AI-enabled medical device software can enhance diagnostics, treatment planning, patient monitoring, and workflow automation. However, these benefits will only be fully realised if relevant stakeholders can trust, and thus adopt this software. Each stakeholder group faces unique trust challenges, for example clinicians can have a lack of trust in AI decisions and particularly in black box AI models, leading to fear of liability and damage to reputation. Similarly, patients have concerns about their safety in addition to concerns about data security and privacy of their sensitive health data. To address this issue, we have developed a process reference model and a process assessment model for AI-enabled medical device software, the purpose of which is to assist organisations to develop trustworthy and regulatory compliant AI-enabled medical device software. These models help ensure compliance with industry standards and best practices and improve process maturity by identifying gaps and areas for improvement. In this paper we present an overview of the software level processes contained within a new process reference/assessment model for AI-enabled medical device software. This new process reference/assessment model contains amendments to 8 existing traditional software development and support processes, and the addition of 3 new processes.