Last week, we showed how our design expert Tina fared in our little experiment. The goal was to develop a functional web application (mood tracker) for mobile use—including the front end, back end, and design—using AI tools. In this post, we’ll take a look at the approach and results of our backend expert, Danny. We’ll also examine the strengths and weaknesses of AI tools such as ChatGPT, Vercel v0, and GitHub Copilot Workspace.
Backend Expert Becomes a Designer
Danny began developing the concept for the app using ChatGPT. He also utilized custom GPTs specifically tailored to user experience design to define the basic layouts and core functionalities of the mood-tracking app.
v0
Once the concept was finalized, our backend expert enlisted the help of v0 for the design and frontend development. The AI tool v0 was developed by Vercel and is capable of generating executable front-end code from text and image descriptions. Danny fed the tool natural-language prompts that described various layouts and how they functioned.
With additional input, Danny refined the target design until v0 reached its limits and even further iterations failed to deliver the desired results. The solution? GitHub Copilot Workspace.
Iteration 1 (Source: v0)
Iteration 11 (Source: v0)
GitHub Copilot Workspace
GitHub Copilot Workspace is an AI-powered development environment. It enables developers to plan and implement projects—from concept to completion—using natural language. Compared to GitHub Copilot, the tool offers a task-oriented, collaborative environment. This enhances collaboration through real-time interaction with the AI, as well as specification-based planning and execution.
Once v0 had reached its limits, our backend expert created a GitHub repository containing the code generated by v0 so we could continue development using Copilot Workspace. „Certain customizations require direct intervention in the code, since the possibilities for creating prototypes with tools like v0 are eventually exhausted,“ says Danny. Since, as a backend expert, he has limited frontend knowledge, Copilot Workspace was particularly helpful to him. In particular, natural language programming made his frontend work considerably easier.
To start a session with Copilot Workspace, first select the appropriate repository and then begin prompting. Start by defining a task and describing it in as clear terms as possible. After creating a well-structured task and the corresponding specification, Copilot Workspace provides suggestions for implementing the function in natural language. Based on the code from the repository and the task, Copilot Workspace’s suggestions are usually high-quality, but you can still edit, expand, or add subpoints to them. This fosters collaboration between AI and developers and helps prevent misunderstandings during implementation.
Our backend expert recommends keeping tasks small, even though the tool can handle larger tasks just fine. It’s important to group related tasks together and avoid combining different tasks into a single request. A modular approach is the best way to maintain traceability. However, Copilot Workplace also has its limitations and makes mistakes that need to be corrected, which is why a basic technical understanding is required.
While working on the backend, Danny implemented some things on his own, but he also relied more and more on ChatGPT and GitHub Copilot Workspace.
Multiselect (Source: GitHub Copilot Workspace)
Multiselect Result (Source: GitHub Copilot Workspace)
During implementation, our backend expert encountered various challenges, particularly with the correct display of mood-tracking graphs (for example, v0 persistently used months instead of days). These were resolved through manual adjustments and iterative refinements in the Copilot Workspace. The final design improvements were ultimately achieved through targeted code adjustments in the GitHub Copilot Workspace.
Despite the support provided by AI tools, technical expertise and manual adjustments were necessary to achieve a satisfactory result. By using ChatGPT, v0, and GitHub Copilot Workspace, our backend expert was able to create a functional and user-friendly interface, albeit with some limitations.
By combining these tools with his own expertise and through multiple iterations, Danny was able to develop a mood-tracking journal that largely meets the product requirements that had been established earlier.
„Ups & Downs“ Web Application“
Home Page
Statistics
Conclusion and Outlook
The experiment clearly demonstrated the strengths of the various approaches: Tina’s design stood out for its inviting layout and clear focus on user needs, while Danny’s application impressed with its broader range of features. Both Xperts gained valuable insights into the potential and limitations of the AI tools they used.
It became clear that AI solutions are not yet mature enough to fully replace human expertise. While these tools are particularly useful for user research, concept development, and branding, human validation and refinement of AI-generated results remain essential. AI tools were able to significantly speed up processes, particularly in rapid prototyping. However, technical understanding—especially when it comes to coding—remains necessary to successfully implement specific requirements. Nevertheless, it is reasonable to assume that AI tools will be increasingly integrated into design and coding processes in the future, as they noticeably improve both the quality and quantity of output in both areas.
At this point, ethical questions quickly come into play. So far, AI-generated results have been based on extensive datasets that reflect the knowledge of many people. Time and again, certain professional groups have protested because their expertise has been—and continues to be—used to train machines without their consent or compensation. On the other hand, there is the question of to what extent AI should and is allowed to replace human labor. Various approaches exist in this area, such as Human-Centered AI, and there are also efforts at the legislative level to establish clear regulations. As we’ve emphasized in previous blog posts and interviews with our Xperts, we want to highlight the positive benefits of AI. It’s important to view the technology not as an adversary, but as a partner. No one can predict the future, yet we believe that human interaction and collaboration will continue to be important.
Given the rapid advances in the field of AI, we plan to repeat this experiment on a regular basis and further strengthen interdisciplinary collaboration among our teams. To remain competitive, companies must prepare their employees for the future.
