Free lesson

Your First Conversation: Save 15 Minutes With AI Today

Your First Conversation: Save 15 Minutes With AI Today — ChatGPT From Scratch: How to Use AI Every Day. What you are actually talking to; Where to open an assistant; Your first real task, start to finish; Treat the first answer as a draft; Common mistakes before using the result. Practical assignment: One real request and two improvements.

Demo edition. The materials are undergoing editorial review.

Topics covered

  • What you are actually talking to
  • Where to open an assistant
  • Your first real task, start to finish
  • Treat the first answer as a draft
  • Common mistakes before using the result

What you are actually talking to

A language model produces an answer by continuing text from your request and the conversation context. That makes it useful for drafting, rewording, explaining, planning, comparing and generating ideas. A plausible sentence does not prove that its facts are correct: an answer can include an invented date, quotation or detail.

A useful working analogy is a well-read intern who drafts quickly but sometimes makes confident mistakes. Give it language tasks, then verify facts that matter with reliable sources. Lesson 5 covers checking in detail.

Fictional example: Nora, a 51-year-old dental receptionist, had regarded AI as a toy. When someone suggested testing it on a small work task, the buttons were not the difficulty: she simply did not know what to type.

Where to open an assistant

Free access is enough for this first exercise when it is available in your country and account. You can open ChatGPT in a browser or install its official app. Other well-known assistants include Claude, Gemini and Microsoft Copilot; local services may handle similar text tasks. Availability, features and limits change, so check the service’s current terms.

When installing ChatGPT, verify the publisher. OpenAI’s official guidance identifies OpenAI OpCo, LLC in Apple’s App Store and OpenAI in Google Play. This helps distinguish the official app from an unofficial wrapper that may show advertisements or immediately request payment. An account is optional in some supported regions, but signing in is needed to save chat history. As of 16 November 2026, OpenAI’s terms require users to be at least 13 or the local minimum age of consent; anyone under 18 needs permission from a parent or legal guardian.

Your first real task, start to finish

Start with one small, real task: a message, an explanation or a plan. Give the assistant three anchors: the action, the situation and the output format.

Help me [write / explain / make a plan / compare].
The situation: [2–3 sentences about the circumstances and audience].
What I need: [tone, length, list or table].
If important information is missing, ask me up to 3 questions first.

Questions are useful only when the answer genuinely depends on a missing detail; they add friction to a simple, unambiguous request.

Fictional example: Nora asks for a polite message under 80 words. A Thursday 10:00 check-up must move, and the clinic can offer Friday 9:30 or Monday 14:00. The assistant asks whose apology the message should convey. Nora chooses the clinic, checks the times, replaces one stiff word and only then uses the text. She adds the patient’s name outside the chat.

Treat the first answer as a draft

Within one chat, ChatGPT uses the preceding conversation as context, so you need not repeat the entire request. State the exact change: “Cut this to 50 words,” “Make the tone warmer,” “Give me two versions,” or “The time is wrong; it should be Friday. Fix it.” You can also ask, “Why did you choose this wording?” This turns the exchange into editing rather than accepting the first output as final.

Start a new chat when the subject changes. Copy, retry and message-editing actions are also commonly available, although button names and positions can change. A microphone supports dictation, while a “+” or paperclip can attach a file or photo when that feature is available. No magic wording is required: state the goal, known details and constraints.

Common mistakes before using the result

The first mistake is typing only “write a message.” Without the audience, reason, tone and length limit, the assistant has to guess. The second is demanding clarification for every tiny task; that creates unnecessary steps when the request is already clear. The third is copying dates, names, amounts or rules without checking them. The fourth is pasting unnecessary personal data when “the patient,” “the client” or “my landlord” is enough for a draft.

Before sending anything, read the output as if you wrote it. Confirm that its meaning is intact, every factual detail is correct and the tone suits the recipient. The assistant can reduce drafting work; the person using the result remains responsible for it.

Practical assignment

One real request and two improvements

Continue to learning

Course outline

  1. Your First Conversation: Save 15 Minutes With AI Today
  2. What to Ask: 10 Everyday Tasks and a Simple Request Formula
  3. Safety and Privacy: What to Set Up in 5 Minutes
  4. Photos, Files and Voice: AI That Sees and Hears
  5. When AI Gets It Wrong: A Confident Tone Isn't the Truth
  6. Memory, Instructions and Projects: Give ChatGPT Useful Context
  7. AI at Home: Money, Food, Trips and Paperwork
  8. AI at Work and in Study: The Rules and Being Honest
  9. Your 10-Task Kit and What Comes Next