Chain-of-Thought
Chain-of-Thought Prompting (CoT)
In one line
Chain-of-Thought (CoT) prompting asks the LLM to walk through intermediate reasoning steps before giving a final answer — a simple change that meaningfully improves accuracy on harder problems.
Going deeper
Chain-of-Thought prompting nudges the model to spell out intermediate reasoning before answering — sometimes literally with phrases like 'Let's think step by step'. Multiple papers have shown sizable accuracy gains on math, logic and multi-step problems.
Marketers do not invoke it constantly, but it is the rationale behind asking AI for analysis 'with reasoning shown'. Demand only a conclusion and you get plausible-sounding guesses; ask for the steps and consistency improves.
Reasoning-tuned models (GPT-5's reasoning modes, Claude's extended thinking, etc.) now perform CoT internally. You no longer need to spell out 'step by step' — the model will, by default, when the task warrants it.
Sources
Related terms
Prompt Engineering
Prompt engineering is the practice of crafting inputs that steer an LLM toward better outputs — a way to dramatically change result quality without retraining the model.
LLMFew-shot Prompting
Few-shot prompting includes a small number of example inputs and outputs in the prompt itself, letting the LLM imitate the desired format or style without any retraining.
LLMSystem Prompt
A system prompt is the instruction sent to an LLM before any user message, defining the assistant's role, tone and rules — effectively the AI product's character.
LLMAI Alignment
AI alignment is the field — and the practical work — of making AI systems behave in line with human intent, values and safety constraints.
LLMStructured Output
Structured output forces an LLM to reply in a predefined JSON or schema shape instead of free text — essential when you need to plug AI reliably into other systems.