Apr 23- Sep 24
Smart’s new CGI car configurator is powered by AI and is the key tool behind the marketing and sales initiatives. As the previous workflow was clunky and the interface outdated, we modernized the entire platform, integrating AI to streamline customization for campaigns and client presentations.
Context
Smart originally introduced the CGI configurator as a quick way for marketing and sales teams to generate car visuals for campaigns and client presentations. But as the tool was initially developed by an external agency that did not conduct any research on their end users, it became more of a bottleneck than a supporter.
When we were asked to redesign it, it was quite slow, filled with errors that users were often confused about, and overloaded with unnecessary steps. Producing a single set of visuals often meant delays and inconsistent quality, leaving the marketing team often frustrated and less focused on the storytelling and brand messaging.
Goal
Our goal was not just to refresh the look of the configurator, but to completely rethink how it worked. We wanted to make it faster, simpler, and more reliable, while also modernizing it with AI-powered image generation. The vision was to give marketing teams an intuitive, dependable platform where they could configure vehicles, generate high-quality visuals instantly, and ultimately spend more time on creativity and less time being frustrated with the tool.
Impact
Team
Product discovery
Before jumping into redesign, I spent time digging into how the configurator was actually used day to day: I started with 8 stakeholder and user interviews, followed by a two-week internal survey with marketing and sales teams who relied most on the tool. This mix of conversations and structured responses gave us both numbers to quantify the issues and stories that highlighted where the tool was causing the most pain.
Heuristic Evaluation
To complement this, I ran a heuristic evaluation of the platform, which quickly surfaced some major pain points: inconsistent design elements, limited flexibility, confusing terminology, and no clear guidance when users got stuck.
User Personas
To make sure our redesign stayed grounded in real needs, we built three core personas based on interviews and survey:
One focused on platform reliability and smooth operation, which cared about performance, security, and making sure the tool worked across teams without unnecessary complexity.
One responsible for day-to-day content production: his biggest need was ensuring the configurator delivered consistent, high-quality imagery that aligned with Smart’s brand, without requiring endless fixes.
One working hands-on with campaigns: she wanted the platform to be fast, flexible, and able to target audiences with creative assets that didn’t break under technical constraints.
Wireframes
I started with low-fidelity sketches and wireframes: these quick iterations helped us map out the end-to-end flow of the car configurator, test navigation patterns, and define the essential steps users needed to take without getting distracted by visual details.
The wireframes acted as a shared blueprint and, paired with the personas we had developed from research, they ensured that every decision was tied back to real user needs and pain points, not just assumptions.
AI Flows
I worked closely with prompt engineers to explore how AI could accelerate production without compromising brand quality: together, we tested flows for backplate generation, upscaling, and relighting, validating how AI outputs could integrate with solid 3D bases. These investigation sessions and experiments helped us identify exactly where automation could save time, and where human control was still essential.
Design System
We built a modular design system with reusable patterns and UI specifications so, instead of fixing issues one by one, we defined components for the entire workflow (from navigation elements to visual selectors) ensuring consistency and efficiency across the configurator.
Closing notes
The redesigned configurator turned what had been a frustrating bottleneck into a fast, intuitive platform. Marketing teams could now produce campaign-ready visuals 30% faster, with error rates down by 12% and user satisfaction up by 22%. Beyond the numbers, the biggest win was also trust, since teams no longer worried about the tool slowing them down or producing inconsistent results, and could focus on creativity and brand storytelling, which is exactly what Smart needed the tool to support.
From the team
