UX Research
Manufacturing Data Quality Challenges
Transforming manufacturing efficiency through deep user research that revealed critical workflow gaps and saved millions in unnecessary digitalization costs.
Industry :
Manufacturing
Client :
Bosch
Project Duration :
July 2025 - Present
Tools :
ChatGPT, Gemini, Miro, Figma
Summary :
Challenge
Bosch Long Thành faced severe data quality and digitalization issues across their manufacturing plant that threatened efficiency and scalability.
Target Audience
Manufacturing operators, plant managers, and data stakeholders struggling with distributed data systems and unclear business objectives.
Impact
Pivoted the entire design approach, optimized integration costs, and provided evidence-backed decisions that prevented expensive, unnecessary upgrades.
THE REAL PROBLEM BEHIND THE WORK
Bosch wanted to digitalize their manufacturing plant, but something didn’t add up. Data quality issues kept surfacing, yet no one could pinpoint where things were breaking down. We needed to look past surface-level assumptions and see what was really happening on the factory floor.
Our early analysis revealed fragmented data across multiple systems and a lack of clear business objectives. The deeper challenge wasn’t just technical, it was human. Operators had quietly developed their own workarounds that bypassed the system entirely. Without understanding these hidden behaviors, any new digital solution would have failed before it even launched.
HOW WE SOLVED IT
We began with hypothesis mapping to document what we thought we knew about data flow and operator behavior. Then we went on-site for contextual inquiry, observing operators as they worked. Watching someone try to use a tablet with industrial gloves told us more than hours of meetings ever could.
Our biggest breakthrough came during synthesis. We used AI tools to help cluster patterns and summarize findings at scale, then reviewed everything manually to ensure depth and accuracy. This hybrid approach gave us both speed and sensitivity, letting us process hundreds of notes while still catching the human nuances behind every behavior and frustration.
THE IMPACT WE MADE
We uncovered a major persona mismatch. The system was built for office workers, not factory operators. That single realization shifted the entire direction of the project and saved Bosch from rolling out a costly, misaligned solution. Instead of forcing people to adapt to technology, we redesigned technology to adapt to them.
Once we had clarity, cost optimization became simple. We showed the team which integrations truly mattered and which were unnecessary. By highlighting where data gaps genuinely impacted production, we helped Bosch focus its investment on high-impact improvements instead of blanket digitalization.
THE EXTRA VALUE WE BROUGHT
Beyond research and analysis, we also prepared and facilitated a series of three strategic workshops on data quality awareness. Our goal was to help leadership understand how poor data entry can cripple AI systems. As we often say, “Garbage in, garbage out.”
These sessions brought senior leaders together to define goals, vision, and strategic actions around data governance. It was a rare and proud moment, as UX researchers, we weren’t just improving interfaces, we were shaping organizational mindset. This level of trust and UX maturity within Bosch made the work deeply rewarding.
WHAT WE LEARNED
Real insights live where systems fail. Observing workers directly showed us that their “mistakes” were often smart improvisations that kept the plant running. These weren’t errors to fix but signals of what the system needed to support better.
AI accelerated our synthesis by nearly four times, but human judgment still carried the final word. Machines can catch patterns, but they can’t capture frustration, pride, or ingenuity. The real magic happens when AI and human empathy work together, that’s where better products, and better decisions, are born.
