To reduce the number of software tests without losing confidence in a release, prioritise tests by business risk. Look at how likely a scenario is to occur and how serious the consequences would be if it failed. This helps teams focus testing on critical workflows instead of treating every test as equally valuable.

In his webinar, Tobias Müller challenges a familiar assumption in software testing: that more tests automatically mean better coverage. A large suite can produce plenty of green checks and still leave teams unsure whether a release is ready. The more useful question is whether the tests cover the failures that would matter most to the business.

Why test volume does not tell the whole story

Test counts are easy to measure. Release readiness is harder. A thousand tests may sound reassuring, but the number alone does not show which business processes they cover, how important those processes are, or what a failure could cost.

Traditional test coverage can show which functions have been tested. Business risk coverage adds another question: have the tests focused on the parts of the product where failure could cause the greatest harm?

That shift is at the heart of risk-based testing. We’ve also explored why, despite its appeal, risk-based testing can be difficult to put into practice.

Prioritise tests by likelihood and impact

Tobias illustrates the approach with a bank account opening process. An identity check and a risk score calculation can have serious consequences if either fails. A problem with marketing emails may be inconvenient, but it is likely to carry a different level of business risk.

By considering how often each scenario occurs and the potential impact of failure, teams can make more deliberate choices about what to test first. In Tobias’s example, the first three prioritised tests cover close to 80% of the identified business risk. That figure belongs to the example, but the lesson applies more broadly: a smaller, well-prioritised test suite can offer more useful release insight than a larger suite with no sense of priority.

Use generative AI to identify risks, not just generate tests

Generative AI can help teams analyse specifications, identify scenarios and map possible risks. But generating test cases is only one step. Teams still need to decide whether the tests cover the right things, and keep them useful as the software changes.

This matters because a growing suite can also create more work to review and maintain. We look at that challenge in “AI-generated test cases are easy, maintenance isn’t”. In the webinar, Tobias demonstrates how an AI agent can start with product documentation and help teams build a prioritised view of scenarios and risks.

What comes next for test automation?

The session also explores visual AI testing. Instead of relying only on technical identifiers, visual systems interpret what appears on screen, including labels, layout and interface elements. Tobias demonstrates how a model can recognise elements and account for values that change between runs.

He also discusses how teams can test AI-generated actions. When an AI system can produce different outputs, checking for one exact response may not be enough. Teams can define expected behaviours and evaluate whether the generated actions meet them. For more on visual testing, see our article on Visual Sense and UI test automation.

The webinar’s practical takeaway is that test automation should help teams answer a real release question: which important risks have we covered, and what do the results tell us?

Watch the webinar:

Frequently asked questions

How can teams reduce the number of software tests without reducing confidence?

Prioritise tests based on the likelihood and impact of failure. Keep tests that cover critical business risks, and review whether lower-impact tests add useful information about release readiness.

What is business risk coverage?

Business risk coverage considers whether testing addresses the failures that could have the greatest impact on an organisation. It adds business importance to the view of which functions or scenarios have been tested.

How can generative AI help with risk-based testing?

Generative AI can analyse specifications and help identify scenarios and potential risks. Teams can use those results to prioritise test cases, while reviewing the AI’s suggestions and maintaining the tests over time.