An embedding is a list of numbers that stands for the meaning of a piece of text, an image or other data. Items with similar meanings get similar numbers, so a computer can tell they belong together even when the wording is different. For example, a help request saying “my laptop won’t turn on” and another saying “the computer doesn’t start” would get embeddings that sit close to each other, even though they share almost no words.
Tools use embeddings to search by meaning, to suggest related items, and to hand relevant documents to an AI model through retrieval-augmented generation. The embeddings are often kept in a vector database.
Why it matters
Embeddings look like meaningless strings of numbers, but they can sometimes reveal information about the text they came from. If donor records, case notes or staff emails are turned into embeddings, treat those embeddings as being as sensitive as the original files. Before you connect a document search or chatbot tool to your files, ask the vendor where the embeddings are stored, who can access them, how long they are kept, and whether they are deleted when you delete the source documents.
Published Updated