GRUPO KFC
MaxPoint had been in production for 10+ years —running 20+ Grupo KFC brands across 5 countries— without ever being tested with the people who used it. I led the first real-user design process in its history: research across 4 countries, 70+ participants and validation with 46 users before a single line of production code.
MaxPoint had been in production for 10+ years running 20+ brands in 5 countries, without a single test with real users.
Modernise the POS to cloud-native without breaking peak-hour operations or the muscle memory of thousands of cashiers.
I led the first real-user design process in its history: 70+ participants, 4 countries, a complete kitchen-to-counter journey redesign.
A faster, fault-tolerant POS, validated with 46 users before the first line of code.
MaxPoint ran point-of-sale, menu publishing, reporting and staff training for Grupo KFC across Latin America. It was 10+ years old, built in PHP, and had never been designed around the people who used it.
The outcome was predictable: cashiers zoomed the browser just to operate the POS. Reports took 15+ seconds, blocking real-time decisions. Publishing a menu needed multiple approvals with no progress indicator. Managers called IT for tasks that should be routine. Across 4 countries and 20+ brands, the same friction repeated thousands of times a day.
The business risk was real: the platform's fragility hit operational efficiency, staff retention and the ability to scale into new markets. A full redesign wasn't a nice-to-have — it was critical infrastructure.
- 4 countries with operational variations across chains and franchises
- 20+ brands with different configuration needs
- Heterogeneous devices: touch POS terminals, tablets and back-office computers, all 10+ years old
- Change resistance flagged as a risk from day one
Grupo KFC staff need to complete critical operational tasks —publishing menus, closing the till, generating reports, taking orders— inside MaxPoint during their shifts, without relying on IT calls or workaround processes.
The system made it hard. An interface built for the system, not for people: no responsiveness, no visual states, no training mode.
I designed and led a CX LAB —a multi-country research and co-creation process grounded in Design Thinking. It revealed that staff had built parallel processes around the failures: the "real" system was MaxPoint + IT calls + WhatsApp groups + manual spreadsheets.
Built from real behaviour patterns — not stakeholder assumptions.
Jessenia López
29A faster till close; a tool that helps her grow.
Juan Sebastián Pérez
37Room to train and deepen operations.
Isabel Espinal
34Stable bank connection; a more efficient nightly close.
José Carlos Fernández
24Real, up-to-date data to decide, without hand-editing reports.
CX LAB — Research + co-creation
Empathy maps by country and role, a Customer Journey Map of the full operational flow, a User-Centered Design Canvas with the core team, the "Boat" activity (pains + wishes + commitments), qualitative personas and sketching sessions where users designed their own ideal interface. Key decision: structured co-creation over interviews alone, so the design mirrored their real mental models.
Synthesis and design principles
Every finding translated directly into a concrete design response (see table below).
Figma design — BackOffice · POS · Mobile app
BackOffice: 3-step menu wizard, drag-and-drop assignment, province/city visibility, role administration. POS: product grid with imagery and icons, order customization, upselling, multiple simultaneous orders, mixed payment. Mobile app: full redesign of order-taking for tablet/mobile, with upselling, grouping, cart review and QR capture.
Validation with real users
Formal usability test with 8 real users in Ecuador —5 critical tasks— before a single line of production code was written. The pivot: MaxPoint had never been tested with users before launch. That single difference —validation before development— was the strongest argument to the client.
| Documented pain | Design response |
|---|---|
| No visual feedback | System status visibility on every action (Nielsen heuristic #1) |
| Zoom required on the POS | Fully responsive interface, optimized for touch screens |
| 15+ second reports | Real-time dashboards, segmented by user profile |
| Excessive approvals | Simplified approval flows with built-in status indicators |
| No training mode | Sandbox environment replicating production without touching real data |
Two rounds of usability testing were run on platform mockups.
Round 1 — Ecuador · 46 participants
| Rating | |
|---|---|
| Ease & friendliness | 4.2 / 5 |
| Design / Look & Feel | 4.0 / 5 |
| Usability | 4.3 / 5 |
| Findings identified | 24 |
Round 2 — Multi-country
| Administrative | Operational | |
|---|---|---|
| Ease & friendliness | 3.3 / 5 | 4.3 / 5 |
| Design / Look & Feel | 3.7 / 5 | 4.6 / 5 |
| Findings identified | 28 | — |
Operational users (cashiers, store managers) found the system more intuitive and satisfying than administrative ones. Strengths in order-taking and customization; improvements in corrections and split payments.
- 10+ year legacy system replaced by a responsive, role-aware platform
- Parallel processes eliminated: IT calls, WhatsApp groups and manual spreadsheets
- Centralized menu management for every location with geographic restriction controls
- Built-in training mode — no IT dependency for onboarding new staff
- Native upselling in the POS flow — enabling revenue growth at point of sale
I'd push harder for post-launch metrics: real reduction in time-per-task, IT ticket volume before and after, onboarding time for new staff. The usability validation confirmed learnability — those numbers would have made the business case even stronger.
Service Design mapped the frontstage well, but the backstage (IT flows, support, system administration) deserved a dedicated blueprint. Some pains were downstream of process problems the redesigned UI couldn't fully solve.
