Skip to main content

​
Lesson 4: Context is King

​
The Missing Ingredient

In Lesson 3, you learned the Task-Context-Format framework. Now we’re going to dive deep into the most commonly neglected pillar: context. Here’s a scenario. You tell a colleague, “Can you handle the Johnson situation?” If they’ve been in every meeting with you for the past month, they know exactly what you mean. But if they just transferred from another department? They’re lost. Every AI conversation starts fresh. It doesn’t know your job, your project, your preferences, or what happened five minutes ago in a different chat. When you skip context, you’re asking AI to help with “the Johnson situation” when it has no idea who Johnson is. The good news? Once you understand what context AI actually needs, providing it becomes second nature.

​
Core Concepts

​
Why AI Can’t Read Your Mind (And What To Do About It)

Let’s get one thing straight: AI is incredibly powerful, but it’s not psychic. When you start a conversation, the AI knows only what you tell it in that moment. It doesn’t have access to:
  • Your previous conversations (unless you’re in the same chat thread)
  • Your company’s internal knowledge
  • Your personal preferences
  • The project you’ve been working on all week
  • The email chain that prompted your question
Think of it this way: every prompt is like meeting someone for the first time. They’re smart, helpful, and eager to assist, but they’re starting from zero. If you want them to help effectively, you need to bring them up to speed. The fix? Tell the AI what it needs to know. Be the narrator of your own situation.

​
The “New Coworker” Test

Here’s a mental trick that will instantly improve your prompts: Before you hit enter, ask yourself: “Would a brand new coworker understand what I’m asking?” Imagine someone just started at your job today. They’re smart, they’re motivated, but they don’t know the internal lingo, the backstory, or the unspoken expectations. If you walked up to their desk and said exactly what you’re about to type to the AI, would they be able to help you? If the answer is “probably not,” then you need to add more context. Let’s see this in action: Without context (confusing for a new coworker):
“Can you fix the intro?”
With context (now they can actually help):
“I’m writing a fundraising email for our nonprofit’s annual campaign. The intro feels too formal and we want it to sound warm and personal. Can you suggest a more conversational opening that still communicates urgency?”
See the difference? The second prompt tells the AI:
  • What you’re working on (fundraising email)
  • Who it’s for (nonprofit, annual campaign)
  • What the problem is (too formal)
  • What you want instead (warm, personal, conversational)
  • Any constraints (still needs to communicate urgency)
That’s not extra work. That’s giving the AI everything it needs to actually help you.

​
The Four Types of Context That Matter Most

Not all context is created equal. When you’re deciding what to include, focus on these four categories:

​
1. Audience

Who is this for? A CEO reads differently than a teenager. A technical expert needs different language than a complete beginner. Examples:
  • “This is for small business owners who aren’t tech-savvy.”
  • “My audience is parents of middle schoolers.”
  • “I’m presenting to our executive team who has limited time.”

​
2. Purpose

What are you trying to accomplish? Information? Persuasion? Entertainment? Problem-solving? Examples:
  • “I need to convince my manager to approve this budget.”
  • “This should help customers troubleshoot on their own before calling support.”
  • “I want to make a complex topic feel approachable and interesting.”

​
3. Tone

How should it sound? Professional? Casual? Urgent? Playful? The same information can be delivered in wildly different ways. Examples:
  • “Keep it professional but not stuffy.”
  • “This should feel like advice from a trusted friend.”
  • “Match the tone of our other marketing materials, which is upbeat and encouraging.”

​
4. Constraints

What are the boundaries? Length limits, things to avoid, required elements, formatting needs? Examples:
  • “Keep it under 200 words.”
  • “Don’t mention competitors by name.”
  • “Must include a call-to-action at the end.”
  • “Format as bullet points, not paragraphs.”
You don’t always need all four, but asking yourself “Have I covered audience, purpose, tone, and constraints?” is a great checklist before you send a prompt.

​
Structuring Context So AI Doesn’t Get Lost

Here’s something that trips up a lot of people: they dump a huge wall of text at the AI and expect it to sort everything out perfectly. Sometimes it works. Often, it doesn’t. AI processes information in order, and it can get confused when context is scattered, contradictory, or buried in the middle of a long prompt. The solution? Structure your context clearly. A simple structure that works: 
BACKGROUND: [What the AI needs to know about the situation]

TASK: [What you want the AI to do]

REQUIREMENTS: [Any specific constraints or preferences]
Let’s see this in practice: Unstructured (harder to follow):
“So I have this blog post I wrote last week about productivity tips and my editor said it’s too long and also we need to make sure it appeals to remote workers specifically, not just anyone, and can you help me shorten it while keeping the main points and maybe make it sound a bit more energetic? Oh and it should be under 800 words. Here’s the post: [text]”
Structured (much clearer):
“BACKGROUND: I wrote a blog post about productivity tips. My editor says it’s too long and wants it to appeal specifically to remote workers. TASK: Help me shorten this post while keeping the main points intact. REQUIREMENTS:
  • Target audience: remote workers
  • Tone: energetic and engaging
  • Length: under 800 words
Here’s the current post: [text]”
The content is the same, but the structured version is dramatically easier for both you and the AI to work with. Pro tip: You don’t have to use labels like “BACKGROUND” and “TASK.” The key is putting related information together and separating different types of instructions clearly. Even simple line breaks and logical ordering help a lot.

