You're using AI but it's not working like you hoped it would. Something small is off. It's costing you more than you think. This simple scorecard shows you exactly what it is and how to fix it.
If you're not getting what you want from AI, you're not alone.
Few people do.
The reason why is because most people use AI wrong.
They treat it like a vending machine with psychic powers.
They think they can just press a button and AI will give them exactly what they want.
That's not going to work.
AI isn't a vending machine (much less a psychic one.)
It doesn't know what you want unless you tell it.
Telling it what you want is called "directing."
If you get good at directing AI, you're golden.
This will show you how well you're currently directing your AI.
Plus you'll see exactly why AI isn't giving you what you want …and how to fix it.
Each question is scored 1 to 5. Answer honestly based on what you actually do right now — not what you think you should do.
—
Add up your scores. The maximum is 50. Here's what each bracket reveals.
Your approach is solid. You're probably already getting above-average output from your AI and you know it. The next move isn't to change what you're doing. It's to systematize it. Take the directions that are producing your best work and turn them into templates you can reuse. (Or even better, turn them into "Powers" inside of oJoy.)
Then look hard at Questions 5 and 7. Objection handling and real evidence are where even strong directors tend to leave conversion on the table.
You've got the basics down but you're missing one or two things consistently and those gaps are showing up in your results. Your copy is probably decent. Just not as reliable as it could be.
Check your lowest-scoring questions. If Question 3 (desired outcome) or Question 8 (CTA) are in there, fix those first. They're the highest-ROI changes you can make right now.
You're using AI to produce output faster. That's not the same as producing better output. Right now you might actually be getting results that are worse than what you'd write yourself, just quicker.
Start with Questions 10 and 2. Getting the framework and format right will do more for your output quality than anything else at this stage.
Your AI doesn't have enough to work with. When you don't give it context, it fills the gap with generic patterns and generic patterns don't convert.
Start with Question 1 and Question 10. Who you're writing for and what structure you're using. Every single direction should start there, every time, before anything else.
Your score tells you how well you're directing. But the questions you scored lowest on point to a specific pattern. Most people's patterns fall into one of these five.
You're writing for "everyone" which means you're writing for no one. Your AI has no idea who it's talking to, so it talks to the average of everyone it's ever been trained on. The fix is simple but it requires discipline: before you write a single direction, write four sentences about the specific person reading it. Who they are, what they want, what's stopping them, and what they're comparing you to.
Your copy informs but doesn't convert. There's no clear ask, or the ask is so vague it creates no momentum. "Get them interested" is not a conversion goal. Decide exactly what you want the reader to do, why they'd do it right now, and what's making them hesitate. Then put all of that in your direction.
Your AI is making things up because you didn't give it anything real to work with. It's not lying on purpose. It's filling in gaps based on what it thinks should be there. Gap-filling produces copy that sounds real but isn't. Keep a running document of your actual proof points: real numbers, real results, real customer language. Drop it into your directions and watch the output change.
The pieces are there but they're in the wrong order. Copy without structure is just words. AIDA for mid-funnel. PAS for cold traffic. Story-Bridge-Offer for email. Pick a framework before you write a single direction and tell your AI which one to use. That one change will do more for your output than almost anything else.
You're accepting output instead of diagnosing it. The first draft isn't supposed to be perfect. It's more like a warm up. If you think it could be better, tell your AI how you think it could be better and get it to make it better. Ask it for stronger headlines and better bullet points. Ask it to read through it and find where readers might have objections or become skeptical, then ask it to fix those points. Stuff like that only takes a few minutes, but those few minutes could be the highest paying few minutes of your life.
Three side-by-side comparisons. Same task, same AI. The difference is how it was directed.
"Write a subject line for a marketing email about oJoy."
"Write 5 email subject line options targeting established business owners who are already using AI tools but not getting results they can use. Their biggest frustration is that the copy sounds generic and takes forever to fix. Write one using a curiosity gap, one using a counterintuitive statement, one using a social proof formula, and two using whatever approach fits best. No hype words."
"Write a headline for oJoy's landing page."
"Write 5 landing page headlines for oJoy. Audience: business owners who've tried ChatGPT and other AI tools but the output sounds like everyone else's and doesn't convert. Key differentiator: oJoy is trained on 25 years of real direct response — sales letters, emails, ads, webinars, infomercials. It doesn't just write copy, it analyzes your business, finds your highest-impact opportunity, and executes it. No prompts required. No templates. Use transformation frame and specific differentiator formulas. No buzzwords."
"Write a Facebook ad for oJoy."
"Write a Facebook ad for oJoy targeting established business owners who tried AI tools but are frustrated and overwhelmed. AI has given them so many choices that they don't know what to do, and when they use AI, they don't get the results they want. They're not seeing an increase in revenue, and they see other people using AI and crushing it. PAS framework. Problem: they're overwhelmed by AI, they feel like they have more to do and they're not getting the results they want. Agitation: they're not seeing an increase in revenue, it seems like using AI has made it so they're actually having to do more work rather than less. Solution: oJoy is trained on 26 years of direct response, it analyzes their specific business, then creates and implements a plan that can move the needle the most. It's like having a CMO and a professional copywriter working on your team 24/7. Tone: direct, honest, peer-level, no hype words."
Be honest for a second.
AI hasn't made your business easier. It's made it busier.
You're spending time learning tools, writing prompts, fixing copy that almost works, and watching your revenue stay flat.
And somewhere on your feed there's a guy talking about how AI is crushing it for him. (Not an ideal feeling.)
The difference isn't effort. You're putting in plenty of that. The difference is the AI.
oJoy was trained on 26 years of direct response marketing. Not writing in general. Marketing specifically. It has a conversation with you, analyzes your business, figures out what's going to move the needle the most right now, and then actually does it for you.
It's like having a CMO and a professional copywriter on your team full time. (But without the costs.)
Try it free for 7 days.
And now you know exactly where yours is strong and where it isn't.
But there's a version of this where the direction is almost effortless.
oJoy doesn't expect a perfectly structured prompt. It has a conversation with you. Asks the right questions. Figures out your business, your audience, and your highest-impact opportunity right now. Then creates a plan to generate the most revenue the fastest. And then it implements the plan for you.
You just talk. It does the work.
It was trained on 26 years of real direct response. Sales letters, emails, ads, webinars, infomercials. The stuff that actually makes people buy. And it took almost two years to build. There's genuinely nothing else like it.
Try it free for 7 days. If you don't love it, cancel with one click and pay nothing.