Future military forces and platforms will need to operate within increasingly complex operational environments where the threats are more diverse and increasingly challenging. This will increase the volumes and speed of data and information that platforms and their commanders will need to capture, process and respond to. This inevitably leads to the consideration of greater levels of automation and the wider use of machine intelligence techniques across the Command and Control (C2) space. Research in this area to date has focused on the integration of intelligent machine agents, and automated systems, with human operators in specific or focused capability areas. To generate true operational advantage from the growing and diversifying data and information available to military commanders, it will be critical to build on these developments by addressing the design and operation of teams of multiple intelligent machine agents (e.g. as collaborative Artificial Intelligence (AI)) and to enable and optimise the integration of humans within those teams to form effective human-machine teams. The UK Defence Science and Technology Laboratory (Dstl - part of the UK MOD) has funded the Intelligent Ship project, which aimed to explore these future Human-Machine Machine-Machine Teaming (HM3T) architectures and relationships, and the approaches needed to enable them. This was undertaken through the combined developments of a range of component intelligent machine agents and through the development, evaluation and demonstration of those agents, with human operators, within a systems level architectural ‘sandpit’ known as the Intelligent Ship AI Network (ISAIN). This work was delivered by a multidisciplinary team of Dstl and 10+ suppliers and evaluated within Dstl’s Command Lab facility. This paper will review the project’s aims, delivery approach, lessons learnt and challenges identified in its second phase, which completed in March 2022. This phase included a series of evaluation events, with each event growing in complexity and in the number of interacting agents and operators. This paper will overview the system level architectures developed and highlight the range of agents developed (providing threat evaluation through to system control) and integrated. Finally, it will review future development needs.