We want to build on our existing vector similarity logic in the platform and see where else we can use images to automate some areas of the site that are