Prompting 101: The Complete Guide to Writing Prompts That Actually Work

If you've ever typed something into ChatGPT and got back a generic, useless wall of text, you've experienced the gap between using AI and prompting it well. The good news: prompting is a skill, not a talent — and it's one you can learn in an afternoon.

This guide synthesizes the free "Prompting 101" track from OnePlaceForAI (10 lessons, ~118 minutes of material) into one practical reference you can keep open while you work. By the end, you'll know exactly how to structure any prompt, which technique to reach for when output goes wrong, and how to build a personal library that compounds your productivity over time.

---

Lesson 1: What Is a Prompt? (Stop Searching, Start Briefing)

Most people treat AI like a search engine — type a few keywords, hope something useful comes back. But an AI assistant is fundamentally different. A prompt is not a query. It's an instruction, a brief, a set of directions.

A useful mental model: imagine you have a brilliant new intern who has read almost everything ever published, but has no idea what you specifically need right now. They'll do exactly what you ask — no more, no less. Vague input → generic output. Clear direction → excellent work.

Compare these three prompts for the same topic:

- Vague question: "Tell me about marketing." - Clear instruction: "Write a 200-word product description for a standing desk aimed at home office workers." - Fully briefed: "I am launching a coffee brand targeting remote workers aged 25–40. Write me three tagline options that feel energetic but not corporate. Keep each under ten words."

All three are "prompts." The third consistently wins — because it gives the AI more to work with. That's the entire game.

One practical note: most AI tools have no memory between sessions. Each new conversation starts fresh. Within a single conversation, though, the AI remembers everything — so great results come from a back-and-forth, not a single exchange.

---

Lesson 2: The Anatomy of a Great Prompt (The 4-Element Framework)

Great prompts follow a structure. Once you know it, you can apply it to almost any task. The four elements:

Element	What it does	Example
Task	The core instruction — a clear verb	"Write a marketing email promoting our summer sale."
Context	Background that makes output relevant	"Audience: founders overwhelmed by complexity."
Format	How you want it structured	"Short hook, three benefits, CTA. Under 200 words."
Constraints	What to avoid	"Avoid buzzwords like 'game-changing' or 'revolutionise'."

You don't always need all four. But when output misses the mark, one of these is almost always missing — the framework is a diagnostic tool.

Weak: "Write me a LinkedIn post about our new product."

Strong: "Write a LinkedIn post announcing our new project management tool for small businesses. The audience is founders and team leads who are overwhelmed by complexity. The tone should be confident but conversational — not corporate. Use a short hook, three benefit sentences, and a call-to-action. Under 200 words. Avoid buzzwords like 'game-changing' or 'revolutionise'."

The second takes thirty seconds longer and often works on the first try.

---

Lesson 3: Context — The Most Powerful Ingredient

Of the four elements, context is the one beginners leave out most — and the one that makes the biggest difference. It transforms generic output into something tailored to your situation.

There are three types of context:

1. Audience context — Who will read this? The same topic explained to a ten-year-old looks nothing like the same topic for a PhD. "The audience is small business owners with no marketing background." 2. Purpose context — Why does this exist? A product description for a homepage serves a different job than one in a sales email. "This email is a follow-up to a warm sales call — the lead isn't committed yet." 3. Background context — What does the AI need to know about your specific situation? "I run a small bakery in Edinburgh specialising in gluten-free goods. My customers are regulars aged 35–60."

Rule of thumb: if the context would change the output, include it. The AI is excellent at filtering what it doesn't need — but it can't invent context you never provided.

---

Lesson 4: Giving AI a Role (The "Act As" Technique)

Role prompting — assigning the AI a persona before the task — is one of the most effective techniques there is. A role shifts vocabulary, depth, priorities, and framing all at once.

Generic role: "You are a marketing expert." → generic output. Specific role: "You are a direct response copywriter with fifteen years of experience writing email campaigns for e-commerce brands. You write copy that is clear, benefit-focused, and always leads with the reader's problem." → distinctive, useful output.

The specific version gives the AI a fully formed character — a point of view, a style, a set of priorities — not just a job title.

Pro move: you can also give yourself a role. "I'm a first-year law student. Explain this contract clause in simple terms." This pitches the response at exactly the right level.

Keep roles grounded in real expertise. "World's greatest copywriter who has never failed" is weaker than "experienced direct response copywriter." Inflated roles produce weaker output.

---

Lesson 5: Controlling Format and Length

When you don't specify format, the AI guesses — and often guesses wrong. Format instructions are among the highest-return additions you can make.

Structure: "Present this as a bullet list." / "Format as a numbered step-by-step guide." / "Use a table with three columns: Feature, Option A, Option B."

Length: Be concrete. "Under 100 words" works far better than "be concise." The AI interprets "be concise" loosely; it hits "under 80 words" reliably. Use word counts, sentence counts, or item counts.

Form: "Write this as a professional email." / "Format it as a LinkedIn post." / "Give me a script I can read aloud." Form changes register and conventions, not just structure.

Plain text: If you're pasting output into Word, a CRM, or an email editor that doesn't render markdown, add "No markdown formatting, plain text only" — the output arrives clean.

Multiple format instructions stack reliably in a single prompt.

---

Lesson 6: Constraints That Actually Work

Telling the AI what to avoid is as powerful as telling it what to do. "Avoid," "do not," and "never" all work reliably.

Cut filler: Left alone, the AI opens with "Certainly! I'd be happy to help" and closes with a summary it just wrote. One constraint each: "Do not start with a preamble. Get straight to the output." and "No closing summary."

Tone constraints name specific patterns to eliminate: "Do not use corporate buzzwords like 'synergy' or 'leverage'." / "Avoid hedging language — no 'it seems' or 'one might consider'."