​
The Goldilocks Zone: Not Too Little, Not Too Much

Is there such a thing as too much context? Actually, yes. If you write a 2,000-word preamble before asking a simple question, the AI might lose focus on what you actually need. Worse, it might start hallucinating connections that aren’t there. The goal is relevant context. Ask yourself: “Does this information help the AI give me a better answer?” If yes, include it. If not, leave it out. Too little context:
“Write an apology email.”
(Apologizing for what? To whom? How serious is this?) Too much context:
“Last Tuesday at 3:47 PM, I was in a meeting with Sarah, who works in accounting, and she has two kids and just got back from vacation in Florida, and during the meeting I accidentally interrupted her while she was presenting Q3 numbers, which reminded me of something my grandmother used to say about patience, and anyway I need to write an apology email…”
(Most of this isn’t relevant. The AI doesn’t need to know about Florida or your grandmother.) Just right:
“I accidentally interrupted a colleague during her presentation in a team meeting. I want to send a brief, sincere apology email. Keep it professional but warm, and not overly dramatic since it was a minor incident.”

​
Try It Yourself

Time to practice! Here are some exercises to build your context-giving muscles.

​
Exercise 1: The Context Upgrade

Take this vague prompt and rewrite it with proper context: Original: “Help me write a message to my team.” Before rewriting, ask yourself:
  • What kind of message? (announcement, update, request, celebration?)
  • What’s it about?
  • How should it sound?
  • How long should it be?
Write your upgraded version, then try both prompts and compare the results.

​
Exercise 2: The New Coworker Test

Think of a task you actually need help with right now. Write your prompt, then read it out loud as if you’re explaining it to someone who just started working with you today.
  • Did you use any jargon they wouldn’t know?
  • Did you assume knowledge they wouldn’t have?
  • Is there backstory they’d need?
Revise accordingly, then test it.

​
Exercise 3: The Four-Part Checklist

Pick any writing task (an email, a social media post, a summary) and before prompting, explicitly write out:
  1. Audience: Who is this for?
  2. Purpose: What should it accomplish?
  3. Tone: How should it sound?
  4. Constraints: Any rules or limits?
Then incorporate all four into your prompt and see how the output improves.

​
Common Pitfalls

​
Pitfall 1: The “You Know What I Mean” Assumption

What it looks like: “Make it better” or “Fix this” without explaining what “better” means to you. Why it fails: “Better” could mean shorter, longer, funnier, more formal, more detailed, simpler… the AI will guess, and it might guess wrong. The fix: Be specific. “Make it more concise” or “add more technical detail” or “adjust the tone to sound more confident.”

​
Pitfall 2: Assuming the AI Remembers

What it looks like: Referencing something from a previous conversation that happened in a different chat session. Why it fails: Each conversation typically starts fresh. The AI doesn’t remember that last week you told it you work in healthcare or that you prefer bullet points. The fix: Re-state important context at the start of each new conversation. Yes, it feels repetitive to you, but the AI genuinely doesn’t know.

​
Pitfall 3: Burying the Important Stuff

What it looks like: Writing three paragraphs of background before mentioning what you actually need, or putting critical constraints at the very end. Why it fails: The AI processes your prompt in order. Important details can get lost in the shuffle. The fix: Lead with the most important information. State your task clearly near the beginning, then provide supporting context.

​
Pitfall 4: Contradictory Context

What it looks like: “Make it short and comprehensive” or “Keep it casual but very professional.” Why it fails: The AI will try to satisfy both requirements and usually end up doing neither well. The fix: Think through what you really want. If there’s a tension (short but detailed), acknowledge it: “Keep it concise, but make sure to cover X, Y, and Z specifically.”

​
Level Up: The Context Challenge

Here’s a prompt that’s missing critical context:
“Write a paragraph about dogs.”
Your challenge: Rewrite this prompt three different times, adding context that would result in three completely different paragraphs. Version 1: Make it educational for children. Version 2: Make it persuasive for someone considering adopting a rescue dog. Version 3: Make it humorous for a casual blog audience. Test all three and notice how the same basic topic transforms based on context alone. That’s the power you’re learning to wield.

​
Key Takeaway

The more relevant context you provide, the less AI has to guess, and guessing is where things go wrong. Use the four types of context (audience, purpose, tone, constraints) as a checklist. Structure your context clearly so AI can follow it. And remember: the gap between what’s in your head and what’s in your prompt is where misunderstandings live. Close that gap, and watch your results transform.

​
What’s Next

You now know how to set the stage with context. But what if you could do more than describe what you want… what if you could show the AI exactly what success looks like? In Lesson 5: Show, Don’t Just Tell, you’ll learn the powerful technique of few-shot prompting, where giving the AI examples transforms vague instructions into crystal-clear guidance. Because sometimes, the best way to explain what you want is to demonstrate it.