Zero Shot, Hero Shot: The 7 Key Prompts To Get AI To Do What You Want It To Do
Welcome to The Art of Getting What You Want from AI- No Safe Word Required
May 13, 2025 • 10 min read
So, you want AI to deliver content that’s as hot as your best-selling offer and as on-brand as your favorite pair of Lulus. But every time you throw a prompt at ChatGPT or Claude, what comes back feels about as personal as a grocery store receipt.
You’re tech-savvy, yes. You’ve dabbled in advanced prompting. Maybe you’ve tried zero shot prompting, one shot prompting, and even a little chain of thought to spice things up. Yet, you’re still stuck editing bland AI outputs that scream, “I was written by a robot who’s never had a wild night.”
If you’re tired of second-guessing your value or your prompt skills, I've got you. I’m going to break down 7 essential prompting techniques here- zero shot, one shot, few shot, chain of thought, meta prompting- plus a playroom of advanced strategies (think tree of thought, thread of thought, and more) that’ll make your AI outputs feel like it’s been working your brand’s pole since day one.
Understanding these prompting techniques isn’t just for AI nerds. It’s for anyone who’s fed up with endless revisions, existential dread, and the deep, nagging fear that their brand will never stand out in a sea of soulless content. You could be sipping your coffee, watching your inbox fill with leads, and wondering how you suddenly became the person everyone asks for marketing advice. Or you could keep wrangling with AI that doesn’t get you, your clients, or your vibe. Your choice. Let’s get you that hero shot- without selling your soul (or your Saturday).
1. Zero shot Prompting: The One-Night Stand of AI Instructions
Zero shot prompting is the AI equivalent of handing over your number and expecting the universe to deliver. No backstory, no complicated instructions, just: “Classify this as positive, negative, or neutral.” You’re betting that the AI has seen enough to know what you want, no foreplay necessary. This is perfect for those tasks that are so basic even your mother-in-law couldn’t mess them up: translation, summarization, quick content moderation.
In other words, zero shot prompting is your ticket to efficiency, but it’s not the answer to every problem. If you find yourself frustrated because the output is about as exciting as a corporate HR memo, you’re not alone. This is where you start, not where you finish. Use it for quick hits, and save your more nuanced demands for the next round.
Example:
Write a subject line for an email announcing a summer sale.
Result:
“Sizzling Summer Savings- Don’t Miss Out!”
Why it works:
No context, no examples, just a straight-up demand. Fast, easy, but don’t expect fireworks.
The catch? If you’re looking for something with a little more flavor- like that signature brand sass or a topic that requires some actual expertise like market positioning- zero shot will leave you high and dry.
Why does this matter? Because sometimes, you just don’t have the time or energy to spell everything out. You want results, not a back-and-forth that rivals your last relationship. The key to zero shot prompting is clarity. Be direct. Cut the fluff. Don’t ask the AI to recite Shakespeare if you only want the CliffsNotes. If the task is simple, this technique works like a charm. But if you need something more... custom, you’re going to need to put in a little more effort (and maybe a few examples).
2. One Shot and Few Shot Prompting: The AI Booty Call- With Benefits
Sometimes, AI needs a little more than just your phone number. You want it to understand your love language, your favorite playlist, and how you like your eggs in the morning. Enter one shot and few shot prompting. This technique is all about showing, not telling. Toss in one or two juicy examples, and suddenly the AI knows exactly how you want those customer service replies- empathy first, solution second, and never sounding like a robot with an inferiority complex.
Example:
Write a playful out-of-office reply. Here’s one I like:
“Hey! I’m currently out of the office, probably drinking something with an umbrella in it. I’ll get back to you soon- unless I decide to go off-grid for good (kidding… sort of).”
Now, give me another one in this style.
Result:
“Hey! I’m off pretending I have a life outside of work. I’ll respond when the vacation gods release me. Until then, keep those emails coming in hot.”
Why it works:
You gave the AI a wink, a nudge, and a prime example. It gets the vibe and delivers something in your lane.
Why does this matter for someone who’s sick of endless tweaks and has better things to do with their time than play "not this, more like that" with Chat GPT all day long? Because your brand’s authentic tone is your lingerie- it’s what keeps people coming back for more. By feeding the AI your best work (or at least what you wish you’d written on your best day), you teach it how to please you. Two or three good examples will give the AI enough context to mimic your style, your format, and your subtle “nobody else like me” vibe.
But don’t get sloppy. Consistency is key. If your examples are all over the place, the AI will be too. Treat your prompt like a first date: show your best side, be clear about your standards, and don’t be afraid to give examples to ask for exactly what you want. This is how you get outputs that sound like you- without the need for endless editing or that creeping suspicion you’re just selling a slightly fancier version of what everyone else is offering.
