B2B
B2B
PepsiConnect
PepsiConnect
PepsiCo Global · 2023 – Present · Enterprise B2B
This isn't a product overview. It's an account of designing a global B2B platform across 15+ markets as a UX Analyst — building systems under real pressure, making hard tradeoffs, failing forward, and learning what ownership actually means.
When I joined, it was mid-expansion, under-resourced, and carrying more design debt than anyone had fully mapped.
Today I lead digital ordering, customer engagement, and design systems globally — across Poland, Turkey, Spain, Egypt, Saudi Arabia, the UAE, Thailand, and New Zealand.
PepsiCo Global · 2023 – Present · Enterprise B2B
This isn't a product overview. It's an account of designing a global B2B platform across 15+ markets as a UX Analyst — building systems under real pressure, making hard tradeoffs, failing forward, and learning what ownership actually means.
When I joined, it was mid-expansion, under-resourced, and carrying more design debt than anyone had fully mapped.
Today I lead digital ordering, customer engagement, and design systems globally — across Poland, Turkey, Spain, Egypt, Saudi Arabia, the UAE, Thailand, and New Zealand.
Role
UX Analyst / Product Designer
Role
UX Analyst / Product Designer
Timeline
Sept 2023 - Present
Timeline
Sept 2023 - Present
Project Scope & Markets
Global · 19+ Markets
Project Scope & Markets
Global · 19+ Markets
Markets I owned
Global · 8+ Markets
Markets I owned
Global · 8+ Markets
Deep Dive
Where the thinking went deepest.
Built a complete RTL design system for Arabic speaking markets — components, logic, global files, and two live market deliveries.
Built a complete RTL design system for Arabic speaking markets — components, logic, global files, and two live market deliveries.
Field research in India, behavioural analytics across live markets, and a WhatsApp-first ordering experience grounded in real user behaviour.
Field research in India, behavioural analytics across live markets, and a WhatsApp-first ordering experience grounded in real user behaviour.
My Contributions
The work that made the difference
01 ——
Design System Governance
Maintained and evolved the global PepsiConnect DS — component standards, naming conventions, cross-market consistency. One update flows to every market automatically.
01 ——
Design System Governance
Maintained and evolved the global PepsiConnect DS — component standards, naming conventions, cross-market consistency. One update flows to every market automatically.
01 ——
Design System Governance
Maintained and evolved the global PepsiConnect DS — component standards, naming conventions, cross-market consistency. One update flows to every market automatically.
02 ——
App Performance
Audited and optimised components, simplified screen structures across key screens. Contributed to a 40% performance improvement across the platform.
02 ——
App Performance
Audited and optimised components, simplified screen structures across key screens. Contributed to a 40% performance improvement across the platform.
02 ——
App Performance
Audited and optimised components, simplified screen structures across key screens. Contributed to a 40% performance improvement across the platform.
03 ——
AI Agent
Built an internal AI agent on Copilot — trained on Figma files, BRDs, and research outputs. Context that took hours now takes minutes. Less dependency, faster decisions, better output.
03 ——
AI Agent
Built an internal AI agent on Copilot — trained on Figma files, BRDs, and research outputs. Context that took hours now takes minutes. Less dependency, faster decisions, better output.
03 ——
AI Agent
Built an internal AI agent on Copilot — trained on Figma files, BRDs, and research outputs. Context that took hours now takes minutes. Less dependency, faster decisions, better output.
04 ——
PepGPT · Smart Easy Order
Designed an AI-powered ordering experience where store owners scan their shelf — the AI reads the rack, identifies what's running low, and generates a suggested order list automatically. No manual input, no catalogue browsing. Point, scan, order.
04 ——
PepGPT · Smart Easy Order
Designed an AI-powered ordering experience where store owners scan their shelf — the AI reads the rack, identifies what's running low, and generates a suggested order list automatically. No manual input, no catalogue browsing. Point, scan, order.
04 ——
PepGPT · Smart Easy Order
Designed an AI-powered ordering experience where store owners scan their shelf — the AI reads the rack, identifies what's running low, and generates a suggested order list automatically. No manual input, no catalogue browsing. Point, scan, order.
