
Client
Job: ComprehensionEntry is one number and its story: total value, day’s gain, then allocation, accounts, movers. Depth on demand — Performance, Realized Gain & Loss, Positions, Transactions, Income Earned, Projected Income.
Case Study · Enterprise B2B SaaS
A White-Label Wealth Platform in 90 Days
A greenfield B2B enterprise SaaS for private wealth management — shaped in weekly working sessions with BlackRock’s team, built by Clearwater Analytics, and sold white-label to wealth firms like Chilton Trust. First wireframe to engineering handoff in ~3 months.
By Anubhaav Vermaa

Led by me, reporting to PMs and Director-level stakeholders at Clearwater Analytics and BlackRock.
Enterprise, annual subscription.
IA → wireframes → hi-fi → design system & tokens → dev QA.
June 2024 — Sept 2024.
Overview — The Idea
Private wealth management still runs on spreadsheets, emailed PDFs, and quarterly binders. This platform replaces that — a greenfield B2B enterprise SaaS built with BlackRock as design partner, engineered by Clearwater Analytics, and sold white-label to firms like Chilton Trust.
There was no product on the market that modeled the actual relationship at the center of wealth management — adviser on one side, client on the other, the same money between them. So every screen was argued into existence from scratch.
Wireframe to engineering handoff in three months.
The Spreadsheet Era
Advisory firms manage high-net-worth client relationships through spreadsheets, emailed PDFs, and quarterly binders. Nothing on the market modeled the actual relationship — adviser on one side, client on the other, the same wealth between them.
Clearwater Analytics set out to build that product from zero, with BlackRock as the anchor design partner — packaged as white-label enterprise SaaS and sold firm-to-firm to wealth and asset managers like Chilton Trust.
And it’s B2B at its most layered: the buyer is a firm, but the users — advisers and their clients — never chose the software. Every design decision had to survive people who didn’t opt in.
Before
The obvious references — Addepar, Orion, Black Diamond — are analytics and data engines. Strong at numbers, silent on the relationship. There was no incumbent to copy, which meant every convention had to be argued into existence.
Discovery by Proximity
No comparable product meant no teardown to lean on. Discovery ran on access instead: two to three working sessions a week with BlackRock’s advisory and product teams, for the entire engagement. Every flow was pressure-tested by people who had spent careers inside the workflow we were replacing. Design reviews doubled as research.
One client — many legal entities: IRAs, revocable and irrevocable trusts, joint accounts. The architecture has to roll up household → entity → account → position without losing anyone.
An adviser’s first question is “who needs me today?” A client’s is “how am I doing?” Same data, opposite geometry.
Statements get printed, exported, mailed. Print and download live permanently in the app bar — never buried in menus.
Two Users, Opposite Instincts

Entry is one number and its story: total value, day’s gain, then allocation, accounts, movers. Depth on demand — Performance, Realized Gain & Loss, Positions, Transactions, Income Earned, Projected Income.

Entry is the book, not a client: active and inactive users, new documents, new form responses. Signals first; any client one gesture away.
We didn’t design two products. We designed one product with two front doors — one data model, one component library, two entry geometries.
Signature Decision
An adviser who can open any client’s portfolio can also confuse one client for another. In wealth management that isn’t a usability bug — it’s a compliance incident. The answer had to be ambient: felt before it’s read.
The shell itself is dyed by role — two distinct environments, rendered green for clients and gold for advisers in the base theme. Peripheral vision does the compliance work before a single word is read. Under a buyer’s brand, the hues re-map to their palette at onboarding — the colors defer to the brand, the two-environment pattern doesn’t.

When an adviser views the product as a client, a persistent banner pins the truth to the top of the screen — “You���re currently viewing a preview from John Smith” — with the client switcher inside the banner and a one-click exit back to the adviser’s own view.

A switcher moves the adviser across every client without re-navigating the product. The interface answers “whose money is this?” before anyone thinks to ask.

Density Without Dread
Financial software fails in one of two ways: it hides the numbers, or it drowns you in them. The middle path is structural, not decorative.
Entity rows expand to asset classes, classes to positions — allocation, value, MTD/QTD/YTD/since-inception in one scan line. Built on MUI’s data grid, bent to the household model.

Every dense module ships with a chart/table toggle — charts persuade, tables convince. Time-weighted return is charted against the S&P 500 and Russell 2000 in place: performance is a comparison, not a number.

The same model re-set in card grammar; a year of transactions becomes a calendar heatmap you read in a thumb-scroll. One All-Accounts selector governs every module — a single mental model for slicing the constellation.

An All-Accounts selector governs every module — a single mental model for slicing the constellation.
Every dense module ships with a chart/table toggle. Charts persuade, tables convince.
A 12-month projected-income forecast — dividends, taxable and tax-exempt interest — turns a statement into a planning instrument.
The Loop I Closed
Mid-build, I flagged a dead end: the platform let a client see everything and act on nothing. Spot a position you want to sell — then what? Leave the product and go hunting for an email address?
I proposed and designed the adviser channel: text and call, embedded on web and mobile, connecting each client directly to their adviser. Insight → conversation → action, without leaving the platform.
It closed the product’s one open loop — the human one.

A System That Works With Any Brand
White-labeling gets treated as a paint job. It’s actually a governance problem: what may a buyer change, and what must survive them? The token architecture splits into two castes.
Primary palette auto-derived from the logo at onboarding, or ingested wholesale from the buyer’s brand guidelines: color, typography, lockups.
Grid, spacing, elevation, component behavior, information architecture — the framework itself, sealed at onboarding. Buyers own every hue; they never touch the structure.
Onboarding is the theming event: logo in → palette derived → guideline override if supplied → framework sealed. After onboarding, buyers restyle color — never structure.
The lock isn’t aesthetic conservatism — it’s the business model. One immutable framework is what lets a single team ship updates and security patches to every branded instance at once, on one annual subscription.
Iterate Where Iteration Is Cheap
Ninety days, greenfield, a three-designer team. The budget math was simple: spend nothing inventing primitives, everything on architecture.
Material Design + MUI as substrate. Data grids and financial charting are the two hardest things to build well fast, so we didn’t build them — we bent them.
The countless rounds happened in wireframes, validated inside the weekly stakeholder cadence — so high-fidelity ran close to single-pass. Residual iteration tracked shifting requirements, not indecision.
The mandate was explicitly minimalist: an instrument, not an experience. Every pixel of expressiveness deferred to the buyer’s brand anyway.
Full token set and documented design system to engineering — then I stayed embedded through the build to hold the line on what shipped.
What I Know, and What I Don’t
The contract ended at development completion; post-launch numbers belong to the client. What I can vouch for: a two-persona, N-brand platform designed from zero, delivered to engineering inside the window, on a token architecture built to outlive the team that made it.
Contact