As systems change and grow more complex, there will always be more that we COULD test. Trying to script more automated tests will never scale to test everything, and nor can we crank more manual tests. There will never be enough testers, enough SDETs, or enough time. To avoid devastating bugs, testing must target what it tests based on time and risk. This talk will discuss how I’ve used collaborative, model-based testing to: 1) Identify the impact of new changes made to complex systems, even when they are poorly communicated in requirements. 2) Automatically identify the optimal set of test cases needed to de-risk these changes. 3) Generate and maintain the test scripts and data needed to execute the targeted tests. I will draw on my work in EdTech, insurance, healthcare and banking, showing you how: 1) Requirements modelling shifts quality “left”, catching potentially costly bugs before they enter code. 2) The same models generate the smallest set of test cases needed to “cover” the requirements, reducing test volume without introducing risk. 3) Test generation creates automation scripts on demand, reusing automation libraries to generate automation from the models. 4) Change impact analysis identifies which tests to generate and run as the requirements change, avoiding redundant testing and focusing on at-risk logic. Join me to discover how you can fix quality issues in your requirements, avoid test creation bottlenecks, and execute the tests that are most likely to find bugs in-sprint!