Write Better Prompts: The Results-Based Technique (2 Simple Steps)

Want to write better prompts for AI tools like Google Gemini or ChatGPT? Getting truly great outputs often comes down to the prompt itself. But many approaches focus too much on telling the AI how to do something.

My business partner Mercer found a more effective way: results-based prompting. This technique shifts your focus to clearly defining the end result you want and the intention behind your request. It's a simple but powerful change that dramatically improves your prompts by leveraging AI. 

In the video below, Mercer explains and demonstrates exactly how this prompting technique works. 

Watch how easy it is to write better prompts in 2 simple steps

Ready to see how Mercer uses AI to improve his prompts? In the video, he breaks down the result-based technique to reduce AI hallucinations and get faster to desired results, every single time.

Key takeaways

• Prompt for results: Define your desired end result and the intention behind it, rather than telling the AI how to do its job.
• Check if AI understands what you want: Get feedback from AI before generating content. 
• Let AI optimize your prompt: Leverage the AI's power to help you write the perfect prompt based on its understanding of your needs. 
• Save time & reduce frustration: This results-based approach minimizes AI errors and misunderstandings.

Why your current AI prompts might be falling short

If you've ever felt frustrated by AI output, it often boils down to a simple misunderstanding in communication.

The way most of us communicate with AI is fundamentally flawed, and that's often why we get frustrating or unexpected results. – Jeff Sauer

The most common prompting mistake is telling the AI exactly how to perform a task, step-by-step. 

We treat it like a simple instruction-following machine. 

The problem with this technique is that we might not actually know the best process ourselves!

When we prescribe the ‘how,' we limit the AI's ability to use its vast knowledge to find the most effective path to the result we truly need.

This leads to situations where the AI gives you precisely what you asked for, but it's still not what you needed or expected. 

These ineffective prompts often result in responses that are

• generic, surface-level
• slightly off-target 
• filled with AI hallucinations 

They leave you to tweak and re-prompt endlessly. 

Understanding this limitation is the first step towards dramatically improving your prompts.

Introducing results-based prompting to improve your prompts

So, how do we break free from those limiting, instruction-based prompts? 

Mercer found the solution in a different approach to AI communication: Results-Based Prompting.

Instead of telling the AI the exact steps to take, his technique focuses on clearly defining two key things:

1. The desired end result: What does success look like? What specific outcome are you trying to achieve?
2. The hidden intention: Why do you want this result? What's the underlying goal or problem you're trying to solve?

By making sure the AI has clarity on the ‘what' and the ‘why,' you empower it to use its full capabilities to determine the best ‘how.' 

This technique transforms AI from just a fun tool or ‘fidget spinner' into a genuinely powerful partner for emails, sales copy, and so much more. – Mercer

This fundamentally shifts the dynamic. 

You're treating the AI less like an order-taker and more like a knowledgeable collaborator. 

This is the core principle that will help you significantly improve your prompts and unlock far better outputs.

A simple 2-step guide to writing better prompts 

As Mercer showed in the video, this powerful method for improving your prompts involves two key steps. 

It might seem like an extra effort at first, but it quickly ensures the AI is perfectly aligned with your goals before it generates the final output.

Step 1: Clarify your goal for better prompt understanding

The first phase is all about making sure the AI truly understands what you want and why you want it.

1. Write your initial prompt 

Start by writing your prompt as you normally would. Don't worry about perfection here; just get your initial request down. Mercer sometimes calls this the “vomit draft”: just get the idea out.

2. Ask AI to analyze and rate your prompt

Instead of running your prompt directly, ask the AI to analyze it, as Mercer shows in the video. 

He does this by using his own “What’s the Result Prompt?”

You can copy and paste Mercer’s first super prompt in Gemini (or any other AI chat app): 

Please analyze the following text and rate your understanding of the desired result on a scale of 1 to 10 where10 is you fully understand the desired result. Explain your reasoning. Then rate your understanding of the hidden intention behind the result.Explain your reasoning. During a rating, if you did not rate yourself a 10, list the questions you have and the most useful answers for those questions. Repeat your analysis until you have a 10/10 understanding for BOTH desired result and hidden intention behind the desired result. 

[PUT YOUR ORIGINAL IMPERFECT HUMAN PROMPT HERE]

The part between square brackets is where you enter your very own prompt.

3. Review and refine 

The AI will explain what it understood and give a score. You need to review this critically before continuing.  

Does it accurately capture your goal and why you want it?

• If yes (e.g., it gives itself a 9 or 10 and you agree), great! Move to Step 2.
• If not, provide clarification. Tell the AI what it missed or misunderstood regarding the result or intention, and ask it to confirm its updated understanding until you're both aligned.

Repeating this process is important and Mercer calls it the ping pong prompting technique.

Below is a screenshot of Mercer showing how Gemini AI analyzes his human prompt to write the copy for a sales page for an AI training of ProfitSchool.

Step 2: Let AI help you refine and optimize your prompt

Once the AI confirms it fully understands your goal (a 10/10 understanding), you leverage its power to craft the optimal prompt.

