Interactive eBook BuilderInteractive eBook Builder
$59 · unlimited access, unlimited eBooksGet it for $59

· By · 13 min read

Prompt Engineering Examples for Beginners

promptsaiwritingebook

Prompt engineering examples are short, copy-ready instructions that show what to include—the reader, the job, the source material, the format and the limits—so you can see why a vague request fails and a specific one works. OpenAI’s prompt engineering guide describes the work as writing effective instructions so a model consistently generates content that meets your requirements. That is the standard these examples are built around: a result you can check, not a paragraph that merely sounds finished.

For eBook and workbook projects, Interactive eBook Builder is a $59 one-time Notion AI template with Spec Builder, Drafter, Polisher and SEO BOT already set up, so you can practice these techniques inside a working book workflow instead of starting from a blank chat. Notion AI access and usage are separate from the purchase. If you want the underlying idea before the examples, see our plain-English guide to prompt engineering. The examples below are teaching prompts written for this article. They are not paid Builder prompts, and they are not recorded model outputs.

What should a beginner prompt include?

A useful beginner prompt makes the decisions you would otherwise leave to guesswork. Anthropic’s prompting best practices recommend being clear and direct, and treating the model like a capable new colleague who does not already know your norms. Google’s prompt design strategies make the same point from the other direction: give clear, specific instructions, and include the context the model needs instead of assuming it already has it.

These five parts cover most early writing tasks:

PartWhat you decideWhat happens if you skip it
ReaderWho will use the result, and in what situationThe tone and level drift toward a generic audience
JobWhat the reader should be able to do afterwardYou get a topic summary instead of a usable resource
NotesThe facts, examples and limits the model may useThe model fills gaps with plausible but uncheckable detail
FormatHeadings, length, lists, exercises or a tableYou spend the next turn asking for a rewrite of the shape
LimitsWhat to leave out or leave unchangedExtra tips, tools or claims appear that were never in your notes

You do not need all five parts for every request. A one-line cleanup can be short. A chapter draft, a workbook exercise or a packing checklist usually needs all of them, because each missing decision is a place the model can invent a default.

Example 1: turn notes into a chapter outline

A vague version is:

Make an outline for my productivity book chapter on morning routines.

That request names a topic and nothing else. The model has to invent the reader, the chapter’s job, the length and whether this is a pep talk, a research summary or a how-to. A more useful version supplies those decisions and keeps the model inside your notes:

Create a chapter outline for a short eBook for new freelancers who want a
weekday morning routine that survives an uneven workload.

Use only these notes:
- The reader has 45 minutes before the first client call, three days a week.
- They already have a calendar and a simple task list; they do not want a new app.
- The chapter should help them choose three repeatable actions, not a rigid 5 a.m. schedule.
- Do not include supplement or biohacking advice.

Return:
1. A working chapter title
2. Five heading-level sections
3. One bullet under each heading describing what that section must accomplish
4. One workbook question the reader answers before the chapter ends

Keep the language plain. If a point is not in the notes, leave it out and add a
line labeled "Need from me" instead of inventing it.

A possible target layout—not a recorded model response—is a title such as “A 45-minute weekday start you can keep,” followed by sections on the constraint, the three actions, what to skip on heavy days, a short example morning, and a closing question such as “Which of the three actions will you protect first when a client call moves earlier?”

Now the review has substance. Did every heading serve the 45-minute constraint? Did a 5 a.m. miracle routine sneak in? Did the outline invent an app? Those are observable misses, which is the point of a beginner example: you can see which instruction failed.

Example 2: draft a short teaching section

Outlines are easier than prose. A second example shows how the same parts work when you want a draft, not a plan. A vague version is:

Write the section on choosing three morning actions. Make it engaging.

“Engaging” is not a test. A more useful version sets the job, the source and the shape of the section:

Draft the section "Choose three actions, not a full lifestyle change" for new
freelancers. Use the approved outline and notes below, and do not add new
habits, tools or research.

Reader: a freelancer with 45 minutes before the first client call, three days
a week.
Job: help them pick three repeatable actions they can finish even on a heavy day.
Length: 250 to 350 words, then one short example of a completed trio.
Tone: calm and practical. No hype, no "crush your morning," no medical claims.

Notes:
- Candidate actions from the author's own list: review today's calendar, write
  the first task in one sentence, and prepare the workspace for the first call.
- The reader may swap an item, but the trio should stay under 45 minutes total.
- The example should use the three actions above, not a different routine.

Format:
- Open with the constraint, not a story about successful founders.
- End with the example trio written as three numbered steps, each with a
  time estimate that adds up to 45 minutes or less.

The comparison is easier if you keep the vague prompt’s output and this one side by side. Look for added habits, invented time claims, and openings that sell a lifestyle instead of teaching the choice. If the draft runs long, shorten the length instruction before you rewrite the whole tone. Change one variable at a time so you can tell what helped. Our guide to improving a prompt when the answer is generic walks through that revision loop in more detail.

Example 3: write a workbook exercise

Workbook pages fail in a different way from chapters. The prose can sound helpful while the activity is impossible to complete. A vague version is:

Add a worksheet about morning routines.

