Global Custody Deal Model

Custody ops, sales, legal, and product teams relied on spreadsheets and email chains to price, compare, and approve deals. The process was slow, error-prone, and left no clean audit trail.

As UI lead over a year, I co-designed with a UX partner and stayed in weekly contact with custody ops, sales, legal, and product teams. Detailed Figma specs and critique cycles kept design and engineering aligned through the build.

I designed a five-step deal creation wizard, surfacing only the inputs needed at each stage. Real-time validation caught errors early, and an auto-generated summary supported client sign-off.

The key challenge was supporting simple and complex deals without overwhelming either. Adaptive form complexity revealed fields as deal scope increased.

Decision 1. Five-step wizard, not a single long form.

Options considered: single long form, a wizard with stage-gated inputs, a decision tree, spreadsheet upload.

What I chose: five-step wizard.

Why: Deal creation is inherently sequential and the old process was hours long. A wizard enforces the logical order, lets me hide inputs that are not yet relevant, and makes the linear progression the interface’s spine. That shape is what got us from 2 hours to 20 minutes.

Decision 2. Adaptive complexity, not a simple-or-complex mode toggle.

Options considered: explicit simple-or-complex mode toggle, conditional field visibility tied to earlier answers, fixed form for all deals, user-selectable sections.

What I chose: progressive reveal, where fields appear as deal scope grows.

Why: A mode toggle forces users to classify their deal upfront, and they are often wrong. Fixed forms overwhelm simple deals. Progressive reveal lets the interface stay minimal for small deals and expand naturally for complex ones, supporting both audiences without asking them to declare which one they are.

Decision 3. Real-time validation plus auto-generated summary.

Options considered: traditional submit-then-validate, manual summary entry, PDF-only output.

What I chose: inline validation during entry plus an auto-generated summary for client sign-off.

Why: Error cost climbs the longer a user has moved past the error. Catching errors in the field where they happen keeps the deal flow clean. Auto-generating the summary from the captured inputs removes a manual step that used to require a separate rewrite, and it produces the audit trail the compliance team needed anyway.