PRODUCT DESIGN / AUTOMOTIVE
CDK GLOBAL
Product Designer / 2022–2024
One service visit. Two very different jobs.
I designed for both sides of dealership service: a Customer Portal that helps car owners make decisions and NXT Inspect, where technicians and service teams manage the work behind those decisions.
AT A GLANCE
2 years
Across dealership products
Customer Portal
Scheduling, approval and payment
NXT Inspect
Communication, parts and service orders
The work below is a selection from the engagement. The figures and identifiers shown inside design artifacts are examples, not outcome metrics.
01 / CONTEXT & ROLE
The visit starts before the car arrives.
A service visit includes an appointment, a handoff at the dealership, an inspection, a decision about recommended work and a return trip to collect the vehicle. To the car owner it is one visit. Inside the dealership it is work distributed across people and tools. The interface must make clear what is happening and who needs to act next.
I took design ownership of improvements to NXT Inspect, an existing product used to work through repair orders, and contributed to Customer Portal flows across desktop and mobile. I worked with product and engineering and used findings supplied by a dedicated researcher. I turned those findings and business rules into user stories, wireframes, prototypes and interface designs. The research was a team input, not a solo research claim.
The tension was deciding where to simplify. Customers may only open the portal around a visit and need a guided path. A technician may work across many repair orders in a day and needs to see enough detail to act without reopening several screens. A single “simplify everything” rule would have made the second job harder.
01 / CAR OWNER
Arrange, approve and collect
02 / DEALERSHIP
Inspect, coordinate and repair
03 / CAR OWNER
Review the bill and pick up
02 / RESEARCH & WIREFRAMES
Translate complaints into decisions.
Research surfaced confusion about basic portal actions, friction arranging vehicle pickup and limited opportunities to present relevant recommendations. Those findings described three different design problems: knowing where to start, knowing what will happen at the dealership and having enough context to decide on additional work.
We worked through user stories before committing to a screen hierarchy. In the scheduling wireframes, I established the customer and vehicle before asking what needed servicing. That sequence matters for returning customers with multiple vehicles: choosing a service first can hide which car it applies to. Establishing the vehicle also gives later screens a consistent reference.
Exploration went beyond booking. I worked through alternatives for the first scheduling screen, arrival/check-in and the vehicle tracker. Those moments ask different questions: “Can I arrange this visit?”, “Did my check-in go through?” and “Where is my vehicle in the process?” Giving each moment a clear job was more useful than reusing the same action hierarchy throughout.
FIG. 01
Research and user stories

The original board shows team research, feedback and user stories feeding the design work.
FIG. 02
Appointment structure

Early wireframes establish the customer and vehicle before selecting work.
FIG. 03
Scheduling, check-in and tracking alternatives

Exploration board from the original portfolio. These are options considered before higher-fidelity design.
03 / CUSTOMER PORTAL
Give the customer a next step at every handoff.
At booking, a returning or new customer starts the appointment, selects a vehicle and sets preferences such as drop-off, pickup and transportation. The review screen gathers the date, vehicle, advisor, transportation and services in one place, with edit actions beside the relevant information. That lets someone correct one detail before scheduling without restarting the visit.
After check-in, the design confirms completion and gives the customer a way to track progress. When the dealership recommends additional work, the customer sees the recommendation beside the evidence and the price, then can approve or decline. The design supports a commercial decision without treating a recommendation as already accepted.
Checkout is a different decision: confirm what is being charged, choose a payment method and know what happens once the payment succeeds. The confirmation supplies a pickup location, not just a success checkmark. It links the digital payment to the next action in the physical dealership.
FIG. 04
Customer Portal: eight original mobile screens

Read left to right across the first row for appointment entry, vehicle selection, preferences and review; the second row shows check-in, recommendations, checkout and pickup confirmation.
04 / NXT INSPECT / COMMUNICATION
Make the repair order the reference point.
Technicians and service teams needed to follow notes, system updates and chat while moving between the repair order and individual service lines. Combining the three into one undifferentiated feed would make it harder to see which messages applied to the current repair.
I worked on a unified message workspace with the order and service lines visible alongside it. Type filters distinguish notes, messages and system events. A person can read across the order when they need its history or narrow the view when working on a particular line. The interface keeps the unit of work visible rather than asking the technician to reconstruct it from memory.
The design had to respect the density of an established tool. The change was to bring related information together and expose scope clearly, not replace the surrounding workflow with a separate messaging app.
FIG. 05
Unified notes and messages

Original NXT Inspect design. The order and service-line context sit beside the message history.
05 / NXT INSPECT / CORRECTIONS
Make moving a part visible and reversible.
A part can belong to the wrong service line. In a dense repair order, a change can affect totals and what colleagues see later. I designed a drag-and-drop interaction inside the service editor so that the source, selected rows and destination remain available while making the correction.
The sequence shows parts selected in the source, the move in progress, then a confirmation in the receiving line with an Undo action. Confirming at the destination answers “where did it go?” and gives the person a way to recover from a wrong move. It is the correction and its feedback together that make the interaction useful.
I also built a product-specific component gallery for table headers, row types, views and layouts that the broader design system did not cover. Row indicators needed to support the meaning of the row without replacing the service information. Reusing those patterns gave later designs a consistent starting point.
FIG. 06
Selected → moved → confirmed

Original part-transfer sequence. The receiving service line confirms the change and offers Undo.
FIG. 07
NXT Inspect component gallery

Original product-specific collection for tables, rows and layouts; a shared reference for continued iteration.
06 / DELIVERY & REFLECTION
Measure clarity for each role.
The work delivered Customer Portal designs and ongoing NXT Inspect improvements across communication, part transfers and reusable patterns. It documents capabilities and my contribution; it does not establish a feature-level change in NPS, revenue or technician productivity.
If I were validating the work again, I would ask car owners to reschedule one detail or find pickup instructions without an advisor. For dealership staff, I would measure whether they can locate an order-level message, correct an incorrectly assigned part and recover after selecting the wrong destination. A completion rate averaged across those tasks would conceal where the design actually helped.
What I took away: a guided portal and a dense service workspace can both be clear. The right level of detail depends on what the person is trying to decide, how often they do it and what it costs to make a mistake.