At Getty we identified an escalating business problem concerning asset inspection costs. Their were too many assets coming into our pipeline to keep inspecting queues for good images manually. We hypothesized that if we created an interface where inspectors could search for bad assets among a filtered set of suspect images rather than looking through all images for good images inspection counts would increase dramatically.
To achieve this we started scoring all assets before they got to the inspection stage, assets above a certain threshold are auto-approved and those below the threshold are put into an inspection queue. The inspection queue is then reviewed by a Getty Images inspector looking for "bad" images. Bad images are flagged and sent off to another process. The remainder of images in the queue are then bulk approved and published. This project is saving Getty Images over one million dollars in 2018 and inspector feedback has been overwhelmingly positive.