ISO 29119-8 Enters Publication: What the New Model-Based Testing Standard Means for QA Teams

Model-based testing is getting a formal place in the ISO 29119 family

Teams have used model-based testing for years to derive test cases from models of system behavior, workflows, states, and business rules. The approach is now moving toward a formal international reference point. ISO/IEC/IEEE 29119-8, titled Software and systems engineering, Software testing, Part 8: Model-based testing, entered the publication stage on September 28, 2026. ISO currently lists the standard as under publication while final production work continues. ([iso.org](https://www.iso.org/standard/89908.html?utm_source=openai))

The listing is not a minor status update. The document is not yet a fully published standard that organizations can cite as a completed edition. But stage 60.00 means the technical work and approval process have moved well beyond an early proposal or committee draft. ISO says final production can take up to seven weeks, so publication could follow soon if the remaining steps proceed as expected. ([iso.org](https://www.iso.org/standard/89908.html?utm_source=openai))

What ISO 29119-8 is designed to cover

The new part sets out requirements and guidance for using model-based testing within the test processes defined by ISO/IEC/IEEE 29119-2. It also explains the relationship between model-based testing and the test documentation practices in ISO/IEC/IEEE 29119-3, as well as the test design techniques in ISO/IEC/IEEE 29119-4. ([iso.org](https://www.iso.org/standard/89908.html?utm_source=openai))

The intended users are teams that use a behavioral or structural model to generate testware. That model could describe permitted states in an account-management system, transitions in a medical device, user journeys in a web application, or combinations of rules in an embedded system. Teams then use it to derive test conditions, test cases, or other test artifacts.

ISO's description assumes automated test execution, but the document does not require a particular automation framework or modeling tool. The generation algorithm's implementation is outside the standard's scope, as is the choice of a specific model-based testing product. ([iso.org](https://www.iso.org/standard/89908.html?utm_source=openai))

What the standard does not mean for tool buyers

29119-8 does not endorse a vendor, notation, or generation algorithm. It will not direct organizations to a particular platform, and it will not make every generated test useful by default. Tool selection remains an engineering decision. The model's quality still determines much of the value the process produces.

That point matters as testing vendors market automated test generation more aggressively. A standards-based process can improve consistency and documentation, but teams still need domain analysis, risk assessment, review, and maintenance. An incomplete state model may generate thousands of tests while overlooking the behavior that matters most to customers.

What QA leaders should do now

• Inventory existing models. Find the state diagrams, workflow definitions, decision tables, and domain models that development and testing teams already use.
• Choose one bounded pilot. Start with a business-critical workflow that has clear states and transitions rather than attempting to model an entire product.
• Measure more than test count. Track meaningful transition coverage, defect detection, execution time, and the effort needed to update the model after a product change.
• Separate generation from judgment. Automated test creation can expand coverage, but testers must still decide which scenarios are realistic, risky, redundant, or unsafe to execute.
• Review tool claims carefully. Ask vendors how their products represent models, preserve traceability, handle change impact, and export evidence for audits or release decisions.

ISO/IEC/IEEE 29119-8 will not replace exploratory testing, API testing, security testing, or production monitoring. Its value is narrower: it gives organizations a common framework for applying model-based testing alongside the established ISO 29119 processes.

For QA teams already experimenting with generated tests, the next move is not to automate everything. Make the model explicit, connect it to risk, and test whether the resulting cases find important failures. After the standard is fully published, that discipline will give teams a firmer basis for comparing internal practices, supplier claims, and model-based testing tools.