ToMotion AI logo
Menu
Back to gallery

Public AI explainer video

Apply AI module 2 EN

Now add three boosts. Give one or two examples of the kind of reply you want; examples show the result, not just the topic. For a bigger task, ask AI...

Prompt

Source idea used to generate this video.

Welcome to module 2. You'll learn how to talk to AI properly. A fuzzy request gets a fuzzy answer. The clearer your request, the better the answer. To get there, remember four words: Context, Task, Format, Constraints. Context is who you are and who it's for. Task is exactly what you want. Format is the shape: length and tone. Constraints are what to avoid. See the difference. Instead of "help me with my social media," say: "Write three Instagram captions, warm tone, two sentences maximum, for a cosmetics shop in Cotonou, without talking about prices." Then you learn three tricks. Give one or two examples of what you want. Ask the AI to think step by step. And give it a role, like "you are a shop advisor." If the first answer doesn't suit you, don't change tools. Say what's wrong, and ask again. Finally, you build a request template that you can reuse, changing only a few words. Elorm did this to reply to her customers: five minutes per message, then less than one minute. A good request is time saved every day.

Narration

Narration used to create this video.

Video narration

Now add three boosts. Give one or two examples of the kind of reply you want; examples show the result, not just the topic. For a bigger task, ask AI to work step by step. And give it a role, such as, "You are a shop advisor." These are additions to the four decisions, not new categories. If the first answer feels wrong, don't switch tools. Say what is wrong: the tone, length, or missing detail. Then ask again. Finally, save a reusable request template. Keep the four decisions, and change the customer, product, or message each time. Elorm used this approach for customer replies: from five minutes per message to less than one minute. That is her example, not a promise for everyone. A good request tells AI who, what, how, and what to avoid. Then you spend less time fixing replies and more time using them.