2023 Microsoft Edge & MSN Sustainable Design
2023 · Visual
Introduction
The E-tree design is an initiative aimed at creating a gamification experience on MSN Weather that encourages users to grow virtual trees by checking the weather. When these virtual trees reach maturity, MSN Weather will help plant real trees in Kenya. The goal is to form users' habits of daily weather checking in a delightful and engaging way.
Team member
Hyejung Bae, Alejandra Lan, Sophia Choi,
Bobby Bernethy, Yan Gao
My role
Key Visual
Experience design
01. Visual upgrade
At the beginning of the design process, we were fortunate to be working right after the release of Mid-J. I saw this as a huge boost for visual creation. In the past, we spent a lot of time searching for creative direction, often dedicating significant effort to exploring different artistic styles. This was our first time using AI to help us explore artistic directions, allowing us to quickly produce results that helped us better understand user preferences.
02. Exporing in AI
Although AI helped us quickly express the initial style, one of the key challenges we faced for such a large user base was ensuring that the styles generated by AI were ones users would actually like. Which artistic styles are most popular among the majority of users? How can we assess the differences between styles, and what metrics should we use?
03. User test
I realized that AI couldn't fully meet the requirements for high-quality visual expression. For example, when combining elements with backgrounds, issues like incorrect camera angles, inconsistent lighting, scale mismatches, or extra unwanted details often appeared. In my view, relying entirely on AI to generate core visuals still comes with challenges that are hard to solve. For more refined and detailed visuals, tools like Cinema 4D may still be one of the best choices.
04. User preference & motive
Qusition
Which experience do you prefer?
Insights
Realism is preferred. Concept B, the more realistic rendering of the tree, was most successful and won each matchup it was in. Judges noted that they preferred the more detailed visual of the tree and that it felt like a more pleasant, appealing visual.
This trend was apparent in other matchups as well particularly those that featured Concept D (bubble tree).
Judge's focused on individual details in some of the versions, particularly Concept A. The stump of Concept A stood out to users as an element that they didn't feel blended well with the background image.
Judge's preferred Concept C to Concept A (though it was inconclusive). The one outlier in the results was judge's preference of Concept C (digital) over the Concept A, with judges citing the more playful, fun, game-like treatment of Concept C.
Perception of "sustainability" mostly maps to original preference. There was a slight trend in judges choosing the more literal representations of trees as communicating the concept of sustainability more clearly, but the general takeaway from the results is that judges aligned in their response to this question with their original response to which version they preferred.
Perception of delight and joy maps largely to original preference as well.
Not enough variation in age responses to gather any significant insight. Majority of responses from ages 21-40. Not much variation across other brackets due to smaller judge pool (100 total).
05. Expression
This was my first time using AI to help improve my work efficiency, especially during the early idea-pitching stage. I found that AI could quickly generate a wide range of visual concepts, which helped me explore different directions much faster than before.
However, I also started to think more deeply about the role of AI in the design process. Instead of seeing AI as a final solution, I began treating it as a creative partner—one that’s great for rapid exploration, inspiration, and breaking creative blocks. It shifts the designer’s role from just creating to also curating and guiding.
This experience also made me reflect on how to balance AI-generated content with human judgment. While AI can offer quantity and speed, the quality and emotional resonance of a design still require human intuition and experience. Knowing when to use AI, and when to rely on traditional tools or manual adjustments, has become an important part of my design thinking. 
