Get in touch with us
We at accompio will be happy to help you.
02.10.2026
Test automation can ensure software quality and significantly accelerate regression testing. At the same time, creating and maintaining automated tests is time-consuming: From a technical description such as „Log in and then open the shopping cart,“ a large number of technical information must first be derived.
This is exactly where it starts AI-based test automation Instead of having to describe test cases in a rigid syntax or with detailed technical specifications, AI should generate executable and at the same time maintainable test code from professionally formulated test steps.
Daniel Horn has been working for many years in quality assurance and test automation. As a product owner, he is responsible for the development of accompio. QAptain, a tool that is intended to bridge the gap between manual test cases and automated test code.
In the full expert interview speaks Daniel Horn about QAptain and how AI can bridge the gap between technical test cases and maintainable test code.
In a classic case Test automation A technical test description often isn’t enough. A test automation engineer must figure out for each individual step which element is being addressed, how it is technically identified, and what code is needed for that.
With longer end-to-end tests, this effort can become significant.
AI can serve as a bridge here: The test case is still described in technical terms – for example, that a user should log in and then open the shopping cart. The AI then takes over part of the translation and integrates it into the technical test automation.
The goal is explicitly stated no, replacing the technical description with a new technical syntax. Instead, natural language should be brought closer to the executable test.
Automatically generated code is not automatically good code. Test automation code must also be maintainable in the long term. Therefore, similar requirements apply to Daniel Horn as to conventional software code: Readability, clear structures, reusability, and meaningful abstraction.
The Page Object Pattern plays an important role in this context. Technical details such as the precise identification of a login field are stored centrally. The actual test can therefore be formulated at a higher technical level. Instead of simply representing individual technical clicks, the test can, for example, simply invoke the process of „logging in with these data“.
This not only makes automated tests more understandable, but also reduces the workload for maintenance. For example, if the technical ID of an element changes, this adjustment ideally only needs to be made at a central location – and not in hundreds or thousands of test cases.
„Test automation code should meet the same quality requirements as my application code.“
Daniel Horn, Principal IT Consultant at accompio
During the generation of test code, human control remains an important component of the process. The code generated by the AI is initially reviewed and, for example, incorporated into the existing development process via a pull request. Only after this review is it accepted.
The reason is simple: Code generation can be prone to errors. The person must continue to assess whether the test actually does what is expected from a professional point of view.
This also changes the role of the test automation engineer. Less time is spent on the purely technical creation of individual code lines. Instead, more resources can be invested in developing more meaningful test cases, reviews, and analyzing risks.
With QAptain This very same connection is intended to be created. Technical test cases can still be described in a way that is understandable to a manual tester. The AI then translates this description into the technical structure of an automated test.
The goal is not to replace all test automation with AI. Rather, it aims to Technical code creation accelerated will be.
In the future, this approach could be even more integrated into the software development process: Test cases and acceptance criteria could already be defined during the requirements analysis or the user story refinement phase. When development subsequently takes place, the corresponding test code could be generated automatically from this process.
In the long run, this will also change the role of quality engineering. The technical implementation of test automation could move more towards the development teams. At the same time, the professional analysis is gaining importance: What requirements exist? What scenarios need to be tested? What risks exist?
Each user story and acceptance criterion can be considered a starting point for a test. The goal is a more integrated process in which Development and quality assurance no longer operate separately, But tests are part of software development from the very beginning.
accompio supports companies in software testing and offers expertise for the future in AI-based test automation.
AI can accelerate a significant portion of the technical work involved in test automation. Automated code can be created from technically described test cases without requiring any manual intermediate steps.
However, for this approach to work, it takes more than just one AI generator: Maintainable code structures, established automation concepts, and human quality control remain crucial.
The future of test automation therefore does not lie in completely removing humans from the process. Rather, AI can help reduce technical routine work and give quality engineering teams more time for tasks that require technical understanding and experience.
Daniel Horn is a Principal IT Consultant at accompio and, as a Product Owner of the in-house developed solution QAptain, he is the contact person for AI-based test automation.
AI-based test automation uses artificial intelligence to generate automated test code from professionally described test cases. This can speed up the technical implementation of test cases.
AI can generate test code. However, to ensure that it is maintainable in the long term, quality principles such as clear structures, reusability, and appropriate abstractions must be considered. Therefore, human review remains important.
The Page Object Pattern abstracts technical details of a user interface. Elements and their technical identifiers are managed centrally, while test cases can be formulated at a more technical level.
No. AI can primarily speed up the technical code development process. Technical testing analysis, review, risk assessment, and verifying whether a test actually ensures the expected quality remain important tasks for Quality Engineering teams.
