These three terms get thrown around a lot in AI articles. They sound technical but the ideas are simple: how many examples do you give the AI, and do you ask it to think out loud? This tutorial demystifies all three and shows you exactly when each one shines.
There is no single "right" way to prompt. Some tasks need a quick instruction. Some need examples. Some need the AI to slow down and reason. Knowing which style to reach for is how you go from a casual user to a confident one.
Zero examples in the prompt. You just tell the AI what to do and let it figure things out from its existing knowledge. This works best for tasks the model has seen many times during training — translation, summarisation, classification of common things, basic Q&A.
A small number of examples — typically two to five — showing the input and the kind of output you want. Then you give a fresh input and ask the model to follow the same pattern. This is the easiest way to teach the AI a specific style, tone, or output structure that it would otherwise not produce on its own.
Ask the model to show its reasoning step by step before giving a final answer. The simplest version is to add the phrase
Let's think step by step.
or "Walk through your reasoning before answering." This dramatically improves accuracy on tasks that require logic, maths, or multi-step decisions.
If you use Zero-Shot for a task that needs reasoning, the AI tends to jump straight to an answer — often wrong. If you use Few-Shot but give bad or inconsistent examples, the model will copy your inconsistencies. If you use Chain-of-Thought for a simple translation, you waste tokens and time. Picking the right style is half the battle.
Classify the sentiment of this review as positive, negative, or neutral.
Review: "The food was great but the service made me never want to come back."
The model has seen thousands of sentiment-classification tasks. It will respond with a label without needing any examples.
Convert product names into a friendly two-line description.
Product: Bluetooth Earbuds
Description: Pop them in, walk out the door, and your music goes with you. Light, wireless, and ready when you are.
Product: Insulated Steel Water Bottle
Description: Cold drinks stay cold for hours. Hot drinks stay hot. The bottle does the work so you don't have to.
Product: Foldable Travel Backpack
Description:
The AI sees the pattern: short, friendly, two lines, present tense, ends warmly. It will produce a new description in exactly that style.
A library has 240 books. 30% are fiction. Of the fiction books, half are
mystery novels. How many mystery novels does the library have?
Think through it step by step before giving the final answer.
The model will write out: 30% of 240 = 72 fiction books. Half of 72 = 36 mystery novels. The reasoning makes the answer trustworthy and easy to check.
Note: Modern reasoning models (like recent versions of GPT and Claude) often "chain-of-thought" automatically behind the scenes. You may not always need to ask for it explicitly — but on tasks that look mathematical or multi-step, doing so still helps.
Pick a classification task you care about (e.g. categorising your emails as urgent / not urgent). Try it Zero-Shot first. If the results are inconsistent, add three labelled examples and try Few-Shot. Compare.
Ask an AI a multi-step word problem without saying "think step by step". Then ask the same problem with that phrase added. Notice how often the answer changes — or how the explanation becomes clearer.
Build a Few-Shot prompt that teaches the AI to rewrite plain headlines into clickable but honest ones. Give it three good examples and one new headline to rewrite. Pay attention to how well it copies your style.
Sign in to join the discussion and post comments.
Sign inPrompt Engineering for Developers
Use AI as your coding co-pilot. 18 tutorials on writing prompts to generate clean code, debug faster, write tests, build APIs, and ship better software.
Prompt Engineering for Business & Productivity
Use AI to work smarter — automate tasks, make better decisions, and communicate professionally. 12 practical business prompt tutorials for professionals.
Prompt Engineering for Education & Learning
Use AI as your personal tutor. Learn how to study faster, create lesson plans, generate practice questions, master languages, and prepare for competitive exams with smart prompts.
Prompt Engineering for Specific AI Tools
Tool-by-tool mastery — deep dives into ChatGPT, Claude, Gemini, GitHub Copilot, Midjourney, Stable Diffusion, and more. Learn the exact prompting techniques each platform rewards.
Prompt Engineering for Data Science & Analytics
Supercharge your data workflows with AI. 15 practical tutorials on using prompt engineering for data cleaning, EDA, machine learning, SQL, visualisation, and more.
Advanced Prompt Engineering Techniques
Master the powerful techniques AI experts use every day. Chain-of-thought, RAG, agents, function calling, prompt evaluation, and much more — 20 deep-dive tutorials.