A prompting technique that asks Claude to explain its reasoning step by step before giving a conclusion. For tasks requiring multi-step judgment, this produces more accurate results than jumping straight to an answer — and makes it easier for you to spot where the reasoning went wrong.

Providing 2–5 example 'input → output' pairs in your prompt so Claude learns your desired format, tone, and logic from examples, then applies the same pattern to your actual task. Faster and more accurate than describing 'what format I want' — examples are understood better than rules.

Instead of treating Claude as a 'machine that must give perfect answers in one shot,' treating a conversation as a multi-round collaborative process — get an initial output, then use follow-up instructions to refine iteratively until you reach what you actually want. Like having a back-and-forth discussion until satisfied, rather than expecting someone to say exactly what you need in their first sentence.

Explicitly telling Claude in your prompt what your answer should look like — including format (bullets, table, paragraph), length (word count, point caps), and structure (which sections to include). This is the single biggest variable affecting how immediately usable Claude's output is. More specific specification = more directly usable output.

A pre-structured prompt framework with placeholder variables, enabling consistent high-quality outputs by simply swapping in new content each time.

Explicitly telling Claude 'who it is' in your prompt, so it thinks and responds from a specific identity, professional perspective, or role. A well-crafted role setting dramatically improves output quality — far more effective than just saying 'please help me...'

Instructions given to Claude before the conversation begins, setting its role, rules, and behavioral framework. It stays active throughout the entire session and shapes tone, scope, and style.

Temperature is a parameter controlling the 'randomness' of AI output, ranging from 0 to 1 (or higher). Lower values produce more stable, predictable output; higher values produce more creative but less consistent output. Workplace users typically encounter this parameter only when using the API.

Zero-Shot Prompting means giving Claude a task instruction without any examples, letting it complete the task based on its own training knowledge. Most everyday Claude use is Zero-Shot — you ask a question or give an instruction directly, without needing to provide examples first.