Story-led case study · Strategy · Leadership · Outcomes

Seven years building the platform two million frontline workers use every day.

YOOBIC · Head of Product Design · 2018 — 2025

Not just the screens — the strategy, the org and the craft bar behind them. I turned a fragmented 30-module platform into an AI-first frontline experience, and held every decision to a number: adoption, speed and engagement all moved.

B2B SaaS · Frontline retail, hospitality & manufacturing Web · iOS · Android London, UK

This page tells the story first. The visuals below are a representative glimpse — out of respect for YOOBIC's competitive confidentiality, deeper screens, performance metrics and unreleased work remain available only on request.

YOOBIC Frontline Employee Experience Platform — Communicate, Learn, Work and Measure around the 'y', serving roles from store associate to Operations Director
One platform, three jobs to be done — Communicate · Learn · Work — measured end to end, for every role on the frontline, from store associate to Operations Director.
7
years tenure
30+
modules across Web & mobile
2M+
users across 80+ countries
350+
global brands
200+
design system components (YOBI)
3
product pillars: Work · Communicate · Learn

What the design decisions actually moved

35%
faster task completion, via redesigned workflows
40%
training engagement with AI-powered learning (NEO)
25%
shorter onboarding through the UX overhaul
20–30%
design–dev gap closed via specs, handoff & QA
Frontline workers across retail, pharma, logistics and grocery using YOOBIC on tablets and phones
Behind every metric: retail, pharma, logistics and grocery teams across 80+ countries.

Why this story matters

Most portfolio case studies are screenshots stitched together with retroactive narrative. This one isn't. Over seven years I led design at the same company through three distinct strategy eras — IPO-track growth, pandemic-era frontline criticality, and an AI re-platforming — and in each one the job was the same: read where the business needed to go, then set the design direction to get it there.

That meant working hypothesis-first — using research, analytics and customer insight to decide what to build, not just how it looked — and being accountable for the result: faster task completion, higher training engagement, shorter onboarding. The screens changed three times. The discipline of tying every design decision to a measurable business outcome didn't.

Three bets, and what they returned

Each started as a hypothesis, was validated with research and analytics, and was held accountable to a number. Strategy first — the system and the screens followed.

The problem · a fragmented 30-module platform

Reset the information architecture around the user's task, not the org chart

The hypothesis: people weren't slow because the product lacked features — they were slow because nothing shared a mental model. We unified navigation and redesigned the core workflows across modules, then validated against task-time analytics.

35% faster task completion
The bet · from task platform to AI-assisted experience

Lead design through the AI re-platforming instead of bolting AI on

I drove the exec-level conversation on where the assistant (NEO) should suggest, decide or stay silent — and embedded it inside learning rather than beside it, so it changed behaviour instead of decorating the UI.

40% training engagement (NEO)
The leverage · 20–30% of design lost in handoff

Make the design system the single source of truth, with specs and QA

YOBI went from a Figma library to 200+ governed components with documentation, tokens and accessibility baked in — turning design–dev friction into shared velocity and cutting the time it took new joiners to ship.

25% shorter onboarding · gap closed

The arc, in five chapters

01 2018

A 30-module platform with no map

I joined YOOBIC as the platform was already feature-rich: thirty-plus modules spanning task execution, communication and learning. Functionally rich, structurally fractured. There was no real navigation system. No design system. No design org in the modern sense. The first job wasn't to redesign anything — it was to understand why a product this useful felt this hard.

02 2019 — 2020

Foundations: YOBI, hiring, the IA reset

Three workstreams in parallel. YOBI, the design system, started as a Figma library and grew into 200+ components across six libraries with documentation, naming conventions and accessibility standards baked in. The team went from a small design function to a structured org with clear seniority ladders, rituals and a shared craft bar. Information architecture got its first principled rewrite — a unified navigation logic across modules so a store associate and a regional manager weren't fighting two different mental models.

03 2020 — 2022

Pandemic: frontline becomes the product

When stores closed and reopened weekly, frontline workflow went from a business-improvement product to a business-continuity product. The mobile experience had to absorb that pressure overnight: faster task surfaces, clearer communication primitives, learning content that worked offline. We compressed a multi-quarter mobile roadmap into months without breaking the design system or the team. That period set the bar for how the org operates under pressure.

04 2023 — 2024

The AI pivot

The product strategy moved from "task platform with reporting" to "AI-assisted frontline experience". That's not a feature — that's a re-platforming. Design had to lead a different conversation: what does an assistant feel like inside a checklist? When does AI suggest, when does it decide, when does it stay silent? How do you keep the design system honest when half the new patterns don't have an established language yet? The answer wasn't a single launch — it was a year of pattern-building, careful ethics work, and teaching the org to design around the model rather than on top of it.

05 2025 — now

Mature platform, new horizons

Today, the design organisation runs on a system, a craft bar, and a culture of taste that didn't exist when I joined. The product has shipped through three distinct strategy eras with the same foundational language. That continuity — not any single screen — is the work I'd point a hiring manager at.

