← All work
UX Research · Product Design · Design System

Fábrica de Créditos

End-to-end UX for Crédito 20 Minutos, the in-store credit line behind Alkosto and Ktronix. Research, flows, a design system and a tested functional prototype.

Client
Corbeta (Alkosto · Ktronix)
Role
Lead UX/UI Designer
Team
Small product team at KPaz
Year
2026
Deliverables
Research · Flows · Design system · Prototype
Tools
Figma · Lovable · Claude
Fábrica de Créditos cover
The brief

A credit answer in 20 minutes needs a factory behind it.

Crédito 20 Minutos lets Alkosto and Ktronix shoppers get financing at the counter. Behind it is an internal tool, the credit factory, used by four roles: store advisors who start the application, analysts who study it, leaders who balance the queue and admins who set the rules.

I owned the UX end to end, from benchmark and research to flows, design system and a functional prototype the client tested before any API was connected.

14Credit products benchmarked
4Role-based experiences designed
322Use cases traced and validated on screen
The problem

People trust the advisor more than their phone.

Clients had to repeat their data, never saw the total cost, and when something failed nobody told them why or what came next. Advisors and analysts worked with fragmented information under time pressure. The result was abandoned applications, rejections that felt unfair and a rising risk of identity fraud.

Insight

Speed is table stakes. Clarity is the edge.

The benchmark (Nu, Bancolombia, Davivienda, Rappi, Addi and others) showed that “approved in minutes” is already the market standard. So I reframed the goal: be the best assisted credit, with one shared file from store to analyst, verified identity, an OTP signature on the client's phone, and an answer that's explained even when it's “not for now”.

Proto-persona
One of five proto-personas, built from the benchmark and replaced with real interview data later.
Hybrid customer journey
Hybrid journey from home to store to payment, with emotions and opportunities at each step.
Key decisions

What I designed, and why.

  • One file, four roles: a single application ID shared by the store, the study and operations, so nobody re-enters data
  • Automated checks first: credit bureau, identity, risk engine and PEP screening lead the analyst's view, so attention goes where a human is needed
  • A readable risk traffic light: green approves, red rejects with a visible reason, yellow goes to the analyst
  • Unhappy paths designed on purpose: an explicit contingency mode if the risk engine is down, and failed calls return to the queue instead of getting lost
  • SLA as a countdown on every case, because in a 20-minute promise time is the product
Credit study flow
Credit study flow with UX notes: risk traffic light, PEP screening, verification call and every exit path.
Design system

Foundations, components and principles, documented in Figma.

Semantic color tokens with WCAG-checked contrast, a responsive type scale, and a component library with every state: buttons, fields, status pills, SLA chips, alerts, the risk traffic light, validation rows, KPI cards, role-based sidebars and a 30-icon set. All of it sits on 10 UX principles and a microcopy guide.

Foundations
Foundations: tokens, type scale, elevation and motion
Button variants
Buttons: 6 variants × 3 sizes
Alerts
Alerts
Risk traffic light
Risk traffic light
KPI cards
KPI cards
Validation rows
Validation rows
Prototype

Tested by the client before development.

I built the Figma designs into a functional prototype in Lovable with Claude. Every screen showed its use-case ID and a scope status, so client feedback mapped straight back to the 322 use cases: 140 approved, 161 refined and 21 excluded.

Leader supervision dashboard
Leader view: analyst roster, workload and SLA risk at a glance.
New application
Advisor: new application, starting with the ID
Credit study
Analyst: automated checks first
Result

Decisions settled before a single API call.

Corbeta validated research, flows and a working product instead of static screens. Development received a documented design system and flows already tested with the people who will use them, so engineering started from decisions that were already validated.