In modern manufacturing, agility and adaptability are crucial for meeting diverse production demands in high-mix, low-volume settings while ensuring quality. Visual inspection plays a vital role, with robotic inspection cells providing a scalable, automated solution. However, optimizing these systems remains challenging, particularly in minimizing inspection time while accommodating various workpiece geometries. This research presents an agile robotic cell for visual inspection, where the input is a 3D model of the workpiece with predefined inspection poses. The system optimizes the placement of two cameras to minimize total inspection duration, defined as the time required for the robot to present all inspection points to at least one camera. The inspection process is formulated as a state-space optimization problem, where each state represents a combination of an inspection point and a camera. For each state, we compute the required workpiece pose and derive the corresponding robot joint configuration. A fully connected, undirected, weighted graph is constructed, with nodes representing states and edges corresponding to transition times between robot joint configurations. A shortest path algorithm determines the optimal sequence of transitions, ensuring each inspection point is visited exactly once while minimizing movement time. To optimize camera placement, an iterative algorithm adjusts camera locations to minimize inspection duration. The system’s agility enables automatic reconfiguration for different workpieces and inspection needs, making it adaptable to various manufacturing scenarios. Experimental validation demonstrates that the proposed optimization approach reduces inspection cycle time while ensuring full coverage.