simplify comlicated product for diverse users
B2POS lending platform redesign
how we simplified a complicated flow for purchase loan aggregator
Quick summary
PosCredit is a B2B lending platform used in large федераль retail chains (electronics and appliance stores).
At the point of sale, a cashier or credit manager submits a customer application. The system sends requests to multiple banks, aggregates responses, and allows the customer to choose a loan offer on the spot.
The product is a part of a larger ecosystem that also included a direct-to-consumer app.
This case focuses on the in-store POS system.
Problem
The existing process was outdated and operationally heavy.
Key issues:
  • No unified design system — inconsistent UI across flows
  • Overcomplicated step structure
  • Redundant statuses visible to frontline staff
  • Too many editable fields leading to frequent input errors
  • High cognitive load during live customer interaction
Credit managers had to navigate unnecessary system states that were relevant for underwriters — but meaningless at the checkout counter.
In a high-pressure retail environment, this slowed everything down.
old version of form of applicants data
Context
  • Partners
    12 connected banks
  • Dynamic
    Asynchronous responses (offers appear progressively as banks reply)
  • Legal
    Regulatory constraints, including government ID verification flows
  • Form-factor
    Desktop and mobile versions used directly in-store
Approach
The redesign focused on operational clarity.

Not feature expansion.
Not visual refresh.
Pure workflow optimisation.

Together with analytics and product, we rebuilt the CJM and identified friction points across the full application lifecycle from data entry to offer selection.
key decisions
  • Scenario restructuring
    The application flow was rebuilt around a linear, predictable structure.
    Each step had a single clear purpose

    Where possible, steps were merged

    Where necessary, actions were locked to prevent invalid edits.
  • Progressive bank responses
    instead of system-driven statuses, the interface focused on what the manager actually needed to know:
    • N Offers received
    • Waiting for banks
    • No approval
    Internal underwriting states were removed from the main surface.
  • Field control and validation
    High-error fields were either:
    • prefilled through government ID service integration
    • or constrained with stricter validation
    This significantly reduced correction cycles.
  • Status decluttering
    Dozens of informational statuses — useful for backend teams but irrelevant at checkout — were removed or relocated to secondary layers.
    The visible UI became decision-oriented, not system-oriented.
  • Design system foundation
    A unified component library was introduced across desktop and mobile.
    This reduced UI inconsistencies and accelerated further development.
Outcome
  • fast
    Average application time reduced from ~18 minutes to ~11 minutes
  • predictable
    Data entry error rate decreased by ~27%
  • simple
    Onboarding time for new credit managers reduced by half
  • native
    Support tickets related to “application stuck” or unclear statuses dropped by ~40%
Operationally, the product shifted from being a complex financial interface to a controlled sales tool.

For retail staff, it became simpler to use.

For the business, it became more predictable and scalable.
Photo by Jacob
Photo by Kolya
Photo by Oliver
Photo by Leo
Photo by Paul
Photo by Lea
Photo by Fabrice
Photo by Adam
Photo by Arnaud
Photo by Leopold
Photo by Katie
Photo by Tiana
Photo by Mohd
Photo by Mohd
Photo by Mohd
Photo by Mohd
Photo by Mohd
Photo by Hal
no prototype available due to nda
more