UX Research
Product Design
DesignOps
Storybook
Enterprise
B2B
"Atlas" — Where the system proved itself.
Real complexity. Real stakes.
Company
Role
Senior Product Designer
Period
Aug 2025 - Present
Industry
Retail / B2B Operations
24+
Active modules
18+
Context
Atlas is Worten's central backoffice — running everything from commercial proposals to logistics and access control. Dozens of development teams build inside it simultaneously, each making independent decisions about components and patterns.
Atlas" is a pseudonym. The real product name has been changed for business confidentiality and GDPR compliance.
1
Start with the most painful module, not the most visible
Chose the CMP — Atlas's most critical and most broken module — as the pilot. Not because it was easy to show, but because fixing it would prove the approach worked under real pressure.
✓ Full stakeholder buy-in
✓ Zero roadmap conflicts
2
Research before redesign — always
Ran structured interviews across commercial teams and approvers before touching a single screen. Mapped the AS IS journey with painful precision.
✓ 95% adoption across teams
3
Make knowledge infrastructure non-negotiable
Built the Confluence documentation layer in parallel with design — not after. Because good design that disappears is just expensive waste.
✓ Accessibility standards
Track 1: Understanding the problem
Before designing anything, I needed to understand what was actually broken — from the people living inside it every day.
We started with the CMP — one of Atlas's most critical and most painful modules. It handles everything from B2B proposal creation to stock validation and approval workflows. Multiple user profiles, high stakes, and a backlog of friction accumulating for years.
Track 2: Designing the TO BE
→ Fewer clicks, more efficiency
Every flow should be as short as possible. Context and filters persist to avoid repetition.
→ Real-time visibility
Stock and status always visible before any submission. No more blind decisions, no more parallel communication channels.
→ Automation wherever possible
Auto-fill client data from NIF. Remove articles without stock automatically. Stop making users reimport Excel files to fix problems the system created.
→ Immediate, integrated feedback
Centralised notifications inside Atlas. Clear error messages when something fails. No more silence.
→ Consolidated, contextual flows
Avoid dispersion. Bulk approvals, parent/child proposal structures, multi-select actions — all in the same window.
-> Flexibility without overload
Allow exceptions and business variations without breaking the base experience or overwhelming the user.
These six principles became the design contract — the shared language between me, product, and engineering for every decision that followed.
Track 3: RISE meets reality
-> Every aesthetic adjustment became a User Story in Jira
Every decision captured in Confluence — not as a snapshot, but as a living document any designer or developer could consult, question, or build on.
Track 4: Building the knowledge infrastructure
The biggest risk in a platform as complex as Atlas isn't bad design. It's good design that disappears.
When dozens of teams build inside the same ecosystem, decisions made in one module get reinvented two sprints later in another. Without a shared reference point, consistency is impossible — not because people don't want it, but because they don't have the tools to find it. I built that reference point.
→ Product documentation in Confluence
Became the single source of truth for the entire Atlas ecosystem: product vision, domain overview, atomic components, UX principles, writing guidelines, and access profiles. Each section designed to be found, understood, and acted upon — not filed and forgotten.
For the atomic components — tables, forms, buttons, modals, alerts, filters, wizards, and 18 others — I documented not just how they look, but how they behave: states, accessibility rules, filter types, progressive loading patterns, fallback scenarios, keyboard navigation.
The kind of detail that stops a developer from making the wrong call at midnight.
__
How I work
Every new product or feature follows a defined lifecycle — from proposal to development. This ensures quality, consistency, and maintainability.
Understand the goal
1-2 weeks
Start with user pain points and product metrics. Define the exact problem before touching Figma.
7 Inputs
7Outputs
Tech alignement
1-3 days
Meet engineering early. Surface constraints, backend trade-offs, and what's feasible before any design work.
7 Inputs
7Outputs
Selected UI Work
A selection of UI-focused projects from agency work — pitches that won accounts and delivered implemented digital products for recognised brands.
RISE — The system Worten didn't know they needed.
Dozens of product teams. Zero shared foundation. I saw the problem, proposed the solution, and led it for three years. That system is RISE.