NEO, YOOBIC's AI assistant, drafting a post and surfacing weekly reports inside the frontline app
The headline artefact of the AI era: NEO, an assistant designed to sit inside the workflow — not on top of it.

What "Head of Product Design" actually meant here

Less "ran the design team", more setting direction and influencing the organisation — the part of the role that lives beyond programme delivery and systems work.

Org & people

  • Hired, levelled and grew the design function from a small team to a structured org
  • Set the rituals: critique cadence, design reviews, quarterly portfolio planning
  • Owned career frameworks, performance and craft-bar calibration

System & craft

  • Built YOBI from zero to 200+ components, six libraries, accessibility-first
  • Closed the historical 20–30% design–dev gap through specs, handoff and QA
  • Led the IA reset that unified navigation across the three product pillars

Strategy & influence

  • Shaped roadmap direction and product bets with Product, Engineering and CS — not just the design of them
  • Set and defended the design direction through the AI re-platforming at exec level
  • Drove cross-functional alignment, and represented design in exec planning, escalations and pre-sales

Methodology

Design strategy & process

Every solution was grounded in user research and validated before it shipped — a structured, iterative loop with clear objectives at each phase, not a linear handoff.

1Research & Discovery
2Define
3Ideation
4Design
5Validation
6Implementation
Six-phase design process: Research & Discovery → Define → Ideation → Design → Validation → Implementation, with objectives for each phase
The process ran as a loop, not a line — validation findings fed back into define and ideation rather than stopping at launch. Each phase had an explicit objective and exit criterion.

Mind Mapping & Workflow Mapping

Mapped the entire YOOBIC ecosystem — how Work, Communication, and Learn modules connected across industries and roles. This surfaced navigation bottlenecks and redundant flows before a single wireframe was drawn, and made the IA problem legible to product and engineering at the same time.

YOOBIC ONE full ecosystem mind map — all modules, navigation paths and cross-pillar connections across Work, Learn and Network
Full ecosystem map of YOOBIC ONE — all 30+ modules, their nav paths, and the cross-pillar gaps that made the platform feel broken despite being feature-complete.

Low-Fidelity Wireframes

Quick sketches and wireframes explored multiple navigation layouts and Newsfeed configurations before committing to a visual direction. Fast to discard, fast to test, fast to align across product and engineering — no pixel investment before the structure was validated.

A glimpse of the surface

The story is the point — but here is what the three pillars actually feel like in the hand, and the tools the team built behind them. Representative screens, shared with permission.

YOOBIC no-code form creator with a live mobile preview alongside the desktop builder
System & craft: the no-code form builder — where the design system turns into something non-designers can ship.
YOOBIC Insight Manager dashboard showing active users, completion and compliance rates
Measure: the same data spine — active users, completion, compliance — that kept the frontline accountable.

What seven years taught me

  1. Continuity is the leverage you don't see in a portfolio. The compounding effect of the same person holding the system, the bar and the hiring philosophy across multiple strategy cycles is enormous — and almost invisible from the outside.
  2. A design system is a hiring tool. The best engineers and designers I attracted came because the foundations let them ship something they could be proud of in the first quarter.
  3. "Frontline" means under-served by every default. Designing for warehouse staff, retail associates and field engineers forces a discipline most consumer teams never develop: assume bad connectivity, assume interruption, assume someone else's KPI is on the line.
  4. AI design is mostly editorial work. The hardest part of the AI pivot wasn't the prompts or the patterns — it was deciding what the product should not say.
  5. The right time to leave a role is when the next person inherits a system, not a rescue.

Voices from the team

Verbatim LinkedIn recommendations from people I worked with at YOOBIC. See all on LinkedIn →

“For seven years, Arcangelo stayed focused, resilient under pressure, and showed strong commitment to every key project. […] You never refused any of our many requests, no matter how tight the deadlines were.”
Baptiste Horn Senior Product Manager · YOOBIC
“I worked with Arcangelo for 5 years at YOOBIC, where we collaborated on building the company's design system from the ground up. […] Arcangelo consistently focused on design consistency and contributed to key conversations around accessibility, tokens, and visual standards.”
Damien Arondel Design System Lead · Senior Frontend Engineer
“I had the pleasure of working with Arcangelo at YOOBIC on our dashboard and AI agent projects, and I can confidently say he's a standout design leader. His creative vision and knack for turning complex ideas into clear, engaging designs made all the difference.”
Benjamin Franck Senior Product Manager, Boomi · ex-YOOBIC

Some of the brands on the platform

A few of the 350+ global brands whose frontline teams run on YOOBIC.

H&M
Carrefour
Lancôme
Cartier
Balenciaga
Ralph Lauren
Puma
Vans
Lidl
Marriott
Peugeot
Sanofi

Want to see the actual work?

Screens, design system internals and product metrics are available on request, under a light NDA.