BMW i8 Unreal is a personal project aimed at creating a prototype for a BMW i8 car configurator, using unreal engine with my knowledge of 3D design, real time and UX/UI design.

The goal was to create a tool aiming at delighting users and making an appealing experience instead of just a standard configurator.


UI designer / 3D artist

UI design, interaction, prototyping, 3D design.


This project started with a simple goal in mind, prototype a car configurator with emphasis on interactivity, simplicity and aesthetic using unreal engine 4.

With my knowledge of UX and UI design, combined with my experience with real time rendering and 3D designed, I was sure I could create a viable product, or if not learn from it and what make a viable, more interactive and immersive, real time car configurator.

The process

Given my background I had a good knowledge of standanrd car configurators as well 3D design and some grasp on Unreal engine, but putting all together might prove challenging for a single person within limited time constraints.
So I needed to be practical, whenever I could work with pre-existing resources such as models and shaders I decided to take advantage of it; this allowed me to save time and use it for planning, bug fixing and testing.

Bought few 3D models, that unfortunately I still had to fix and optimize for the purpose, and got hold of some pretty looking unreal engine sharers.
I then got to work on a mockup for the UI and how would the configurator work and look like.

The goal

The goal is to create a working prototype, have a chance to test it with user and gather insights on what would make a commercially viable, easy to maintain and delightfully interactive car configurator.

The advantage of unreal engine is a library of assets and shaders that are realistic from the get go, without requiring a huge amount of customisation.

There is still a thing left, there’s a lot of knowledge I’m missing of the tool, but I’m not planning a perfect, fully functioning product. Just needs to be good enough for testing.

Building an MVP

For different reasons such as, time, resources and a lack of understanding of what would make or break such a product, the approach is clear.

I needed to build an MVP.

The configurator as I mentioned initially needs to have a strong focus on visual, needs to feel cinematic, there must be the “configuration” aspect included, and should work without the need for an expensive GPU. For these reasons few things were removed from the original concept, multiple vehicles won’t help me get any insights more than just one, same for having a price calculators and payment flow.

After the objective was defined and plan laid out, I cut some time out to figure out and learn more about unreal engine and its blueprint system.

Prototyping and testing

Once the objective was clearer, I started dedicating time to understand how to work with Unreal engine blueprint system, then prepped the model and scene.

Everything was on track with the envirnment, lighting, shading and baking, worked them up gradually step by step.
Animating the UI and getting it to work turned out to be more time consuming than anticipated, had to make adjustments.

Focusing on the goal back again, got me reconsider the prototype validation phase.
I didn’t need to build a perfectly working prototype, that’s why I went instead for the Wizard of Oz method.

Created a basic UI, got the configuration bound to different hotkeys, planned a standardized flow to share during testing and then got to the testing.

Got some contacts in the automotive industries and some other in the phase of buying a car to check the prototype for me.
Set the stage and got them to “try” the concept and share their thoughts as they went along, as I was using shortcuts to give the impression of a working product.


The process was challenging but I was able to stay on track by refocusing on the original goal, and chaging testing method instead of wasting time on trying to build something which might prove not what people want.

The outcome of the testing was interesting, even though the interactivity novelty brought excitment, it died down due to the lack of more detailed informations or interior view, required for most when deciding on a big purchase like buying a car.

Is also worth mentioning that even trying to optimise the scene, it was still slow on laptops with less powerful GPUs, making this not as viable as standard configurator for most household laptops and desktops.

I can envision this being a nice addition to a dealer floor as interactive tool to enhance the phisical experience.

More projects