A more useful version defines the exercise the way you would brief an editor:

Create one workbook exercise that follows the chapter section on choosing three
morning actions.

Reader: a new freelancer filling this out on paper or in a PDF, in about
five minutes.
Job: leave the page with three chosen actions and one backup for heavy days.

Rules:
- Use only the chapter's three candidate actions, plus a blank line if the
  reader wants to substitute one action of their own.
- Each chosen action needs a time estimate in minutes. The three estimates
  must have a space to total them.
- Include one prompt: "If a client call moves earlier, which action will you
  drop first, and what will you do instead?"
- Do not add scoring, personality labels or advice that is not in the chapter.

Format:
1. A one-sentence instruction at the top
2. Three labeled lines for the actions and times
3. A total line
4. The backup question, with two short writing lines

A possible target—again a teaching layout, not a captured reply—is a page the reader can complete without rereading the chapter twice. If the model turns the exercise into a quiz, or adds a scoring system you never asked for, the miss is in the format and limits, not in “making it nicer.”

This is also a good place to notice a beginner habit: asking for a “fun worksheet” and hoping the model will invent a useful activity. Fun is not a completion test. “The reader can fill this in five minutes and walk away with three timed actions” is.

Example 4: extract a checklist from messy notes

Not every prompt has to produce chapters. Beginners often need a clean list from a pile of notes. A vague version is:

Turn these notes into a checklist for my readers.

A more useful version tells the model what the list is for and what it may not add:

Create a packing-and-setup checklist for a first-time craft-market seller.

Use only these notes:
- The venue supplies one table. There is no electricity at the stall.
- Bring a table cover, labeled stock boxes, price cards, an inventory sheet,
  a cash box and a phone battery pack.
- Do not include items the venue already supplies, and do not add equipment
  that needs a wall outlet.

Return six checkbox items, one for each item to bring, written as short
imperative lines. After the list, add one reminder sentence about the
supplied table and the lack of power. Use plain words. If the notes are
missing something essential, add a single "Need from me" line instead of
inventing an item.

The test is strict on purpose. If the checklist tells the seller to bring a table, or adds a card reader that needs a plug, the prompt’s limits were ignored. If it invents “business cards” you never mentioned, the source boundary failed. Those failures teach more than a polished but uncheckable list.

What these examples have in common

Each improved prompt does four things the vague one skipped:

  1. It names a specific reader and situation.
  2. It points at notes the model is allowed to use.
  3. It describes a finished shape you can inspect—an outline, a section, a page, a list.
  4. It says what not to invent.

That last part matters, but it works best when you also say what to do. Anthropic’s guidance notes that telling the model what to do is often more effective than only listing what to avoid. In the examples above, “return five headings with one bullet each” carries more weight than “don’t be messy,” and “use only these notes” is stronger when the notes are actually present in the prompt.

OpenAI’s guide also organizes longer instructions into identity, instructions, examples and context, and it recommends using Markdown or XML-style labels so those sections stay distinct. You can see a light version of that structure here: a short job description, then a notes block, then a format block. When you are ready to formalize that structure, our guides to writing a prompt template and using delimiters in prompts take the next step. For a different beginner path—building one small project and revising it—see learning prompt engineering by building.

How to adapt an example without copying it blindly

Do not paste an example into a new niche and assume it will hold. Keep the structure and replace the decisions:

If you write more than one chapter, you will get further by turning a working example into a reusable template than by collecting dozens of one-off prompts. The Builder’s Spec Builder, Drafter and Polisher stages are built around that idea: planning, drafting and editorial refinement get separate instruction sets, and SEO BOT can later turn related reader questions into supporting articles on the Auto-Blogger blog, each linking back to your eBook, site and brand. You can keep the supplied setup or adapt the audience, voice and activities. You still review the work before it goes out.

One more caution, also from OpenAI: model responses are not deterministic. A single good result does not prove a prompt will always behave the same way. Try a second set of notes before you reuse it on a whole book.

FAQ

Are these prompts meant for ChatGPT, Claude or Gemini?

The structure works in any of those chat tools. Vendors phrase some advice differently, and model behavior changes, so treat the wording as a starting point and check the result against your notes. For differences that show up in vendor documentation, see our comparison of reusing the same prompt across ChatGPT, Claude and Gemini.

How long should a beginner prompt be?

Long enough to include the decisions the task requires, and no longer. A checklist may need a short notes block and a format. A chapter draft usually needs the reader, the notes, the length and the limits. Adding color that does not change a decision rarely helps. Our guide to how long a prompt should be covers this in more depth.

Should I add “act as an expert” to these examples?

A short role can help with tone and audience, but it does not make the model an expert on your subject and it will not replace your notes. If you use a role, keep it specific—“You are helping a new freelancer complete a 45-minute planning page”—and still supply the facts. Our article on whether you should tell AI to act as an expert explains the tradeoff.

Can I use one example prompt for my whole eBook?

You can reuse the structure, not the filled-in notes. Update the reader, the chapter job, the source material and the output shape, then test the new version. A book-length project also benefits from splitting planning, drafting and editing instead of asking for the entire manuscript in one request.

Before you hit send

Ship niche eBooks with agents already built.

Get the Interactive eBook Builder for $59