Test Automation Improvement Model (TAIM), has been used to guide assessments and describe patterns for aspects of automation in testing. In this paper, we have updated TAIM for the next generation of autonomous software and complemented the model to use as a driver for software quality. We bring some lessons learned and describe the evolution of the model. TAIM is heavily influenced by analytics approaches, i.e. AI or machine learning, as we now strive for autonomous systems. We describe the new levels in TAIM, their focus, quality and cost aspects. Key areas are addressed and some experiences of using the TAIM process and suggestions for further development are proposed.