Thoughtworks · Fortune 50 Client · Extended Contract
Leading Design for
Legacy Transformation
Embedded as Lead Designer on a Fortune 50 Merchandising team — driving user research, product vision, and a new system that saved users 10 hours a week.
Context
The Challenge
At Thoughtworks, the focus was premium technology consulting for large enterprise clients. As a Senior Experience Designer embedded on the client's Merchandising team, the mandate was to find meaningful improvements for Merchandisers — users who plan and manage shelves at scale — while building solutions that could extend across the broader technology organization.
Report Hub
The Solution
User interviews revealed the core problem wasn't the reports themselves — it was discoverability and relevance. The solution pitched to product and approved was Report Hub: a centralized control center for Merchandisers to find, save, and act on the reports they actually need.
- 1A personalized home page surfacing role-recommended and saved reports
- 2Keyword and tag-based search across the full report library
- 3Filters by business task and data source for precision discovery
- 4Save functionality for quick home-page access
- 5Direct contact with report owners for deeper metric exploration
- 6Data visualization pull and multi-report combination
Built within the company's strict design system parameters, Report Hub went from stakeholder pitch → design → usability testing → production pilot (alongside Azure Analytics setup) → full rollout to all Merchandisers.
AI & Modern Practice
Designing With — and For — AI
AI played two distinct roles in this engagement: as a personal tool that shaped how I worked, and as a product feature I was responsible for designing. Both pushed me to think about AI not as a novelty, but as something that has to earn trust from the people using it.
The most tangible AI output in Report Hub was a homepage widget that surfaced AI-generated insights — flagging anomalies and things Merchandisers needed to act on, pulled from their reports and served at the top of their experience before they'd even gone looking.
Designing that widget required thinking carefully about a user base that was not inherently technical — older, process-driven, accustomed to Excel rather than algorithmic recommendations. The design question wasn't just "how do we show the insight?" It was "how do we show it in a way that feels trustworthy enough to act on, without requiring the user to understand how it was generated?"
Clarity over cleverness
Each flagged anomaly needed to be written in plain, action-oriented language not data science output. The design had to translate a model's finding into something a Merchandiser could act on immediately. Trust through context
Surfacing where an insight came from which report, which data gave users a way to verify before acting. For a non-technical audience, traceability is what makes AI feel credible rather than arbitrary.Urgency without alarmAnomaly detection can easily feel like a warning system. The visual design had to signal "this needs your attention" without creating anxiety — especially for users already overwhelmed by busy season.
AI in my own process
Outside the product, I used AI personally to synthesize user research — clustering qualitative interview data into themes and pressure-testing proposed flows before usability testing.
What this engagement taught me is that designing AI features for non-technical users is fundamentally a trust problem. The insight itself is only half the work — the other half is designing the container around it so the right person acts on it with the right level of confidence.
Outcomes
Impact & Results
Report Hub launched to strong reviews. Users saved an average of 10 hours per week, NPS scores rose by 20 points, and the pilot recorded a 100% conversion rate. Analytics also surfaced which reports were most-used and by whom — feeding a continuous consolidation strategy to further reduce costs.
Overseeing the Microsoft Power BI component library — building 12+ components consumed across design and engineering teams — generated an estimated $10,000 in productivity savings per component use.
The product and design teams continue iterating, expanding Report Hub to all branches of the company.
User Pilot
Getting Users to Actually Switch
One of the hardest problems in digital transformation isn't building the replacement — it's getting people to use it. In this case, a replacement tool had already been built and made available to users. Nobody was adopting it, and nobody could explain why. We ran a user research project to find out.
After presenting findings to stakeholders, I proposed a structured 6-week pilot: a small department would use only the new tool — no fallback to the legacy system — while I provided hands-on support throughout the transition. The legacy tool we were replacing had been in use for over 20 years. For many users, it was as automatic as muscle memory.
The hypothesis: transition is possible when users have 1:1 support, clear reference materials, and a structured path through the unfamiliar. The pilot was designed to prove it.
Support System
What We Built Around the Users
- 1A 20-page side-by-side "cheat sheet" mapping old functionality to new — so users could reference familiar tasks without needing to figure out the translation themselves
- 2A dedicated chat channel where users could ask questions at any point, with the new product's tech support available alongside me
- 3Weekly check-ins to gauge how users were genuinely feeling — not just whether they were completing tasks, but whether the transition felt sustainable
- 4Biweekly learning hours with the product owner, who donated time directly to the pilot users for deeper product education
This level of support wasn't just goodwill — it was research infrastructure. For users to participate openly in an extended pilot, they had to feel comfortable and genuinely cared for. That trust was what made the qualitative data reliable. It also meant resisting stakeholder pressure to report things rosier than they were — accurate data was the only data worth presenting.
Long-Term Vision
Designing for a Future I Wouldn't Build
Large corporations move slowly — and as a consultant with a finite contract, I knew some of the most valuable work I could do was point toward what came next. Early in the engagement I ran user research specifically around what users would want from a unified hub experience: a true homepage, dashboard, and control center for Merchandisers. That research sat with me throughout the contract.
Toward the end of my time, I synthesized everything — learnings from Report Hub, the pilot, and the original research — into a conceptual vision for a Merch Hub platform presented directly to senior stakeholders. The goal wasn't to ship it. It was to give the team a north star they could build toward over the next few years.
"Our goal is to create the best place for merchandisers to start work and for MX Product Teams to launch applications — harnessing everything learned from NEXT and Report Hub to enhance the Merch Hub platform significantly."
Skills & Growth
What This Engagement Built
This engagement deepened expertise across research-driven product creation, enterprise design systems, and the organizational side of change management at scale.
Design Case Studies
