Blog

AutomationSTAR 2026: How AI Is Transforming Quality Assurance

Written by Carina Ernst | Friday, 20.11.2026

For example, AI can support teams in quality assurance by preparing or executing test tasks. To ensure reliable results, however, decisions must be transparent and responsibilities clearly defined. This is one of our key takeaways from both practical experience and the AutomationSTAR 2026 conference.

From November 4–5, 2026, we—Jürgen Volles, Benedict Wolf, and Carina Ernst—attended the conference organized by EuroSTAR Conferences in Antwerp. Through presentations, at exhibition booths, and in discussions, we explored AI-assisted testing, new tools, and the requirements for their responsible use.

 

Two tasks: AI in the testing process and testing AI systems.

The discussion about AI and software quality has two sides. On the one hand, teams can use AI to assist with testing tasks. However, they must also test applications that use AI or make decisions themselves. Both of these tasks present different challenges for quality assurance.

When testing AI systems, it is not enough to simply verify the expected outputs of individual components. Test teams must also evaluate how a system behaves under different conditions and the potential consequences of errors. While traditional testing approaches remain important, they must be supplemented depending on the application.

 

From Automated Workflows to AI Agents

Traditional test automation primarily involves executing tests according to defined rules. AI agents, on the other hand, can plan tasks and use tools within specified limits. The degree to which they act autonomously depends on their implementation and the permissions granted.

At the conference, we also explored multi-agent systems, in which multiple agents collaborate to complete tasks. Another key topic was the Model Context Protocol (MCP). This open protocol allows AI applications to connect to external tools and data sources in a standardised way.

This raises a practical question for quality teams: Which tasks can such systems meaningfully support, and at what points must humans review or approve decisions?

 

Greater Autonomy Requires Clear Oversight

AI-supported testing methods can free up teams from routine tasks and help to identify additional test cases. Whether these methods improve test coverage or reduce maintenance efforts depends on the specific use case. AI-generated suggestions and results must also be reviewed.

When agents select test tasks, the basis on which they are acting should be transparent. If they access tools or data, they must have the appropriate permissions and limits. It must also be clear who is responsible for evaluating results ('human-in-the-loop') and making decisions regarding critical actions. MCP alone does not replace these safeguards.

Governance refers to the framework of responsibilities, rules and controls for this application. This does not mean the end of the role of testers, however: human expertise remains essential for assessing risks, designing tests, and interpreting results.

 

Insights for testOFFICE and Our Future Work

The exhibition and interaction with other providers, users and the international testing community has provided us with new perspectives. We’ve also reflected on how we present our added value and which features of testOFFICE to improve, as well as considering which ideas might be of interest to our users.

We will take many of these insights with us to inspire our projects and the continued development of testOFFICE. The most suitable approaches for testOFFICE also depend on the needs of our users.

 

Conclusion: Use AI Strategically and Evaluate Results

AutomationSTAR has shown us just how intensely the testing community is discussing AI agents and new forms of automation. The important thing is not to achieve as much autonomy as possible, but to use AI in a way that aligns with the company’s quality objectives, the risks of the industry and the applicable regulations.

Our practical takeaway: First, define a suitable test task, establish access rights and responsibilities, and review the results. This enables you to evaluate whether AI is actually supporting the team and whether quality assurance remains reliable.

 

Would you like to know more about using AI in testing or testing AI systems?
Talk to us about how we can support your quality assurance and traceability efforts.