05 ——
Arabic Design System
Built a complete RTL design system for PepsiConnect — components, direction rules, naming conventions, global files across CE, DO, and PM. Delivered Saudi Arabia and Egypt end to end. Every Arabic market going forward starts with this foundation.
05 ——
Arabic Design System
Built a complete RTL design system for PepsiConnect — components, direction rules, naming conventions, global files across CE, DO, and PM. Delivered Saudi Arabia and Egypt end to end. Every Arabic market going forward starts with this foundation.
05 ——
Arabic Design System
Built a complete RTL design system for PepsiConnect — components, direction rules, naming conventions, global files across CE, DO, and PM. Delivered Saudi Arabia and Egypt end to end. Every Arabic market going forward starts with this foundation.
06 ——
Research & Validation
Conducted field research in India — store visits, user interviews, real behaviour observed in context. Post-launch used FullStory across live markets for session replays, heatmaps, and drop-off analysis. Design decisions backed by evidence, not assumption.
06 ——
Research & Validation
Conducted field research in India — store visits, user interviews, real behaviour observed in context. Post-launch used FullStory across live markets for session replays, heatmaps, and drop-off analysis. Design decisions backed by evidence, not assumption.
06 ——
Research & Validation
Conducted field research in India — store visits, user interviews, real behaviour observed in context. Post-launch used FullStory across live markets for session replays, heatmaps, and drop-off analysis. Design decisions backed by evidence, not assumption.
Process
Designed. Questioned. Shipped.
#1
Business Context & Requirements
Every feature started with a direct conversation with the BU. Understanding the business need, the market context, and the constraints before touching a frame. Requirements documented, gaps flagged, ambiguity resolved upfront.
#1
Business Context & Requirements
Every feature started with a direct conversation with the BU. Understanding the business need, the market context, and the constraints before touching a frame. Requirements documented, gaps flagged, ambiguity resolved upfront.
#2
Competitive Analysis & Brainstorming
Existing patterns researched, competitor behaviour mapped, ideas explored before committing to a direction. For Arabic markets this meant studying Amazon.sa, Noon, and Carrefour — understanding real user behaviour and navigation patterns.
#2
Competitive Analysis & Brainstorming
Existing patterns researched, competitor behaviour mapped, ideas explored before committing to a direction. For Arabic markets this meant studying Amazon.sa, Noon, and Carrefour — understanding real user behaviour and navigation patterns.
#3
Designing — Screens & Components
Low fidelity to high fidelity. Every component built to global standards, every screen considering what happens at market 12, not just market 1. Arabic DS built in parallel to ensure consistency from day one.
#3
Designing — Screens & Components
Low fidelity to high fidelity. Every component built to global standards, every screen considering what happens at market 12, not just market 1. Arabic DS built in parallel to ensure consistency from day one.
#4
Sign Off
Designs reviewed, aligned, and approved with stakeholders before handoff. Edge cases covered, business logic validated, full alignment with engineering before handoff.
#4
Sign Off
Designs reviewed, aligned, and approved with stakeholders before handoff. Edge cases covered, business logic validated, full alignment with engineering before handoff.
#5
Design QA
Every feature tested against the original design intent. Discrepancies flagged, implementation reviewed, production quality held to the same standard as the Figma file.
#5
Design QA
Every feature tested against the original design intent. Discrepancies flagged, implementation reviewed, production quality held to the same standard as the Figma file.
#6
Post-Deployment Analysis
FullStory sessions, heatmaps, user journeys, drop-off funnels — all reviewed after deployment. Behaviour data collected, patterns identified, opportunities surfaced for the next cycle.
#6
Post-Deployment Analysis
FullStory sessions, heatmaps, user journeys, drop-off funnels — all reviewed after deployment. Behaviour data collected, patterns identified, opportunities surfaced for the next cycle.
#7
Refinement
Insights from analysis fed back into the next iteration. Design decisions updated, components improved, global files kept current. The loop closes and opens again.
#7
Refinement
Insights from analysis fed back into the next iteration. Design decisions updated, components improved, global files kept current. The loop closes and opens again.
Recognitions
The work didn't go unnoticed
Andres Merizalde
Global UX Lead
