gluntyacademy The one skill that changes everything: how you ask. Twelve recipes you can reuse forever. Tap a template to copy it, then fill in the blanks.
A recipe is a shape for a good request. The blanks in [brackets] are yours to fill. Tap any template to copy it.
12 reusable recipes, all yours to keep.
When the first answer feels generic and could have been written for anyone.
Help me with [task]. It is for [who it is for], the goal is [goal], and it has to fit [limit, like 100 words or 30 seconds]. Why it works The AI fills any gap you leave with its most average guess, so every real detail you add pulls the output toward the answer you actually wanted.
When you want real options instead of one take-it-or-leave-it answer.
Give me three different versions of [thing]: one [style A], one [style B], and one [style C]. Then tell me which fits [my goal] best and why. Why it works Asking for a range makes the AI explore instead of locking onto its single most likely guess, so you see ideas you would not have known to ask for.
When you can point to what good looks like more easily than you can describe it.
Here is an example I like: [paste the example]. Make a new [thing] about [topic] that matches its style and structure. Why it works A concrete example beats a pile of adjectives because the AI copies the exact pattern in front of it instead of guessing what you mean by good.
When a draft looks fine but you suspect it could be sharper.
Here is your answer: [paste it back]. List the three biggest weaknesses in it, then rewrite it fixing all three. Why it works Spotting flaws and fixing them are separate jobs, so forcing the AI to name problems first gives it something specific to solve instead of repeating itself.
When you want the answer to come from a certain viewpoint or level of expertise.
Act as a [role, like a patient science teacher or a tough editor]. [Now ask your question or give your task]. Why it works A role loads a whole way of talking and thinking, so one word quietly steers thousands of small choices about tone, detail, and what counts as a good answer.
When the answer must fit a format, a length, or a list of must-haves.
Do [task]. Rules: [rule 1], [rule 2], [rule 3]. If you cannot follow a rule, say so instead of quietly ignoring it. Why it works Clear limits stated up front are easier to hit than fixes made afterward, because you are shaping the output while it is still being written.
When you get an answer but do not understand how it got there.
Explain your reasoning for [the answer or choice] step by step, as if you were teaching me. Point out anywhere you were unsure. Why it works Asking for the steps makes the AI show its work instead of leaping to a guess, and thinking out loud tends to make the answer itself more correct.
When something is close but not right and you want to fix it without breaking it.
Keep everything exactly the same except [one thing]: change that to [what you want]. Do not touch anything else. Why it works A single edit is easy to check and easy to undo, while asking for ten changes at once makes it impossible to tell which one caused a new problem.
When you care how something feels, not just what it says.
Make [thing] feel [three vibe words, like calm, playful, and honest]. Avoid [what you do not want, like sounding like an ad]. Why it works Feeling words plus a short do-not list shrink the huge space of possible tones down to the exact corner you were aiming for.
When you have a plan or a draft and want to catch blind spots before you commit.
Here is my [plan or draft]: [paste it]. What is missing, what did I get wrong, and what would a smart critic point out? Why it works You naturally ask about things you already thought of, so telling the AI to hunt for gaps surfaces the parts you did not know to ask about.
When an explanation is technically correct but you still do not get it.
Explain [topic] like I am [age or level], using an everyday comparison. Then give me one small example I could try myself. Why it works Naming your level tells the AI which words and ideas you already have, so it builds the explanation out of things you know instead of things you do not.
When the answer is wrong, lazy, or misses the point you were making.
That is not quite what I meant. The problem is [what is off]. Try again, and this time [what you actually want]. Why it works The AI cannot read your mind, but it can treat your correction as fresh input, so naming exactly what missed steers the next try much closer to right.
Now go use one on something real.