Task Types: Your Prompt Toolkit

L2 Lesson 4 of 5 — Intentional Prompting

1

2

3

4

5

Six Tasks, Six Prompts

You now know the anatomy of a good Prompt The input you send to an AI model — your question, instruction, or task. (Role, Context, Task, Format, Constraints) and how to iterate and refine. In this lesson, you’ll see how all of that plays out in practice — across six typical tasks from everyday work.

Each example uses the five building blocks from Lesson 01. Pay attention to how each prompt deploys Role, Context, and Format deliberately.

1. Writing

Creating text — from internal emails to customer-facing communications.

You are the head of internal communications at a mid-size company. Write an internal announcement email: Starting June 1, the company is adopting a hybrid work model (3 days in-office, 2 days remote). Tone: positive but professional. Mention HR as the point of contact and reference an upcoming FAQ document. 200 words max.

Why this works: The role (head of internal communications) sets the tone. Context (hybrid model, date, split) eliminates guesswork. Format constraints (200 words, email) prevent a rambling essay.

2. Analyzing

Evaluating data, identifying patterns, deriving actionable recommendations.

Here is our customer churn data for the last 6 months: [paste data]. Identify the 3 most significant factors that correlate with churn. For each factor: evidence, confidence level (high/medium/low), and one concrete countermeasure. Present the results as a table.

Why this works: The task is tightly scoped (3 factors, not “everything”). The table format enforces structure. The confidence rating forces the AI to assess its own certainty — which makes weak spots visible.

3. Summarizing

Reducing long texts to what matters — for people who are short on time.

Summarize the following article (approximately 2,000 words) in 3 bullet points. Each bullet point: one sentence max. Focus on: core argument, strongest evidence, and practical relevance for project managers. Written for someone with 30 seconds to spare.

Why this works: “3 bullet points, one sentence each” prevents a half-page summary. The audience hint (project managers, 30 seconds) steers what’s relevant and what gets cut.

4. Translating

Not word-for-word, but meaning-for-meaning — with sensitivity to audience and tone.

Translate the following product description from English to French. Maintain the marketing tone — natural, idiomatic French, not a literal translation. Technical terms (API, SaaS, Dashboard) stay in English. Target audience: IT decision-makers at enterprise companies in France and French-speaking Europe.

Why this works: “Natural, idiomatic French, not a literal translation” is the critical constraint. Without it, AI often produces stiff, mechanical translations. The audience hint steers the level of formality and word choice.

5. Brainstorming

Generating ideas — with structure, so the list stays usable.

I need to increase engagement at our quarterly all-hands meetings. Current situation: low attendance, negative feedback (“too long, too boring”). Generate 10 creative ideas. Mix of practical/low-cost and ambitious. For each idea: one sentence on implementation effort (low/medium/high).

Why this works: Context (low engagement, specific feedback) prevents generic suggestions. The mix of pragmatic and ambitious gives the AI room to explore. The effort rating makes the list immediately prioritizable.

6. Formulas and Code (for Non-Developers)

AI can solve technical tasks — and explain them to you so you can adapt them yourself.

Write an Excel formula for the weighted average of cells B2:B10 using the weights in C2:C10. Then explain in plain language what each part of the formula does, so I can adapt it for other ranges.

Why this works: The task alone would be enough — but the explanation makes the difference. Instead of blindly copying a formula, you understand what it does. That’s the difference between dependency and building competence.

Try It

1. Pick one of the six task types that fits your next real work project.
2. Use the example as a template and adapt it to your situation — your own role, your own context, your own format.
3. Compare the result with a prompt that has none of this structure. What do you notice?

Think Further

All six examples follow the same pattern: Role, Context, Task, Format, Constraints. The pattern matters more than the individual templates. Once you’ve internalized it, you can apply it to any new task.

In the next lesson, you’ll see this in direct comparison: Good vs. Bad Prompts shows you side by side what makes the difference.