Learn more about the use case:
Test automation using AI
What drove the decision to develop an AI-based use case?
My co-founder and I have been dealing with the problems in test automation for a long time. The initial idea that AI could be a solution came during our studies. When we looked into it more closely, it was clear to us that it would be a game changer.
Specifically, we use AI in the object detection of elements on user interfaces. You can think of it as self-driving cars, which also have to recognize vehicles and people visually - but on applications.
What were your expectations or requirements for the AI application?
First and foremost, it was important to us that the use of AI generates added value for the customer. At the beginning, we even actively refrained from mentioning that we were using AI. For us, AI was more of a tool to solve problems where classic approaches fail. In our case, this specifically includes:
- Object recognition of elements on user interfaces of different operating systems.
- Conversion of natural language into automation steps.
Both object detection algorithms and LLMs are perfectly suited for these tasks.
What frameworks/methods did you rely on for development?
We are a KIT spin-off and the TECO Institute in particular has supported us. The institute was our university partner during our EXIST funding (a scholarship for founders). Even today, we work closely with the institute on a wide range of issues.
The use of AI solutions must also be accompanied by a change in competences. How have you dealt with this challenge in respect of your employees?
We have been an AI-first company from the beginning, and all of our employees have had experience with AI in some way. However, we also actively promote the further development of our employees in communities of practice internally.
In what timeframe did they implement the AI solution in your company/your client‘s company?
Already since our foundation, our AI has been an elementary part of the solution, i.e. since the initial research work in 2020 until today.
A conclusion to your AI solution:
Test automation is the number 1 bottleneck in many software projects today. Using our AI, it is possible to simulate a tester on different devices, independent of the platform, in order to save manual work.
One final piece of advice/tip to other entrepreneurs looking to apply AI:
You should always think problem-oriented and not want to apply AI just because it sounds good. The focus should always be on the customer benefit. A good test for this is to present the application to the customer and not mention anywhere that AI is being used. In this way, it quickly becomes clear whether it is worth using AI or whether classic approaches are better.