1. Ask AI to write the better prompt itself

Now, use a prompt similar to Mercer's second “super prompt”. 

This one, he calls “Create the Better Result Prompt”:

You can copy it below and save it for later reference:

Please write a prompt that is highly likely to produce a result even better than the original desired result and fully satisfies the hidden intention behind the result. Rate your prompt on a scale of 1 to 10 where 10 is your prompt is highly likely to produce a result even better than the original desired result. Explain your reasoning. Then rate your prompt on a scale of 1 to 10 where 10 is your prompt that is highly likely to satisfy the hidden intention behind the even better result. Explain your reasoning. During a rating if you did not rate yourself a 10, list the questions you have and the most useful answers for those questions. Repeat your analysis until you have a 10/10 prompt likely to deliver BOTH an even better result than the original desired result and hidden intention behind the even better result.

2. Review the AI's prompt and rating

The AI will generate what it considers the best prompt. 

It might even rate its first attempt slightly lower (e.g., an 8/10 or 9/10).

3. Request refinement (if needed) 

If the AI doesn't rate its prompt a 10/10 immediately, ask it to refine it with another, simple prompt like this:

“Okay, how could you improve that prompt to make it a 10/10?"

4. Get the final “perfect” prompt 

The AI will provide the refined, 10/10 prompt. 

This is the highly effective prompt you'll use.

In Mercer’s workshop, that looked as follows: 

Now you have an AI-optimized prompt: 

• built on a clear understanding of your goals
• ready to generate a much better final output than your initial attempt likely would have.

Now, it’s time to reap the reward of Mercer’s two-step refinement process!

Putting your improved prompt into action

You’ve successfully guided the AI to create a high-quality, 10/10 prompt perfectly aligned with your goals. Now what?

Simply take that final, optimized prompt generated at the end of Step 2 and use it to get your actual desired output. 

Pro tip: It's best to start a fresh chat or interaction with the AI. 

Paste in your refined prompt, and watch as the AI delivers a response that's significantly more accurate and useful than what your initial prompt likely would have produced. 

In the video, this is what Gemini wrote for a sales page. 

And although it is not perfectly aligned with our offering, this is definitely a solid basis.

Benefits of using results-based prompting to improve prompts

Investing a few extra moments in this two-step prompting technique yields significant advantages that go beyond just getting an answer from your AI.

Mercer demonstrates this in the video and in the screenshot below, you can see that

• AI explains its reasoning.
• Tells where it is missing information to generate the best possible response.
• Fills in the missing pieces itself.
• The AI gives itself a score.

Mercer’s technique has clearly some powerful benefits:

• Write Significantly Better AI Prompts: By focusing on results and intention, you naturally create clearer, more effective prompts that leave less room for AI misinterpretation.

• Get More Accurate & Relevant Results: When the AI fully understands your goal, the final output is far more likely to be precisely what you need, not just what you literally asked for.
• Reduce Errors and Hallucinations: Clearer, AI-refined prompts minimize confusion, leading to fewer nonsensical or “made-up” responses from the AI.
• Save Time and Reduce Frustration: Less time spent re-writing prompts and correcting flawed AI outputs means you get to your desired result faster and with less hassle.
• Improve Your AI Communication Skills: Practicing this technique trains you to think more clearly about your desired outcomes, making all your AI interactions more effective.
• Unlock Deeper AI Collaboration: You shift from simply giving instructions to collaborating with the AI, leveraging its power to help you define and achieve your goals more effectively.

Imagine working with someone who gives you what you needed, not just what you asked for. That's the potential of AI when you focus on the result. – Mercer

Essentially, results-based prompting helps you move from basic instruction-following to truly strategic communication with AI.

What to do next? 

You now understand the power of Results-Based Prompting and the simple two-step process to implement it. Don't just let this be theoretical knowledge – put it into practice! Here’s what you can do next:

• Apply it immediately: The best way to solidify your learning is to use it. Pick a task you regularly use AI for (like drafting emails, brainstorming ideas, summarizing text, or even writing code) and consciously apply the two-step results-based method today.
• Dive deeper: If you want to become an AI first business, you can best watch Mercer’s Gemini Workshop.

Good luck with having AI craft better prompts!

Answers to your most common questions about this technique to write better prompts

Do I have to use the exact “super prompts” Mercer uses in the video? 

No, Mercer's prompts are working great, but you can adapt the wording as long as you ensure the AI analyzes its understanding of your goal (result + intention) and then generates/refines the optimal prompt based on that understanding.

Can I use results-based prompting with any AI model (Gemini, ChatGPT, Claude, etc.)?

Yes, this technique is about clear communication and goal definition, so the core principles apply effectively across different AI language models. But as these are constantly evolving and not all tools give the same results, you may need to test and finetune them yourself.

Why is identifying the “hidden intention” so important for improving prompts? 

Identifying the ‘hidden intention' gives the AI crucial context about why you want something, allowing it to make better decisions and provide more relevant, nuanced results that go beyond just fulfilling the literal request.