What does it take to run a successful lesson using educational robots? Working with schools and robots for over 40 years, I noticed many teachers achieve remarkable results with them, but others did not. I spent years trying to find what made the difference. Then, I discovered the work of Professors Paul Black and Dylan Wiliam on Assessment for Learning (AfL). I found a good match between how experienced teachers worked and AfL methods. Assessment for Learning helps teachers manage lessons using robots effectively—this paper is the 4th in a series where I look at AfL with robots. Here, I review Success Criteria in detail: How do they enrich lessons, and how do successful teachers use them? How do they support learning with robots, and how do some unnecessary bureaucratic problems restrict their effectiveness? We cannot measure achievement with simplistic measures like a grading system; it’s highly personalised and manifests in many different ways. We cannot measure achievement with simplistic measures like a grading system; it’s highly personalised and manifests in many different ways. Working with robots offers a far more nuanced and diverse way of recognising success and progress. Understanding what it looks like guides our ability to develop ways of using robots in education and how to help teachers get the most from the technology.