The ONE thing in prompting that unlocks ChatGPT’s full potential

May 24, 2024
5 minute read

In today’s newsletter, I’d like to do two things:

  1. Ask you to fill a quick newsletter feedback survey.
  2. Tell you the most important thing when it comes to prompting ChatGPT

So first of all, the survey.

I’ve put together a quick 3 question survey, I think you can fill it out in like one minute. As a thank you for your time and answers, I’ll give away one Prompt Master AI course lifetime access (worth $297) among those who submit their responses. (If you are already a member with us and win it, you’ll be able to gift it to someone else.)

Click here to join this giveaway and fill the form.

Okay, now with that out of the way, let’s talk about the most important thing when it comes to prompting ChatGPT.

I did an online workshop yesterday to HR specialists, and within 60 minutes, I got them from zero to knowing how to write great prompts.

They don’t have to rely on prompt libraries ever again.

And at the end of the workshop, I asked: “What was the most useful thing to you on this workshop?”

95% of them said the same thing.

And their answer was the unlock that gives you all the power of ChatGPT.

The one thing you need to say to utilize the entire knowledge of humanity in an instant.

I like to say that simply having access to ChatGPT is like holding Albert Einstein hostage in your basement, and getting high school students to do math for you instead.

But with this one thing, you can use the Einstein in your basement and solve your problems better and faster.

The answer is: The persona element of the prompt.

This is what can change ChatGPT from “helpful assistant” to “theoretical physicist” or “professional business coach” in an instant.

And this is an essential element to a prompt, so if you don’t tell ChatGPT who to act as, it will just use it’s default “helpful assistant” persona, which we call “vanilla GPT”.

It’s useful for a lot of tasks, but with a good persona prompt, you can finally offload tasks to ChatGPT with calm confidence.

After doing an exercise during the workshop about this element, I asked people:

  1. what’s the quality of the response from 0-100% (0% being trash, 100% being perfect, no changes needed)
  2. how much time would it have taken you to get to this level of quality on your own?

The exercise was only 3 minutes, and most responses were:

90% quality (need to change 1-2 things) and 30-45 minutes to do it on my own.

So right on the workshop, people could save this much time by simply providing a persona and a task to ChatGPT.

We teach this in our course but I’ll give it away for free now.

The persona is part of the context element of the prompt, among two others:

  1. Persona/Role
  2. Tone of voice
  3. Primary objective.

So the template is to:

Act as a [persona] who is [tone of voice] and [primary objective].

For example:

Act as a [copywriter] who is [friendly, funny and informal] and [whose primary objective is to sell as many cupcakes as they can].

Note how the primary objective is more like an overarching goal this person is trying to achieve and not the task description itself, which would be the actual task like “write a welcome email for people who sign up to our webinar.”

So just by using a persona, you can improve the AI’s responses by miles, because it can now be a subject matter expert in ANY industry.

And if you want to save time in your work by using AI, click here and become a Prompt Master today. As a Prompt Master, you’ll be able to build AI automations without writing a single line of code. You’ll be a master of LLM prompting, and building custom GPTs.

And with the support of our amazing community, even if you get stuck, someone will help you out. We also offer unlimited DM support and 90 day money back guarantee, so it’s a totally risk-free investment. When you’re ready, become a Prompt Master!

Please don’t forget to fill the survey so I can provide you even more value in these emails.

Think faster,

Dave Talas

P.S: Sorry about the course platform emails, there was a bug in Circle and you might have received a few extra emails. We fixed it since, but can’t unsend the emails.