Embeddings
Embeddings turn text into lists of numbers that capture meaning, so similar ideas land near each other mathematically: "invoice overdue" close to "payment reminder", regardless of shared words. They're how machines search by meaning instead of keywords.
Why it matters
Embeddings power the useful magic in modern tools: search that understands intent, "related articles" that relate, support systems that find the right answer however the question was phrased. If RAG is the dish, embeddings are the knife work.
How it works
A model maps each text chunk to coordinates in a high-dimensional space where distance means semantic difference. Your content gets embedded once and stored; each query is embedded and matched to its nearest neighbours: meaning-based retrieval in milliseconds.
What to do about it
Nothing to buy directly, but when evaluating "AI search" or knowledge tools, ask whether retrieval is semantic (embeddings) or keyword-based. It predicts how well the tool handles humans phrasing things humanly.
