With the increasing scale of offshore wind farm development, maintaining farms efficiently and safely becomes a necessity. The length of turbine downtime and the logistics for human technician transfer make up a significant proportion of the operation and maintenance
(O&M) costs. To reduce such costs, future O&M infrastructures will increasingly rely on offshore autonomous robotic solutions that are capable of co-managing wind farms with human operators located onshore. In particular, unmanned aerial vehicles, autonomous surface vessels and crawling robots are expected to play important
roles not only to bring down costs but also to significantly reduce the health and safety risks by assisting (or replacing) human operators in performing the most hazardous tasks. This paper portrays a visionary view in which heterogeneous robotic assets, underpinned
by AI agent technology, coordinate their behavior to autonomously inspect, maintain and repair offshore wind farms over long periods of time and unstable weather conditions. They cooperate with onshore human operators, who supervise the mission at a distance, via the use of shared deliberation techniques. We highlight several
challenging research directions in this context and offer ambitious ideas to tackle them as well as initial solutions.