Content constraints prevent unwanted inclusions: "Do not mention competitors by name." / "Avoid technical jargon — the reader has no coding background."

A default constraint set worth memorizing for most professional writing: 1. No filler phrases — start directly. 2. No closing summary. 3. Avoid hedging language. 4. No buzzwords.

Constraints sharpen positive instructions — they work together, not as replacements.

---

Lesson 7: Iterating Your Way to Great Output

The first output is a draft, not a finished product. This is the single biggest mindset shift. Professional users write a prompt, review the output, identify what's off, and send a follow-up. That process — iteration — is what separates consistently great results from giving up after one try.

Follow-up prompts are almost always shorter than the original — the AI remembers the conversation. Be precise about what to change: "Make it shorter — under 80 words." / "The tone is too formal. Rewrite it more conversational."

Four iteration modes: - Refine — small adjustments ("Shorten the opening sentence.") - Redirect — wrong angle ("Too sales-heavy. Make it educational.") - Expand — add more ("Expand the second section with more detail.") - Condense — cut fluff ("Cut this to half the length.")

One of the most powerful follow-ups: "Give me five alternative versions of that headline, ranging from direct to playful." AI generates options exceptionally fast.

When to start fresh: if you've sent 4+ follow-ups and it's still wrong, or the output type is fundamentally wrong, abandon and restart with a better initial prompt. Use what you learned.

---

Lesson 8: Few-Shot Prompting (Show, Don't Tell)

Instead of describing the style you want, show it with examples. This is called few-shot prompting, and examples convey style far more precisely than words.

Describing your voice: "Write in a direct, concise style. Short sentences. No filler." — helps, but approximate.

Showing your voice: include 2–3 samples of your actual writing, then ask for new content in that style. The AI models your sentence rhythm, word choices, and paragraph structure exactly.

Structure: 1. The examples (1–3 samples) 2. The new task, clearly separated

Input-output pairs for transformation tasks: "I'll give you a feature, you convert it to a benefit. Feature: '256-bit encryption' → Benefit: 'Your data is protected with bank-grade security.' Now convert: 'Real-time collaboration with 50 users'."

Two to three examples hit the sweet spot — enough to show a pattern, not so many the prompt gets bloated. Few-shot pairs beautifully with role + context for output that sounds like you, not the AI.

---

Lesson 9: Chain-of-Thought Prompting (Make It Think Out Loud)

For complex problems — decisions with trade-offs, multi-step math, logical analysis, planning — the AI performs significantly better when asked to show its reasoning rather than jump to a conclusion.

The magic phrase: "Think step by step." That's it. Add those four words to almost any reasoning-heavy prompt and quality improves measurably. Why? When the AI generates reasoning sequentially, each step informs the next, and errors that would slip past in a direct answer become visible mid-chain — letting it self-correct.

Best for: multi-step reasoning, decision-making, math, planning, analysis. Skip it for: simple lookups, writing tasks, straightforward instructions.

You can make it structured: "Analyse this decision using: (1) List key factors. (2) Identify risks per option. (3) Weigh what matters most. (4) Give a recommendation with reasoning." This combines chain-of-thought with format control — organised, reviewable output.

Note: chain-of-thought produces longer output. That's a feature for complex decisions — you want to see the thinking. Use it intentionally.

---

Lesson 10: Building Your Prompt Toolkit

You now have all nine techniques. The final step is turning knowledge into a personal system that compounds.

A reusable prompt is a template with placeholders for what changes: > "You are writing on behalf of [my name/company]. Write a polite, professional email declining [type of request]. Tone: warm but firm. Acknowledge genuinely, give a brief honest reason, close positively. Under 120 words. No hedging phrases like 'I'm afraid' — be direct."

A prompt library doesn't need to be fancy — a notes doc organised by task type (Writing, Research, Planning, Content, Communication), not by date. When a prompt produces reliably great output, save it immediately. A library of fifty tested prompts is worth more than any course.

Quick diagnostic — when output is wrong, reach for: | Problem | Fix | |---------|-----| | Generic or irrelevant | Add more context | | Wrong tone or expertise | Adjust the role | | Wrong structure | Add format instructions | | Filler or bad habits | Add constraints | | Close but not quite right | Iterate with a follow-up | | Need to match a style | Use few-shot examples | | Complex decision | Add chain-of-thought |

The compound effect: for two weeks, every time you need to write, research, or decide something, open an AI tool and prompt it — even if you think you could do it faster yourself. The point is practice. After two weeks, you'll have a collection of prompts that work for your tasks and the habit of reaching for AI before the blank page.

---

Your Action Plan

1. Start with the framework — Task, Context, Format, Constraints in every non-trivial prompt. 2. Add a role when you need specific expertise or tone. 3. Control format and length with concrete numbers, not "be concise." 4. Add constraints to kill filler and protect your voice. 5. Iterate — treat the first output as a draft. 6. Use few-shot when you need to match a style exactly. 7. Trigger chain-of-thought for anything involving real reasoning. 8. Save your wins to a personal prompt library.

Prompting isn't magic — it's clear communication with a very capable but very literal tool. The quality of your output is in your control. Start where you are, use what you have, and build the habit.

This guide synthesizes the free "Prompting 101" learning track from OnePlaceForAI — ten lessons covering everything from the anatomy of a prompt to building a personal prompt toolkit.

---

Ready to put this into practice? Pick one task you do regularly — a type of email, a report summary, a social post — and write one great prompt for it this week. Test it, iterate until the output is reliably good, and save it. Then pick another next week. In a month, you'll have a toolkit that makes you measurably faster at everything you write, research, and decide.