Project 314 began after one of GM’s top dealers raised concerns that the vehicle purchasing process is overly complex and time-consuming. She had recorded herself completing the entire process and found that a single sale required an average of 314 clicks and nearly 30 minutes. When compared to Toyota's 5–10 minute process, this was deemed unacceptable, sparking an initiative to simplify and streamline the process.
The long-term vision for Project 314 is to unify the many dealer tools currently used to sell vehicles into a single, cohesive system powered by “Data Lake”—a centralized source of customer and vehicle data. This project is extensive and is projected to take at least 5 years to complete. In the interim, product owners were tasked with improving their existing tools while ensuring they could seamlessly integrate into the planned singular ecosystem.
Project Overview
-
UX Consultant
-
Partnered with UX and product owners to define goals and scope
-
Created high-fidelity Figma prototype
-
Presented prototype to stakeholders for feedback
-
Maintained consistency with GM’s digital design system
My Contributions
Project 314
Streamlining the GM Vehicle Purchase Process
During GM's larger initiative to improve the vehicle purchasing process, the Vehicle Locator System (VLS) was identified as a major pain point, with confusing workflows and limited sorting/filtering capabilities. I was tasked with designing a prototype to visualize an interim solution that simplified dealer searches and provided a foundation for future system improvements.
Design Process
After the initial scope of Project 314 was defined, I was asked to lend design support to the business team focused on improving the vehicle location aspect for the project. At the time, they were focused on creating an improved version of the Vehicle Locator System (VLS)—a core tool used by dealers to find and locate vehicles. They had ideas on how to improve the existing tool, but needed a designer who could quickly translate their ideas into an approachable, user-centered prototype and visual aide to portray these ideas to the greater development team.
I began by meeting with the business team and reviewing the system. Together, we walked through the pain points identified through research and dealer feedback. The existing workflow only allowed dealers to search for one brand, model, and model year at a time. When they changed any of these values, the rest of the form was wiped, forcing them to rebuild the entire search from scratch. This meant that if a customer was open to multiple options—such as either a 2024 or 2025 Chevy Malibu—the dealer had to run the search multiple times. It was slow, repetitive, and easy to make mistakes.
Former VLS Search and Results Screens
With these issues in mind, the business team wanted to implement a sort-and-filter feature to replace the search creation screen. We talked through what they envisioned, and I began prototyping a new version of VLS. This initial iteration used a table layout with filter drop-downs for each column, allowing dealers to sort in ascending or descending order and check multiple options for each build item (make, model, year, color, trim level, etc.). While this approach significantly simplified the interface, the full-width table created a new issue: horizontal scrolling. Dealers couldn’t see all the data at once, and there was a real risk they might overlook information hidden off-screen.
Initial Prototype with Horizontal Scroll Issue
After some brainstorming, we reworked the layout and slimmed down the excess of information shown by default. We decided to initially hide the year, make, and model columns, while still allowing dealers to view these columns if they wished by clicking a corresponding eye icon within the filter table. Since these are typically the first filters selected—and therefore already known to the user—hiding them kept the interface clean without sacrificing access to information if needed.
Additionally, certain dropdown filters were simply too large and convoluted, especially when dealers selected multiple makes or models. For example, selecting both the 2500 and 3500 Silverado would give over 20 body style options, not to mention all available colors, trims, engines, etc.. It was overwhelming and made it clear that the dropdown wasn’t the best solution. While researching alternatives, I found an article discussing UX best practices for applying filters against a wide range of criteria. The examples shown immediately reminded me of the Sims 3 “Create-a-Sim” process as well as the Mii Maker on Wii—two ways I killed an embarrassing amount of time as a teen.
These three examples inspired a revised concept: a slimmer table that didn’t require horizontal scrolling on default, paired with a collapsible filter panel. The panel allowed dealers to refine results and toggle additional data columns as they pleased, reducing visual clutter while retaining the flexibility required for more detailed searches.
Create-a-Sim
Mii Maker
UX Example
Create-a-Sim
Mii Maker
Once I had prototyped a few core filters, I presented the high-fidelity prototype to the larger VLS and development teams. The developers helped determine what could be implemented quickly as an interim solution while work continued on the longer-term system redesign.
During this process, some overlap in responsibilities sparked a bit of friction with GM’s formal UX team. There was initial concern that my work might conflict with what they had already been developing. However, once we met to share our design concepts, it became clear that we had very similar insights and ideas that could easily be combined. This alignment strengthened consistency with GM’s digital design system and ensured the solution would scale effectively. Our collaboration ultimately shaped the current live version of the VLS, now used by dealers globally.
Current (as of April '26) Live Implementation
My Final VLS Prototype—Press Spacebar When Keyboard Input is Needed
(Prototype Experience Best on Desktop)
Outcomes
Working on Project 314 was a clear reminder that even small adjustments to layout, filtering, and visual hierarchy can significantly impact usability. Simplifying VLS’s table and filters helped reduce cognitive load and created an experience that better supported how dealers actually search—rather than how the system had previously forced them to.
This experience also highlighted how crucial communication is when multiple teams touch the same product. What started as a tense overlap in responsibilities ultimately turned into shared understanding and a stronger, more unified design direction.
If I could go back in time, I would have started collaborating with the UX team much earlier. We could have bounced ideas off each other and aligned on final concepts more quickly and with less animosity. They’ve also conducted relevant user tests and secured additional funds for future research which I was not privy to. Without access to this data, I had to rely solely on my own UX best practices knowledge and the business team’s limited understanding of user pain points. Reviewing VLS - specific research would have helped inform my decisions and strengthened my contributions to the design.
Overall, this experience demonstrated how thoughtful UX interventions—even for an interim solution—can simplify complex workflows, reduce user frustration, and create scalable solutions that support long-term growth.