AI School › Talking to AI · Lesson 37 of 130

🔁 Correcting and trying again

If you don't like the answer, you don't start over: you fix it by saying what to change.

When you draw, the first line is almost never perfect: you erase a bit, add, improve. With AI it's the same, and it's a huge advantage. You don't have to accept the first answer as if it were carved in stone. You can say "too long, make it shorter", "use easier words", "add a funny example", "that's not what I meant, let me explain better". Each of your comments is like guiding a sculptor: blow by blow, the statue becomes the one you had in mind. The conversation is teamwork.

Let's take an example. You ask for a summary and the AI gives you ten lines, but you wanted three without big words. Reply: "The meaning is perfect, but shorten it to three lines and remove the hard words". In seconds you have what you need. The secret is not to get angry or start over every time: just explain what's wrong. Trying again isn't failing, it's fine-tuning. Like tuning a guitar: you turn the peg a little at a time until the note is right.

🎒 Try it in class (or at home)

Ask the AI something, then improve the answer with three corrections in a row (shorter, simpler, with an example) and watch it grow.

← Lesson 36: Explaining the situation (the context) · Lesson 38: Asking step by step →

From our network

AGORÀ Intelligence: Enterprise AI Governance Platform

Govern AI at scale: policies, adoption and measurable results on your data. Built for boards and C-suite.

Visit agora-intelligence.com →

From the Agora Intelligence blog

More on agora-intelligence.com →

📱 Download the Android app (beta) iOS coming soon

Say what you mean. Get what you need.

Grace Certified, the AI coach that trains and certifies your prompt engineering, by Agora Intelligence.