How it works

The 8-beat operating model

The same systems you run today — made to work as one, starting safely at the edge.

Beat 1 · Why RIAs

The pressure is operational

RIA economics reward capacity: revenue scales with assets, clients expect more, and supervised firms already know how to review work before it leaves the building. CloudLink starts from that operating reality, not from a generic AI feature list.

Beat 2 · AI Ownership

Rent the engine. Own the harness.

A powerful AI model, on its own, is a brilliant new hire on their first morning: no access to your systems, no memory of your firm, no rulebook, and no one checking the work before it goes out the door. The model is the raw talent. What makes it useful — and safe — is everything built around it: a working model of your firm it can actually read, a clear definition of each job, the rule that a person approves anything that matters, and a permanent record of what it did. You rent the talent; you own everything that makes it trustworthy.

See the Ownership Line →

Beat 3 · Overcome Inertia

Start at the edge

Why firms stall

Almost every firm knows it should be doing this, and almost none have. The reason isn't doubt about AI — it's that the obvious version sounds like ripping out systems that work, betting the firm on this year's model, and handing your client data to someone new. So it stalls, because standing still feels safer than any of that.

The way through

You don't rebuild your firm to become AI-native. You stand up a working model of your firm alongside the systems you already run — reading from them, never replacing them, and never becoming the official record. It starts by only reading, at one safe seam, proves itself on low-risk work, and earns its way inward from there. Because it lives on the edge as a working copy, there's nothing to rip out — and nothing to unwind if you ever stop.

First posture

Read-first, low-risk work proves value before any deeper integration. The twin earns its way inward through evidence, not promises.

Beat 4 · Ontology & Architecture

A governed digital twin of your firm

The RIA Ontology

We connect to Microsoft 365, your CRM (Wealthbox/Redtail), portfolio/custodial (Orion/Addepar, Schwab/Fidelity), and planning (eMoney/MoneyGuidePro), and express them as one continuously-synced ontology of your firm: typed objects (households, accounts, positions, communications), governed relationships between them, and the actions each permits — with data sensitivity and source-of-truth rules carried in the types. The ontology is not documentation: it ships as a versioned artifact verified in CI on every change — AI prompts are assembled from it (version-pinned and recorded), the relationship graph is directly queryable, and where a source asserts identity rather than proving it — a custodian file’s account number — the ontology types the difference and records how each claim was resolved. Your existing systems remain the authoritative record.

Agents that do the busywork

Reusable patterns — gather & summarize, draft for review, reconcile & flag, generate from template, monitor & alert, orchestrate across systems — handle the work between human judgments.

A complete audit trail

Every action, draft, approval, and decision is recorded in a tamper-evident, exportable, examiner-ready log.

See the public architecture →

Beat 5 · Governed by Construction

Governed by construction

Human-in-the-loop is the default everywhere. Autonomy is earned per workflow, through measured evaluations — never assumed. Client-facing work always carries a named human approver — the author of record — on a tamper-evident audit trail.

The trust anchor is simple: a named human at your firm approves everything client-facing. See the governance model →

How a number earns the right to be said

Modeled

Built from cited assumptions and conservative calculator inputs.

Sim-measured

Measured on the living firm — a fully simulated advisory firm whose days actually advance.

Pilot-measured

Measured in your environment against your baseline. This is the only rung we present as your results.

Modeled at your firm's numbers; demonstrated live on a simulated firm; measured in your pilot.

Beat 6 · ROI & the AI Graph

The AI Graph compounds

As you turn on more work, it all runs on one connected map of your firm — every household, account, document, task, approval, and result, linked and kept current. That map is the asset. Each new job you automate plugs into the same map, so the tenth costs a fraction of the first, and every approval your people make becomes information you own. The firm gets more capable — and more valuable — the longer it runs, instead of sprawling into another drawer of disconnected tools.

Run the ROI calculator →

Beat 7 · The Live Product

The living firm proves the mechanics

The living firm — a fully simulated advisory firm whose days actually advance: markets move, files arrive late, deadlines age.

Proof momentWhat it showsRung
GraphQueryRecord-by-record concentration and review evidenceSim-measured
Inbox triageMoney movement routed to a named human; never draft-eligibleSim-measured
Audit trailPrompt, draft, approval, and hash check in one chainSim-measured
Value MeterPer-task minutes, dollars, and approval outcome tied to assumptionsSim-measured

Beat 8 · How We Start

Three waves, lowest-risk first

WaveWorkflowsWhy first
W1 Quick winsMeeting notes & follow-ups, review prep, communications drafting, paperworkContained scope; fast proof of value; measured against your baseline
W2 Operational backboneOnboarding & account opening, billing, RMDsHigh impact; deeper integration after the first receipts
W3 Continuous & high-trustCompliance surveillance, governed rebalancingHighest trust boundary; earned through the six pilot gates

The six pilot gates

  1. Readiness — scope sources, data quality, workflow shortlist, and control posture.
  2. Connect — configure connectors read-first, with no write-back yet.
  3. Migrate — reconcile the twin to source systems before anything goes live.
  4. Provision — author workflows from patterns and prove them through evaluations.
  5. Dual-run — compare outputs beside your existing process for a cycle.
  6. Go-live — enable human-in-the-loop workflow operation with rollback tested.
Pricing is published-list, not quote-games: a fixed-fee readiness assessment that credits fully toward a build; a one-time conversion by buildout tier; an annual platform subscription with AI usage included — no token meters, no surprise bills. Founding build partners get materially better economics, earned back by results and candid feedback. If you want current numbers, we'll show you the list — and what founding partners pay against it.

Next step

See the numbers for your firm

Run your ROI, then let's scope a pilot.

Run your ROI →