3. Chain of Thought Prompting: Making AI Show Its Receipts
Ever wish your AI would stop guessing and actually walk you through its logic before it spits out an answer? Chain-of-thought prompting is here to give you that slow, satisfying build-up. This technique is all about telling the AI to work step by step- think of it as the difference between a one-night stand and a full-on, hours-long affair (minus the drama). Instead of the AI blurting out “9 marbles,” you get a detailed breakdown: started with 8, gave away 3, found 4 more. Now you know exactly how the magic happens.
Example:
We need to decide which social platform to focus on next quarter. Think this through step by step:
List the platforms we currently use.
Analyze which one has the highest engagement rates.
Consider where our target audience is most active.
Recommend where we should double down, explaining why.
Result:
The AI walks through each step, showing its math and logic so you can make sure it makes sense in a humanly logical way too- no magic, just receipts all the way.
Why should you care? Because AI isn't the God's gift to marketers it likes to think it is- and if you're smart, you want transparency, proof, and a little reassurance that you know what the hell your AI is actually doing. Whether you’re solving complex problems, explaining your pricing, or just trying to get AI to stop hallucinating, chain of thought prompting forces the bot to show its work (aka 'receipts'). You get fewer mistakes, more accurate results, and a clear path from A to Z.
The trick? Just add a line: “Think through this step by step.” Or, if you’re feeling wordy, “Let’s work this out one step at a time.” Then spell out the exact steps you want it to follow- with an example for each one. You’ll get answers that feel thoughtful, deliberate, and way more trustworthy. This is how you go from “AI wrote this and I’m kinda embarrassed” to “Wow, is there anything you can’t do?”- without needing a PhD in prompt engineering.
4. Meta Prompting: The Power of Telling AI Exactly How You Like It
If you’ve ever wished your AI tools came with a template, structure, or at least a little discipline, meta prompting is your new best friend. Think of it as giving the AI a checklist- Section 1, Section 2, Section 3- so you never get a random, rambling mess. Whether you need product analyses, email outlines, or just a little consistency in your life, meta prompting lays out the structure so the AI knows exactly what you want (and how you want it)- examples optional.
Example:
Write a product review using this structure:
Section 1: Hook (one sentence)
Section 2: Pros (three bullet points)
Section 3: Cons (two bullet points)
Section 4: Final verdict (one cheeky sentence)
Result:
You get a review that’s organized, on-brand, and looks like you spent way more time on it than you actually did.
Here’s why this matters: when you’re running multiple startups, wrangling kids, and trying to look effortlessly competent on Zoom, you don’t have time for AI that needs handholding. Meta prompting is about efficiency. It gives you control over the format, keeps things organized, and means you spend less time editing and more time doing whatever else you actually care about. If you want your outputs to always look tight, professional, and on-brand, this is the technique to master.
Be specific. Tell the AI which sections you want, what order they should come in, and don’t be shy about demanding a little structure. You’ll get outputs that are consistent, reliable, and- dare I say it- sexy in their predictability. No more rolling the dice and hoping for something usable. With meta prompting, you’re always in charge.
5. Self Consistency Prompting: Making AI Check Itself Before It Wrecks Itself
We’ve all been there: you ask AI for help, and it gives you an answer that feels…off. Maybe it’s technically correct, but you can’t shake the feeling that something’s missing. Self-consistency prompting is how you get the AI to double-check its work- like having it solve a math problem three different ways and then pick the answer that keeps coming up. Perfect for those moments when accuracy is non-negotiable and you need the AI to stop winging it.
Example:
I barbecued 12 sausages, but my dog ate half. I then gave away 2 to my son and 2 to my daughter. How many sausages did I have left for me? Generate three different reasoning paths to solve this problem, then select the most consistent answer.
Here, the AI will work through the problem multiple times using different reasoning paths:
Path 1:
Start with 12 sausages.
Dog eats half: 12 ÷ 2 = 6 sausages gone.
Left with 12 – 6 = 6 sausages.
Gave away 2 to son: 6 – 2 = 4 sausages left.
Gave away 2 to daughter: 4 – 2 = 2 sausages left.
So, 2 sausages for me.
Path 2:
Dog eats half of 12 = 6, so 6 sausages left after the canine carnage.
Then, total given away: 2 (son) + 2 (daughter) = 4 sausages.
6 – 4 = 2 sausages left.
Result: 2 sausages for me.
Path 3:
12 sausages.
