Endoscopic spinal surgery (ESS) is an effective technique to address a range of spinal pathologies. 70% of spinal procedures are performed endoscopically in Asia compared with less than 10% in the UK. Endoscopic techniques cause far less tissue trauma and blood loss compared with open surgery, which also increases the number of patients who can benefit to include the physiologically frail and obese. Furthermore, procedures can be performed as day cases which accelerates patient turnaround time and reduces the demand for inpatient hospital beds. This is particularly important, post-covid, as waiting lists remain long. Despite these numerous benefits, adoption is limited. This is attributed to the considerable learning curve, unconventional training requirements, and lack of surgeons possessing ESS skills. Computer vision has the potential to overcome these caveats. Developing a machine learning model that can identify the anatomical structures and surgical tools used in the surgery in real-time will support rapid surgeon learning, reduce complications and create opportunities for related technologies such as surgical robotics and autonomous navigation. The model would provide immediate feedback to the surgeon about the surgical environment, enhancing surgical precision. This will improve patient outcomes, and lead to faster recovery times, increasing department capacity.