Construction projects in mining are often delayed resulting in cost overruns. Workforce productivity challenges are a key contributor alongside operational complexities such as deeper deposits and higher working temperatures. Accurately measuring workforce productivity for comparison against plan is challenging, especially in underground mining projects. Manual and paper-based tracking systems are outdated, prone to errors and data gaps which limit real-time decision-making in project management. Computer vision is a field of artificial intelligence (AI) that uses machine learning and neural networks to derive meaningful information from digital images and videos. Use of computer vision is widespread, and its applications are constantly evolving. Many industries such as manufacturing, healthcare and transportation have seen significant transformation due to its implementation. Productivity tracking on mining projects may also be served by computer vision to provide real-time data and insights with minimal infrastructure and sensors. An extensive distributed system is not required for installation throughout the mine as is the case with other systems such as location tracking. Using standard surveillance cameras, a computer vision model can be trained to detect any item within an image. These items are then tracked frame by frame and backend logic is applied to contextualise the scene and output the actual activity taking place. This automates the process of productivity tracking in an underground setting. Hatch has developed a web-based application called Hatch Mine VisionAI, which includes event based video playback and dashboards for decision-making. It provides stakeholders with real-time productivity insights to maintain timeline and budget. The solution is designed to: • Identify tasks being performed. • Track task duration and count. • Support identification of safety risks. Hatch Mine VisionAI facilitates sustained increases in productivity, reducing project deviations and cost overruns as enhanced monitoring leads to direct improvements in daily performance. It allows for increased project estimation accuracy and planning, reducing risk of cost overrun. Safety risk is also mitigated through reduced reliance on manual surveillance to perform time studies, allowing project personnel to focus on higher value activities.