Dog ate 6 (half). That leaves 6.
Gave 2 to son, 6 – 2 = 4.
Gave 2 to daughter, 4 – 2 = 2.
So I’m left staring at 2 lonely, delicious sausages.
Consistent answer across all approaches:
2 sausages were left for me.
Apparently, the dog and the kids know how to crash a barbecue.
More importantly though- this technique is especially useful for complex reasoning, math problems, or anything where you just can’t afford a mistake. By instructing the AI to generate multiple reasoning paths, you force it to slow down, check its assumptions, and cross-reference its answers. The result? Outputs that are not just plausible but actually correct. Now you can finally relax and trust that your content, analysis, or sales copy isn’t going to embarrass you at the worst possible moment.
Combine this with chain of thought prompting, and you’ve got a one-two punch that’s basically error-proof. This is how you move from “I hope this is right” to “Yes, I’m that good”- and you don’t even have to tell anyone your secret. Just smile, collect your compliments, and let the AI do the heavy lifting.
6. Thread of Thought Prompting: Keeping AI on a Leash (But the Fun Kind)
Ever feel like your AI’s logic is wandering off for a smoke break halfway through your prompt? Thread of thought prompting is how you keep the conversation tight and the reasoning on track. Unlike chain of thought (where every step gets spelled out), thread of thought prompting has the AI focus on a single line of reasoning and follow it through- no detours, no rambling, no getting lost in existential crises about “what is a brand, really?”
Here’s how this works: you instruct the AI to stick with one perspective or strategy from start to finish. Perfect for when you want a consistent argument, a focused breakdown, or just need the AI to stop spiraling into 12 tangents about alternate solutions.
Example:
Write a rant about why every brand needs a bold voice- stick to the argument that it’s the only way to avoid being totally forgettable online. No detours.
Result:
The AI goes full throttle on a single train of thought, so every line is another punch against boring brands. No ifs, elses or buts.
7. Tree of Thought Prompting: Because Sometimes You Want ALL the Options
Some days you want laser focus. Other days, you want to see every possible branch before you pick the ripest fruit. Tree of thought prompting is how you get your AI to explore multiple solutions, angles, or creative takes- then show its work so you can cherry-pick the best ones.
Here’s the play: ask your AI to approach a problem or prompt by branching out into several distinct possibilities. Each “branch” can be a different strategy, perspective, headline, or offer angle. The AI lays them out, walks through each one, and (if you’re feeling spicy) even recommends which branch to follow based on your goals.
Example:
Suggest three angles for a launch announcement based on:
FOMO
Transformation
Humor
Outline the hook for each, then pick the one that’s most likely to get clicks for a high-energy audience.
Result:
AI branches out, lays the options on your metaphorical silver platter, then helps you pick the juiciest one.
This is a power move for campaign ideation, brand positioning, or anytime you want more than one boring answer. You could be reviewing three killer hooks for your next launch, analyzing different pricing models, or just making sure your content never, ever looks like it came from the “corporate beige” department.
Don’t settle for a single, predictable output. Tell your AI to grow a damn tree- then climb whichever branch makes your brand look the hottest.
Let's Wrap It Up with a Bow
Prompting techniques aren’t just for show- they’re the difference between AI outputs and workflows that feel like bad Tinder dates and ones that get you noticed in your niche (in the best way). The secret? Know what you want, be explicit, and don’t settle for less. Whether you’re using zero shot prompting for quick tasks, one shot prompting for nuanced responses, or chain of thought for complex reasoning, the trick is to always lead with clarity. Tell the AI exactly what you want, show it how you like it, and don’t be afraid to demand a little structure.
For the power users reading this, advanced prompting is your playground. Thread of thought, Tree of thought, and ReAct (not covered here) aren’t just fancy buzzwords- they’re how you get AI to do the kind of work that actually moves your business forward. You could be spending less time editing and second-guessing, and more time scaling your business, closing high-value clients, and maybe even sneaking in a guilt-free nap.
So, next time you sit down to prompt your AI, remember: you’re in control. You know how to prompt effectively. And you don’t have to settle for bland, soulless AI outputs ever again.
Note: This is by no means an exhaustive list of all of the prompting techniques out there- or even of all of the best ones. These are simply the prompting techniques that I have personally used and built into heyLevi™ since 2022- and the ones that I have gotten the best results with using AI for product marketing.
-Ruheene Masand Jaura
Book Ruheene for your next event.
Uncover insider tips and strategies for using AI as a brand creator, content generator and most importantly- a brand equity builder.
Ruheene Masand Jaura © 2025 All Rights Reserved