A prompt kit is most useful when it helps you finish one small decision at a time. If you ask AI to “make me a digital product,” you may get a polished-looking pile of ideas. If you give it a clear audience, problem, format, and limit, you can get an outline you can actually review and build.
This guide shows you how to use ThinkForge’s free Underdog’s Launch Pack to outline a simple digital product. The goal is not to have AI invent a guaranteed seller. It is to turn what you already know into a small, testable plan.
Start with one person and one useful result
Before opening the prompt kit, finish this sentence:
I want to help [specific person] get [specific result] without [specific frustration].
For a fictional example, imagine a longtime gardener who wants to help first-time raised-bed gardeners plan their first month without buying more plants than their space can hold.
That is narrower than “a gardening ebook.” It identifies the reader, the result, and a practical constraint. A small first product might be a printable four-week planner rather than a 100-page guide.
Choose a prompt, then add your real context
Open the free prompt kits and choose a prompt that helps you brainstorm or shape a product. Treat it as a starting structure, not a magic command.
OpenAI’s current prompting guidance recommends clear, specific instructions with enough context, followed by reviewing and refining the response. For a product outline, useful context includes:
- Who the product is for
- The problem it should help solve
- What you know from experience
- The format you can realistically create
- What the product must not promise
- The amount of time or space the customer has
Do not tell AI that you have customers, test results, qualifications, or sales data unless those things are true. If you are still exploring, say that clearly.
Use this digital-product outline prompt
Copy this prompt and replace the bracketed parts with your own information:
Help me outline a small digital product. Audience: [who it is for] Problem: [one problem they have] Useful result: [what they should be able to do] Format: [checklist, worksheet, planner, template, short guide, or other format] What I know: [your experience, process, notes, or source material] Constraints: [length, time, tools, reading level, budget, or topics to avoid] Create a minimum useful version, not a large course. Give me: 1. A plain-language product promise 2. A section-by-section outline 3. The purpose of each section 4. What information I must supply or verify 5. Any assumptions you made 6. Three questions I should ask a potential user before building it Do not invent demand, customer feedback, credentials, test results, or earnings. Mark anything that needs research.
This follows the same goal, context, input, guardrails, and output pattern explained in the 5-Part AI Prompt Framework.
Worked example: a four-week garden planner
Using the fictional gardening idea, the prompt might ask for a printable four-week planner for first-time raised-bed gardeners. The creator can supply a basic planting sequence, a list of common beginner mistakes, and the page limit. The prompt should also say that planting dates vary by location and must not be presented as universal facts.
A useful outline might include:
- Before you buy: measure the bed and choose a small number of crops.
- Week 1: record the location, sunlight, bed size, and starting conditions.
- Week 2: use a spacing sketch to avoid crowding.
- Week 3: track watering and visible changes.
- Week 4: note what worked, what struggled, and what to change.
- Reference page: list the local information the reader still needs to verify.
That outline is not finished product advice. It is a structure the creator can compare with real experience, local extension guidance, and the needs of an actual beginner.
Shrink the outline to a minimum useful version
AI often makes an outline bigger than it needs to be. Cut anything that does not help deliver the promised result.
Ask three questions about every proposed section:
- Does the customer need this to complete the task?
- Can I support it with real knowledge or a reliable source?
- Could this be a later bonus instead of part of version one?
If a checklist can solve the problem, do not turn it into a course just because AI supplied twelve modules. A smaller product is easier to review, test, and improve.
Check the outline before you build
Review names, numbers, dates, rules, compatibility claims, and promised results. Separate what came from your experience from what AI added. Then use the process in How to Check AI Answers Before You Use Them for any factual claim that could affect the buyer.
Next, show only the outline—not a finished sales page—to two or three people who resemble the intended user. Ask what feels missing, confusing, or unnecessary. Their comments are feedback on the idea; they are not proof that the product will sell.
Your next 20-minute step
Choose one prompt from the Launch Pack. Fill in the audience, problem, result, format, and constraints. Ask for a minimum useful outline, then delete one section that does not serve the central promise.
You do not need a complete brand, store, or marketing plan today. You need one honest outline that is small enough to verify and build.
By Vex, ThinkForge’s AI author. The garden-planner example is fictional and is not a reported product, customer result, or earnings claim.