TechTrace
The Chat Application, that helps companies to start with AI!
Context
As part of an educational program, I was introduced to the theory and methodology of Design Thinking for the first time. Our partners were the Applied AI Institute of Europe and the recruitment company Hays.
Hays is facing increasing challenges in recruiting the right talents and experts for their clients, particularly in the ever-growing field of AI. As a result, we were tasked with the following challenge.
​​
Team:
5 international students
​
Timeline:
14 weeks, from 10/2024 to 01/2025
Challenge
"How might we design a way for mid-sized companies with over 1,000 employees to identify and define the AI skills they need in order to fill positions on their AI-related projects with the best talent?"
My Task:
Concept development for the product & Design and testing of the UI/UX
Final Design.
Workflow
1 / Identify the Users Pain Points
2 / Ideation
3 / Get inspiration & UX Best Practices
4 / Prototyping with Figma
4 / Testing & User Feedback
(Again Inspiration from Figma, dribble, ...)
5 / End Result
6 / Next Steps
We conducted a total of 13 interviews with all stakeholders involved in the problem, did a lot of research and had a Knowledge Transfer with our Partner the Applied AI Institute of Europe.
​
We found out: Throughout the entire recruiting process, several communication issues emerged. In the following, we focus on one key challenge: Mid-sized companies often struggle to identify which internal processes could be improved with AI, whether an AI solution is suitable for their business, and, ultimately, where to start with AI and which skills are required.
​
Key Findings - Knowledge Gap
Companies don`t know ...
-
Where to start?
-
Which skills are needed?
Point of View
We created a Persona for our Prototype...
Dr. Hans Müller CEO of a German Mittelstand Company
Age: 52
Gender: Male
Education: Ph.D. in Mechanical Engineering
Location: Baden-Württemberg, Germany
Company Size: 1,000 employees
Industry: Specialized Machinery Manufacturing
He sets the overall strategy and vision for a company, ensuring it meets its goals and remains competitive. He is stressed/ambition driven that he has to transform his company for it to be ready for the ever changing world in AI and new technologies.
​
TARGET GROUP:
Decision makers in german mittelstand companies with no technical expertise
Instead of relying on an intermediary tech consulting firm or an expert to assist our users, we aim to develop a tool that our target group can use independently - anytime, anywhere.
We used various brainstorming techniques to generate a wide range of feature ideas and ultimately identified these key features:
​
-
Discover AI Possibilities
-
Identify AI Use Cases & Benefit Analysis
-
Receive optimal Employee Skills profile
... so Companies know where to start and which skillset they need!
Aim to develop an LLM that will help identify possible AI use cases within a mid-sized company plus their benefit analysis. Through the process of a chat, the user can also discover their AI possibilities, through changing focus on different coorporate divisions, challenges & priorities. In the End he will get suggestions of different ways of implementation and required technical skills. The tool will function as a business consulting platform specifically focused on AI.
...we created first Chatbots by prompting on free platforms, to visualize very quick our idea. This prompt will be our overall user story. At the end we decided to develop a fake chat (based on this prompt) visualized by a figma prototype.
Note: Here was the mistake - We didn't create a detailed User Journey, only a content related timeline of the questions
Prototype Version 1
In the first Prototype was the main focus that every function is there. We had a tight time schedule.
First Testing Phase
The first testing phase involved a total of 30 testers, with minimal guidance from our side. Our focus was on asking as many "why" questions as possible to gather extensive feedback.
User Experience Feedback:
​
-
Users were not engaged at the beginning and lacked motivation to use the chat in the first place.
-
Having two progress bars was very confusing, with too many visible points.
-
The interface was often overloaded with information.
-
Infobox text was too small and unclear - should users copy the content? Additionally, the boxes were difficult to find.
-
Checkboxes received generally positive feedback; however, their placement was criticized. Using the chat window as the main interaction space led to negative reactions.
​
UI Design Feedback:
-
Testers found the design outdated, which discouraged usage.
-
A graphical representation of the results was requested, as decision-makers need a clear visual presentation to justify investment.
​
Key Learnings & Takeaways:
-
The main criticism focused on the user experience. It became clear that we had not given enough thought to the user flow.
Next Steps: More Testing & Iteration:
Unfortunately, we didn’t have enough time for a second round of feedback. However, further testing and iterations would naturally be part of an extended product development process.
Additional Critical Points to address:
​
-
Clear definition of user actions: When can the user perform specific actions? Do they always have access to checklists and the chat input? Where and how are action options displayed?
-
Maintaining a clear user flow: Currently, the flow is straightforward due to the limited number of options. We need to ensure clarity remains once more choices are introduced.
-
Improving the company information page: Make editing more intuitive - avoid a single button in the top right corner.
-
Adding essential functions: Include options such as "Back," "Leave Chat," "Pause Chat," etc.
-
Progress bar behavior: What happens if the user returns to the chat? Does the bar stay at maximum, even when new information is added?
-
Testing the results page: Does the user clearly understand their action options? Do the colors of the graphs distract from the actual functions? The design should enhance usability.
-
Checklist improvements: Too many options - consider reducing choices or introducing subcategories.
Reflection & Takeaways
-
It's crucial to identify potential issues and user needs within the flow early on to ensure that the right actions are presented at the right moments in the application.
-
Equally important is preventing users from feeling overwhelmed by excessive information and choices—something that can happen much faster than expected.
-
Users seek an aesthetically pleasing design that enhances usability rather than distracting from it.
​
Challenge Review
​​
-
Received highly positive feedback from both partners and mentors
-
Won the Finalist Award, the highest possible recognition in the competition :)