable skills plus your judgment on top is a great solution to the problem. And you can run a version of that pattern today on the plan you already pay for, with your own files standing in for the connectors, and this week's exercise below builds exactly that.
The Skeptics Have One Good Question. Can It Get Specific Enough?
By Tuesday, WealthManagement had rounded up reactions from analysts and technology providers, and it's full of debate.
Longtime analyst Alois Pirker likes the signal but predicts Claude will run into the same wall Salesforce hit with its financial services rollout, which needed a decade and a rebrand to find its footing, reaching the level of specificity advisors actually need. His example is a firm like LPL, where 30,000 advisors might include 15 genuinely different kinds of practice, each needing different workflows.
Consultant Will Trout pointed out that the one-sixth-of-time-with-clients figure has been true for years, so the test is whether Claude actually moves the number. And Anthropic's Peter Nolan answered the competitive question directly, saying "We aren't trying to replace any of the tools out there" and describing Claude as a conductor helping an advisor's average of roughly seven tools work together.
Anyone who has watched their CRM and their planning software refuse to speak to each other knows how ambitious that job is.
Why you should care: Pirker's point deserves a spot in how you think about every AI tool, because generic and specific is the battle. A model that knows what financial advisors do in general still doesn't know what your practice does in particular, your niche, your service calendar, your compliance rules and the way you explain a rebalance to a nervous client.
Big platforms will spend years closing that distance for enterprises. You can start closing it for your own practice in an afternoon, because the specificity Pirker is describing comes from context, and you own all of your context already.
That's also the real advantage of a small firm. Nobody needs to build 15 workflow variations for you. There's one of you, and you already know exactly how you work.
Jud Mackrill's Response: Whoever Owns the Data Layer Owns the Decade.
On Friday, Jud Mackrill of Milemarker published the smartest reaction to the launch, a piece titled You Don't Need Another Integration. You Need Your Own.
He gives Anthropic real credit, writing that of every major AI company it has shown the most genuine interest in how advisors actually work. Then he raises the point the press releases skip. Connectors show you the slice of your business each vendor chose to expose, on their timeline, at the depth they picked. The things that actually run a firm, the bank, the legacy system nobody will rip out and the spreadsheet holding the truth about your top 20 households, and you know exactly which spreadsheet I mean, aren't on anyone's connector list.
His answer is owning your own data layer, complete and reconciled and under your control, so any model can plug into it. Because if your intelligence lives inside one vendor's product, "their roadmap becomes your roadmap." Worth noting that Milemarker sells exactly this kind of data layer to firms, so yes, he has a horse in this race. But the argument holds up even without the horse, plus Jud is brilliant and has been in the industry for quite some time.
Why you should care: Jud is writing about enterprise data infrastructure, and the principle scales all the way down to a practice of one. The solo version of owning your layer is owning your files. A voice guide, a plain-words compliance profile, a niche description and your client segment notes, kept as documents in storage you control, work in Claude today and in whatever model is best next year.
The advisors in the strongest position after this week's launch are set up so the model is swappable and the context is theirs. That has been the architecture behind everything Amplify for Advisors teaches, and it was nice to watch one of the most respected builders in wealthtech make the same case to the enterprise crowd. When the tools change fast, the durable asset is the part only you can write.
ONE THING TO TRY THIS WEEK
Start your own layer. Not the enterprise system Jud is describing, that takes real infrastructure, but the blueprint of it, which no vendor can build for you anyway because only you know where your practice's truth actually lives. About 30 minutes.
Step 1. Open Claude or ChatGPT and paste this.
I'm a financial advisor mapping where the important information in my practice actually lives, so I can start owning my own data layer. Interview me one question at a time, waiting for each answer. Ask about every place practice and client information lives, my CRM, custodians, planning software, email, spreadsheets, note tools, paper files, and what lives only in my memory. For each place, ask what truth exists only there. Then ask about my niche and how I serve clients, so you understand what information matters most in my practice. Work entirely with system names and categories of information. Do not ask for and do not let me share any actual client names, account numbers, or identifying details, and stop me if I start.
When the interview is done, produce three things. First, a Practice Data Map listing each system and the information that lives only there, with a flagged section for anything that exists only in my head or only in one fragile place. Second, a Source of Record table assigning each category of information one official home. Third, a Household Master Record template, a one-page field list I could fill out for any household covering relationships, goals, key dates, service history, preferences, and connections, tuned to my niche. Treat everything as a first draft for my review.
Step 2. Read the Data Map and go straight to the flagged section. Every advisor who does this finds something important that lives nowhere but their own memory, and seeing it on paper is the whole reason to do the exercise. Your memory is not a system, and now you know exactly what needs a real home.
Step 3. Open your own spreadsheet, away from any AI tool, and fill out the Household Master Record for the one household you know best. Time it. If gathering the truth about your best relationship takes 20 minutes across four systems and your own recollection, you've just felt Jud's argument in your own practice. That template, filled out household by household on your own time, becomes the spreadsheet that actually runs your firm, except now it's deliberate instead of accidental.
Sam Farrington, CFP®
Want the prompts and frameworks that turn this news into action for your practice? That's what Amplify for Advisors is for. New frameworks every Tuesday and Friday.
Explore more at amplifyforadvisors.ai