Generative systems produce designs at high volume and speed, but they also tend to produce higher rates of manufacturability issues, tolerance violations, and documentation gaps compared to designs built by experienced engineers. AI design review tools are well-suited to catch exactly those issues automatically, at the point of generation, before a design reaches a human reviewer. A practical near-term workflow would have generative design produce candidates based on constraints, AI design review analyze those candidates for manufacturability and standards compliance, and engineers evaluate the filtered set to decide which direction to develop further.