UX Design & Research · NDA · Enterprise MES · 1 Designer - 3 Devs - 1 PO · 2 years

Worker assistance for final assembly — legible at distance, operable with a glove.

An assistance system for automotive final assembly: a high-mounted production monitor, a handheld (PDA) for inspection and special steps, a configuration dialog for parametrisation, and a dashboard for oversight. This case study focuses on the production monitor as the core — with the three adjacent parts in overview.

Worker → Constraint → System → Interaction → Outcome
Case Study

The obvious brief was "put the data on a monitor." The real problem was physical — and bigger than one screen.

Assembly workers need the right build information at a glance while working to takt — in a loud hall, often wearing gloves, in front of a high-mounted monitor. But the monitor is only one part: inspection and special steps run on a handheld (PDA), the per-station content is parametrised up front in a configuration dialog, and a dashboard gives the people doing that parametrisation oversight and control. This case study covers the production monitor in depth and the three further parts in brief.

Role
UX Generalist, Concept, UI Design and Developer Handoff
Timeline
My involvement 2024–2026 · Total run 2023–2026
Context
Global automotive OEM, via an industrial-software studio
Output
Production monitor · PDA · Parametrisation · Dashboard
The challenge

Legible at three metres, operable with a glove — and for two very different users.

The monitor is mounted high so everyone can see it — which means only the lower half is reachable, especially for shorter workers. It's read from 1–3 m, often with gloves, in a loud hall.

And nothing may be scrolled away or paged — not for lack of space, but because the plants explicitly want it that way. Anyone working to takt on the line shouldn't interrupt their work to navigate the monitor. Everything important has to be there at a glance.

"The standard steps I do in my sleep. I just need to see instantly when something's different."
— Proto persona "experienced worker" · illustrative
Two users, one screen

An experienced worker knows the motions by heart and does them almost blind. For them the monitor shouldn't show the routine but the exception — deviation, special case, special equipment — instantly and unmistakably. At the same time, the same monitor must guide new workers and working students safely, step by step. The routine recedes, the exception jumps out.

"I'm still unsure — I need every step clear and in the right order."
— Proto persona "new worker / working student" · illustrative
Principle

Guide where it counts, freedom where it doesn't: the default is to inform and orient. The system only guides bindingly for takt, quality or safety — an open inspection step, an imminent line stop. In free work, the decision stays with the worker.

Reframe · prioritised How-Might-We
  1. PRIO 1 · CORE
    How do we show build information glanceably, legible at 1–3 m and operable with a glove — without breaking takt?
  2. PRIO 2 · TWO USER GROUPS
    How do we let the routine recede and the exception jump out — for experienced workers and new ones on the same screen?
  3. PRIO 3 · GUIDE VS. FREEDOM
    How do we guide bindingly where quality and safety demand it — and otherwise leave the decision to the worker?
Main part · Production monitor · Process

From the physical environment to hard, testable constraints.

01

Frame

Contextual inquiry on the line: mounting height, reach envelope, glove use, noise, viewing distance, and the age/height/experience spread of the workforce.

02

Constrain

Turn the environment into hard design constraints: reach zone, no-scroll, distance legibility, no forced actions.

03

Adapt the system

Rescale a Material-based component library for distance and gloves instead of designing from zero; define the reach-zone layout and handheld patterns.

04

Prototype

Self-contained, testable single-file prototypes for both surfaces; validate structure and interaction logic.

05

Pressure-test

Review against distance legibility, glove hit-accuracy, overflow-without-new-space, and the "guide where it counts" principle.

06

Hand off

Document the adapted component set, the reach-zone layout, and the handheld interaction pattern for engineering.

Insights

Four constraints from the floor — each with a design response.

01

The reachable area is only the lower half

The display sits above head height; the top is visible but not reliably touchable.

→ Design responseAll controls live in a fixed bottom reach zone; the upper area is display-only.
02

Gloves and noise defeat precision and sound

Gloved fingertips miss small targets; swipe is unreliable; audio is lost in the hall.

→ Design responseOversized tap targets, no swipe, tap-only; every state change is shown visually, never by sound alone.
03

Space is fixed — no scrolling, no page flipping

There's no time to scroll during takt — everything needed has to be visible at once, on a single screen (plant requirement).

→ Design responseFit-to-viewport; excess elements aren't displayed — they're indicated as a count or through compaction, never via a scrollbar or a new panel.
04

Guide where it counts — don't patronise

An implied automatic consequence creates unsafe obligations on a safety-relevant line.

→ Design responseThe default is to inform; binding guidance only for takt, quality or safety — otherwise the worker decides.
Key decisions

What was kept, cut, and established.

