For independent advisory firms
AI doesn't fix broken data. It scales it.
We make an advisory firm’s data fit to automate — so the tools you already pay for produce something you’d put in front of a client.
Independent RIAs, roughly $100M–$1B AUM, 2–25 people — firms with real operational load and no in-house engineer.
The gap
The tools work. The data underneath them usually doesn’t.
Meeting assistants, notetakers, Copilot — they’re good now. Every one of them assumes the data underneath is clean, complete, and consistent. In most firms it isn’t.
So the output comes back generic, or confidently wrong, and someone rewrites it by hand. Nobody sells the part that fixes this: vendors sell software that starts working on day one, not the unglamorous work that has to happen first for it to be worth anything.
What we build
Five things, in the order they usually matter
-
CRM data health and remediation
A read-only profile of every client record: what’s inconsistent, what’s duplicated, what’s missing, and which fields are quietly recording the same thing in four different places. You get the findings and a ranked worklist.
-
Systems of record that are actually authoritative
Deciding which system owns which fact, then enforcing it. Most firms hold the same information in three places, disagreeing, with no rule about which one wins.
-
CRM to Outlook and Microsoft 365
Client contacts in every advisor’s mailbox and on their phone, kept current automatically. One-way, CRM stays the system of record, every run logged and auditable.
-
Connecting AI to your firm’s data
We build on MCP, the open standard for connecting AI assistants to your systems. One connector works with whichever assistant your firm already uses — Claude, ChatGPT, Microsoft 365 Copilot — and runs inside your own cloud tenant, so client data never leaves your control. Change assistants later and the integration still works.
-
The integration nobody will ever sell you
Vendors integrate with each other’s mainstream platforms. They will never integrate with your homegrown database, your SharePoint site, or the spreadsheet that runs a process no one else has. Someone has to build those by hand.
Proof
What this looks like in practice
1,953
records profiled
9.5%
had an advisor assigned
4
fields recording “advisor”
We profiled all 1,953 contact records at an SEC-registered RIA — read-only, no client data stored. Their data was structurally sound, but an advisor was assigned on 9.5% of client records, with “advisor” recorded in four different fields depending on who entered it. Any automation built on that would have produced confident, wrong output.
We handed them the findings and the worklist, then built the thing the data was for: a scheduled Wealthbox to Outlook contact sync running in the firm’s own Azure tenant, with advisor routing, incremental updates, failure handling, and a no-client-data-in-logs guarantee enforced by automated tests.
Phased and billed hourly, in small pieces, each delivered before the next was scoped.
How we work
Small first
When you need wiper blades you don’t need four tires and a brake job. We scope the smallest useful thing, ship it, and earn the next one.
Your cloud, your data
We deploy into your Microsoft or Azure tenant. You own the infrastructure and the credentials. Nothing routes through us or a third-party vendor cloud.
Auditable by default
Every run produces a structured log of what was read, written, failed, and skipped. For anything touching client records, “it worked” isn’t good enough.
Fails loudly
When something breaks we surface it. We don’t insert a fallback that makes a broken integration look healthy.
Where we fit
- We integrate the platforms you’ve already chosen. We don’t resell them and take no referral fees, so our recommendation is never for sale.
- We move and organize data so every number traces back to its source. What it means for a client stays your judgment — we don’t automate advice.
- We work in defined engagements with a visible end. If what you need is a multi-quarter rebuild, we’ll tell you on the first call.
- Wealthbox
- Redtail
- Orion
- eMoney
- MCP
- Claude
- ChatGPT
- Microsoft 365
- Copilot
- Graph API
- Azure
- C# / .NET
- PostgreSQL
Where we’d start
A read-only data health check
We profile every client record through your CRM’s API — nothing changed, no client data stored — and hand you a written findings document: what’s structurally sound, what isn’t, and a worklist ranked by what it would unlock. You keep the findings whether or not we do anything else.
It’s the cheapest way to find out whether we’re any good.
No pitch deck, no sales process. 30 minutes to work out whether this is worth doing.