AI PROJECT STARTER — LESSON 01
English transcript

Introduction

Welcome to AI Project Starter.

I’m your virtual AI instructor.

Let’s turn your business idea into a clear first step.

Open your worksheet, and let’s get started.

In this lesson, you’ll use an AI assistant to turn that idea into a practical research plan. By the end, you’ll have a short project brief, a list of assumptions, and one next step you can take outside the chat.

We’ll use a fictional video-editing service as our example. You can follow along with your own idea. You don’t need to know how to code.

Start with a vague request

Let’s start with a request you might naturally type: “I want to start a business with AI. Give me a plan.”

Read the answer you receive. Does it know your skills, your available time, or who you want to help? Without those details, it may give generic suggestions, make assumptions, or ask you questions.

A polished answer can still be a poor fit. Our first job is to give the assistant enough context to produce something we can evaluate.

Add context and constraints

We’ll add four things: context, a goal, constraints, and the output we want.

For this example, imagine you can edit basic videos. You have five hours a week and a monthly tool budget of fifty dollars. You’re considering editing short videos from footage that local businesses already own.

These are example details, not facts about you. Replace them with your situation.

Our goal today is to decide what to research before building an offer. We’ll ask for a short brief, assumptions, and questions for potential customers. We’ll also tell the assistant not to invent demand, prices, or testimonials.

Pause here and read the prompt. Notice how specific the requested output is. That makes the answer easier to review.

Check the assumptions

Now we need to inspect the answer. Instructions help, but they don’t guarantee accuracy.

Our answer identifies three assumptions: owners have footage they struggle to edit, some may pay for help, and we can deliver within our limits. Notice that it labels them as assumptions. It has not verified any of them.

Now look at the review table from our second prompt. The editing skills, five-hour schedule, and fifty-dollar budget came from our fictional example. Those are supplied details. Whether an owner would pay is still unknown.

For delivery time, the answer suggests a useful test: complete an authorized sample and measure the entire process, including file transfer, instructions, editing, exporting, and a revision.

The table also questions some research advice. Treat that as a review aid, not an independent fact-check. Past spending provides context, but it does not prove someone will buy our service. Two conversations can offer initial evidence, but cannot tell us how widespread demand is.

Review the answer and take a research step

The next step is to learn from people who might use the service.

The first question asks what happened the last time an owner recorded footage. That gives us a concrete event to discuss.

But question four asks when footage most recently went unpublished. That assumes it happened. Let’s improve it: “Have you recorded any footage that you decided not to publish? If so, tell me about the most recent example and what led to that decision.”

This is an important part of using AI: you can improve an answer even when it looks well structured.

The suggested action is two twenty-minute conversations. Its time allowances add up to three hours and ten minutes, but finding participants may take longer. Treat those numbers as a planning estimate.

For your first step, try arranging one conversation. Record what the person actually says and what remains unknown. If you cannot arrange a conversation yet, review one real competing offer and record its source and date. That can reveal how an offer is presented, but not whether customers buy it.

Your exercise

Now it’s your turn. Write down your idea, who it might help, your relevant skills, and your constraints. Run the first prompt, then review the answer using the second prompt.

Save a one-page brief with the problem you want to investigate, three assumptions, and one next action. If you’re unsure about something, write “unknown.” That is a useful result.

You’ve completed this lesson when you can explain what you’re considering, what you still don’t know, and how you will learn more.

Keep your brief. In the next project, you’ll use it to create sample content while making sure every claim matches what you actually know.