Chain-of-thought prompting tells the model to show its reasoning before giving a final answer. The simplest version is the phrase "Let's think step by step." More structured versions ask for explicit numbered steps or scratchpad reasoning inside XML tags.
Why it works: extra reasoning tokens give the model more "compute" to spend on hard problems. Why to be careful: CoT roughly triples token usage and latency, so use it where it matters (math, multi-step reasoning, code) and skip it where it doesn't (classification, extraction, short answers). Modern reasoning models (o1, o3, Claude with extended thinking, Gemini Thinking) bake CoT in at the model level.
Bring this to your business
Knowing the term is one thing. Shipping it is another.
We do two-week AI Sprints — one term, one workflow, into production by Day 10.