With ChatGPT, hugging face etc, text based searching is already outdated.  I see there are a number of ATS out there racing to use these new tools to do things like upload resumes, scan and split the text and then store them into vector databases to allow for semantic search.  This way we can do natural language search on resumes, rather than field driven.  Combined with candidate communication and recruiter notes, this is a much better and more efficient way to search candidates or have the ATS recommend candidates.  

Benefits: You can use chat to find candidates with certain types of experience in your database, and it will also learn as it goes.  It can also remember candidates preferences and recruiter notes from communications and include that in its recommendations.  It can also enrich the database of candidates by looking up information on their companies, etc.   The chat can also communicate with candidates using chat/email etc using natural language and then update their record based on feedback.    Please tell me this is in the works already, we really don't want to have to migrate to one of these newer LLM based systems.