✓ Kept
Bottom-anchored control bar

The most important actions stay permanently anchored in the lower, reliably reachable zone. Since only the lower half of the display is safely reachable in the production environment, interaction is designed around it.

✕ Cut
Swipe gestures on the handheld

Swipe gestures proved unreliable, especially when operated with work gloves. They were replaced with clearly recognizable, oversized tap targets.

✓ Established
Scaled library components

Existing library components were deliberately scaled and adapted to the demands of the production hall — especially regarding touch targets, legibility, and operability under harsh conditions.

✕ Cut
Tile display on the PDA

The visually elaborate tile layout was dropped in favor of a minimalist, text-based display. On the small screen, fast legibility of relevant information takes priority.

✕ Cut
Over-design without functional value

Decorative or unnecessarily complex UI elements were deliberately avoided. Instead of modern design for its own sake, the principle holds: as simple as possible, as much as necessary — focused on a solution that is genuinely feasible and viable under the given technical and operational conditions.

Further reading

Selected deep dives from the project.

Each deep dive walks through a single design challenge in detail — three as a pattern applied across several surfaces, one as a concrete cross-device flow.

Deep Dive

Part numbers

Integrating new information into an already-shipped UI — best outcome, least change effort.

TouchesProduction monitor · PDA · Config dialog
Deep Dive

Workflow widget

A pattern for work-step status, applied three times.

TouchesProduction monitor · PDA · Config dialog
Deep Dive

Early warning

A new tile on an already-full screen — and a restructuring of the workstation configuration.

TouchesConfig dialog · Production monitor
Deep Dive

Station sign-on

One shared design for two devices whose technical interplay was unclear at the outset.

TouchesProduction monitor · PDA
Part 2 · Handheld (PDA)

The handheld for inspection and special steps.

Not everything happens on the monitor. Safety-relevant stamping, scan-based product-data capture and the fallback mode on device failure run on a handheld (PDA) — one-handed, in short steps, with unambiguous success/error feedback.

  • Focus over fullness: the small display shows only the identifier and relevant status icons — the station name gives way to what matters.
  • Safe stamping: open inspection items yellow, success green, error red — here the system guides bindingly (safety).
  • No silent logout: a visible countdown before auto-logout, so no one is signed off unnoticed.
  • Fallback mode: if a device fails, a replacement "learns" the station and shows the same content — no production loss.

Screens under NDA · description [proto] based on the project documentation.

Part 3 · Parametrisation (configuration dialog)

No parametrisation, no monitor content.

What the monitor shows per station is defined up front: in a desktop configuration dialog, a dedicated role at the customer (the parametriser) sets up workstations, operations, layouts and inspections — along a deep hierarchy (plant → hall → line → station → workstation). A business-critical enabler that decides what reaches the line.

Part 4 · Parametrisation dashboard

Oversight and control — without judging.

To close, a brief look at the dashboard for the parametrisers: "What have I parametrised? Where are the anomalies? What's still open?" At scale (many halls, dozens of lines) it makes the proportions visible — deliberately neutral, with no good/bad verdict.

  • Ranking instead of scoring; the relative share is the central measure.
  • Several chart formats for different users; horizontal bars beat pie/radar at a distance.
  • Drill-down across three levels and a comparison mode (current ↔ next).
  • No traffic light, no score — the system shows, it doesn't judge.

Data-viz concept · separable as a standalone enterprise case.

Outcome

Patterns, not just screens — and in real use.

What shipped

  • An adapted, distance-scaled component set for both surfaces
  • A reusable reach-zone layout pattern (display-only top, interactive bottom band)
  • A glove-first handheld interaction pattern (tap-only, oversized targets)
  • An in-place overflow-without-growth pattern
  • Validated single-file prototypes + engineering handoff docs

Status & rollout

  • Trialled on site during development & design — real environment, real users
  • In fixed production use at two plants
  • Greenfield (new builds) and brownfield (existing plants) — both cases covered
  • Cross-site rollout across all locations in progress / planning

Pilot phase complete, main work done (~80/20), rollout adjustments ongoing. No quantitative adoption or error metrics claimed — [not verified] until field data exists.

Contextual inquiryDesign-constraints specComponent adaptationReach-zone patternHandheld interaction patternInteractive prototypesEngineering handoff
Reflection

The core wasn't to show more information, but the right information: for the experienced worker the routine recedes and the exception jumps out; the new worker is guided safely through the sequence by the same screen. And around the monitor, the invisible work — parametrisation, inspections, oversight — decides whether the right thing reaches the line at all.