What it is

An embedding is a way of turning text (or images, or audio) into a list of numbers that captures its meaning. The word "king" becomes something like [0.2, 0.8, -0.1, ...] and "queen" ends up with similar numbers because they're related concepts. It sounds abstract, but it's how AI actually "understands" things. It doesn't read words like you do. It converts everything into maths and works with the numbers. Two things with similar embeddings have similar meanings.

Why it matters for your job

Embeddings are the reason AI can do things like find documents related to your question, group similar customer complaints together, or recommend products. If your work involves any kind of categorisation, search, or pattern-matching across text, embeddings are the technology doing the heavy lifting. Understanding this helps you see where AI can slot into your workflow and, importantly, where it might struggle.

What to do about it

When an AI tool gives you a weird or irrelevant result, it's often an embedding problem. The AI's mathematical representation of your query didn't match what you actually meant. Learning to rephrase your questions for AI tools is a small skill with outsized impact.