During the last years, EU manufacturing has faced production flexibility challenges by deploying, among others, novel hybrid manufacturing systems, involving collaborative robots and mobile manipulators combined with flexible grippers, vision systems, sophisticated tasks/actions planning solutions and flexible integration platforms. Despite the importance of AI enabled flexible robotic systems, several aspects settle back their wider adoption, and impact on the objectives of the green deal: •Limited cognition/ intelligence: existing solutions support non-trivial tasks but cannot act autonomously. •Insufficient perception and diagnostics: In a circular economy, there is an increased need for understanding the state of products or parts that are being handled, after they have been used. •Decision making is restricted: Current decision-making focuses on process or line level, not taking into account optimization at value chain level or per individual product. •Small scale adaptation of AI due to small number of available data and training needed, to support tailored solutions in high variability context. •Lack of use of explicitized knowledge in AI and robotics. Lifecycle data and knowledge is not used across the value chain to improve decision making after a product’s first life. •Complexity in robot programming and interaction which requires the involvement of skilled engineers, does not provide flexibility in execution, Thus, ROB4GREEN aims to develop easy to use and deploy AI driven collaborative robotic systems, that can reason and adapt to a variety of strategies for processing products after their first life, both hardware and behavior wise, improving existing skills and generating new ones, working autonomously combining data and knowledge. Such systems will be validated at scale and in major industries, showcasing optimization ranging from cell to the whole value chain, towards achieving significant impact on the objectives of the green deal.