NorthPay: Fintech onboarding and KYC concept
Fintech onboarding and KYC concept
Onboarding for a fintech account
Financial products have to collect sensitive data before anyone sees value, which makes onboarding the most fragile part of the flow. Every field is there because regulation requires it. The order those fields appear in is a design decision.
Problem
Users, the business and compliance want different things
Users want to know why it takes so long, the business wants to know where people drop, and compliance needs every field. Cutting fields was off the table, so the work went into how they are presented.
- 18%of US shoppers who abandoned a checkout blame forced account creation (Baymard)
- 3phases: account type, identity, verification
- 6steps, all visible from the first screen
Flow
Choosing the account type before any form
Choosing the account type before any form opens removes the worst failure: switching type mid-form and losing everything already typed.
Wireframes
One job per screen.
Low fidelity locked structure and order first. A rejected version put the account toggle next to the first fields; switching type there would wipe everything already typed. Keeping the decision on its own screen prevents that loss.

Key screens
The first screen asks one question
The first screen asks one thing. Personal details come grouped, with the stepper always showing what is left.


States
Errors and waiting
Errors appear field by field as people leave each input, instead of all at once on submit. Verification has a pending state with a timeline, so people can tell the wait is normal.


First use
A checklist instead of an empty dashboard.
After verification, people land on a setup checklist with one clear next action instead of every option at once.

Reflections
What a real engagement would change.
KYC can span days, so the flow needs save and resume. Compliance was simulated here; with real AML rules I would map legal requirements before locking one question per screen. And the rejected layout deserved a fair build to test against, not a strawman.
Outcome
Designed against the known drop-off points.
This concept never shipped, so there are no live metrics. Each decision is mapped to a documented cause of abandonment instead.
- 1decision before the first field
- 1question per screen
- 6steps visible from screen one
Benchmark: Baymard Institute, checkout abandonment research.