“Do you guys do, like, general ledger stuff in here?” That’s a question everyone asks. Let that sit in your brain. We’ll come back to it.
In post three we made the Snowflake analogy: structured data sits in one canonical place, owned by the client, readable by any rendering tool. This post is about what that future actually looks like for a wealth owner — and why the vendor-lock-in model that has dominated wealth tech for fifteen years is about to break.
This is a future that benefits clients more than it benefits incumbent platforms. So it is being resisted in places where you’d expect it to be resisted. But it is coming anyway, because the technology underneath is coming anyway.
What “render anywhere” means in practice
Imagine a wealth owner — call her the principal of a $300M family office. Today, her financial life is fragmented across:
A reporting platform that shows public holdings. A separate tool for the alts. A QuickBooks instance for the operating company. A trust accounting system at the trust company. A CPA who maintains the version of the books that ties to the return. An estate attorney who has the most recent version of the trust structure. A bill-pay service. The principal’s own assistant, who has the master spreadsheet that ties it all together.
Each of these is a rendering — a view — on top of some subset of her financial reality. None of them is the canonical version. The CPA’s version usually wins in a tie, because the IRS sides with the CPA. But the CPA’s version is in working papers and isn’t queryable by anyone else. And it certainly isn’t updated real-time daily, weekly, etc., and it’s just a tax basis, not an investment basis as well.
Now imagine the same principal in 2030. Her financial data sits in a structured, permissioned data layer that she controls. Her CPA pulls from that layer to do the return. Her estate attorney pulls from that layer to do trust updates. Her wealth-management dashboard pulls from that layer for performance reporting. Her AI assistant — Claude, ChatGPT, Gemini, whatever — pulls from that layer to answer her questions. When her kids come of age, she grants them read access to their share of the structure. When she changes wealth managers, the data doesn’t move; the rendering changes.
Same data. Many renderings. All permissioned by her. None of them is the source of truth — the structured data layer is the source of truth, and everything else is a view.
That is the open-finance vision applied to wealth. And by the way — you can just have this now. Don’t wait until 2030, please. For your own sake.
Why this is technically possible now
Three reasons.
One — MCP. Anthropic’s Model Context Protocol, released in November 2024 and now hosted under the Linux Foundation’s Agentic AI Foundation, gives AI agents a standardized, permission-aware, auditable way to read from external data sources. OpenAI, Google, Microsoft, Bloomberg, AWS, and Cloudflare all back it. It is the closest thing the industry has produced to an HTTP for AI. Moody’s was an early adopter for credit and reference data. Bloomberg and other data vendors are in motion.
Two— the open-banking regulatory tailwind. The CFPB’s 1033 rule and the broader push toward consumer-owned financial data has shifted the legal and regulatory ground under the screen-scrapers and given clients a legitimate claim to their own data, machine-readable, on demand.
Three— the AI rendering layer. For the first time, a UHNW client can credibly say to a software vendor: “I want to ask my data questions in natural language, from any interface I choose.” If your platform doesn’t support that, your platform is going to lose to a platform that does. The “interface” advantage that proprietary dashboards have held is collapsing.
Stop trying to be the dashboard
Go get yourself the data lake. The one that can render anywhere.
If you like the UI at Addepar — great, render your own data lake you love there. If you like the Tamarac rendering — great, render your own data lake you love there. If you want a new one, build it. Or have a contractor build it. Or jank it up in some Excel file that auto-populates from it. Literally render it anywhere. But if you get that data lake first, then the point is you aren’t rendering it. The person who made you that data lake — yes, they can render it, anywhere, for you. And they can migrate it, anywhere, for you, anytime, in quick time, most likely.
This is the key — migration pain is why fintech is so hard to sell. But a data lake that renders anywhere lets you switch from render A to render B. Migrations, solved.
Who benefits when this works
Bucket one— the wealth owner. Lower switching costs, higher control, less key-person risk in the family office, fewer 11pm logins to GP portals.
Bucket two — the CPA. McKinsey’s research suggests AI can save advisors and accountants 20–30% of their time on technical planning work — but only if they have clean structured data to work from. Today they don’t. Tomorrow they can.
Bucket three— the estate lawyer. Today, drafting a trust amendment means asking the family to send over their latest schedule of assets, which is usually wrong or outdated by the time it arrives. Tomorrow, the lawyer pulls from a live permissioned feed. Less error, faster work, fewer surprises at funding.
Bucket four— the next generation. The single largest pain point in the great wealth transfer is that heirs don’t know what they’re inheriting. We’ll spend the whole next post on this. The render-anywhere model fixes the structural piece. Why not have your data lake present itself as PDFs to you on Wednesday at noon because that’s what you want, and your son’s permission him to a legacy UI dashboard he loves because it’s customizable? Both work. Both are great. Just get one lake.
Bucket five— even the incumbent platforms. The ones who lean into this get a much bigger market — they become the rendering of choice for people whose data is somewhere else, instead of fighting to hold the data hostage.
What still has to happen
Three things, none trivial.
One— the structured data layer has to exist for real. That is the work of post two, post four, and post six. Most of the industry hasn’t built it yet.
Two— permissioning standards have to mature. MCP is most of the way there for AI, but human-to-human permissioning (granting your CPA access to a subset of your data for tax season, for example) still needs UX and policy work.
Three— the legal frameworks need to catch up. Trustee fiduciary duties, advisor compliance, custodian data-sharing terms — all need updating to assume a render-anywhere world.
None of this is theoretical. All of it is being worked on right now, in public and in private, across the industry. This is exactly why you need AI × White Glove — there’s still a lot of White Glove needed.
And furthermore — you, $10 millionaire — why oh why oh why can’t you just have it too? These family offices have their own accountant running their IRRs. And you — you can too. Please don’t fight with empty-shelf software. Let’s get yourself a data lake and render it somewhere, anywhere. But don’t make the lake yourself. And don’t let it be an abstraction you can’t show your wife.
The reframe
The next decade of wealth tech belongs to firms that treat client data as the client’s data — structured, owned, permissioned, render-able anywhere. The walled-garden model produced fifteen years of decent businesses. The render-anywhere model produces the next thirty.
Permission your financial life. Render anywhere. The technology is here. The architecture is the next thing.
Now — back to that opening question. Do you guys do, like, general ledger stuff in here? So if the product is the hard work of the database, and everything — yes, that’s right, literally everything — gets in there, and if there are humans with actual real solid finance experience who know what they’re doing, and agents with the grunt ability to do an insane amount of work and pattern recognition, and if we tell you the entire industry, literally everything, is just a rendering… well, tell us — do you guys do, like, general ledger? Yeah. You want it by the top of the hour, or tomorrow, so it looks like it was harder?
But Real Work. Real Moat.