Sign in or Join the community to continue
DIRECT prompting methodology walkthrough
Posted Feb 09, 2026 | Views 92
# DIRECT
# Prompting
Share
SUMMARY
The DIRECT prompting methodology helps you create clearer, more effective prompts by thinking through six elements: Doing, Information, Role/Persona, End Goal, Context, and Tone/Style/Format. By giving tools like ChatGPT the right instructions, relevant information, perspective, purpose, context, and output expectations upfront, you can reduce back-and-forth and get more tailored, actionable results faster.
+ Read More
CONTENT & TRANSCRIPT
0:01 Hello everyone! In this video, we're going to review the DIRECT prompting methodology. Remember, large language models are trained on the world's wealth of information.
0:10 Because of that, when an LLM responds to a question that you ask, it has lots of context to pull from.
0:16 With good prompting, you can get an answer that's exactly what you want faster because you're narrowing that aperture of information that the model can respond to you with.
0:27 So DIRECT stands for:
D: Doing. This is where you describe what you want the LLM to do for you.
I: Information. This can be done in a few ways. One, you can use that @ symbol, which activates the apps that your ChatGPT can connect to.
0:44 So maybe you have contextual information in ClickUp and you want to connect your ClickUp. Or in Gmail, you can connect your Gmail.
0:51 Another way of adding that information is by uploading, for example, a PDF or a spreadsheet file. And then the last approach is that you can just add those informational details directly into the information section of your prompt.
1:05 R stands for Role or Persona. This is where you give the model the perspective that you want it to create its answer from.
1:14 This is where you can describe the personality traits, background experience, or any specific systems or processes that persona is an expert in. Define those in the Role or Persona section.
1:27 E stands for End Goal. You're communicating to the model the purpose, or what you want to be able to do with the output that you're given.
1:35 C stands for Context, where you add any unique details or attributes about that situation.
And T stands for Tone, Style, or Output Format.
1:47 If you have a tone saved in managed memory, you can reference that. But this is also where you specify how you want the output to be formatted.
1:59 Let's dive into an example.
If I go to my ChatGPT, I'm going to start with a bad prompt. For this, I'm just saying, "Hey, I want you to help me create a plan for organizing four company-wide team-building activities starting March 1st, 2026."
2:17 And really, if we look at DIRECT, that's probably just the D of my prompt.
2:24 You can see ChatGPT will give me an answer, but most likely it's very vague. It's potentially not very actionable, and it's not tailored to my company because I'm not giving it any of that context.
2:40 So now let's look at the DIRECT version of this prompt.
I'm going to do this using my threading feature. Remember, with this little pencil, you can create multiple versions of the prompt that you can just flip through.
2:52 So let's do that. I'm going to paste in my DIRECT prompt here.
You can see that in this prompt, I'm now providing more information about my company constraints, including the budget.
3:04 Through my Role section, I'm telling it that I want it to focus on a cross-office, hybrid workforce context. I'm also giving it a very specific output format that I want to see.
3:18 So I'm going to send that, and then we'll compare the first and second outputs.
You can see that because I asked for the event details, this version, for one, is a lot more actionable.
3:35 If we look at that first version, it just gives me some example activities, but it doesn't give me any details about how to implement them.
3:49 Coming back to our DIRECT prompt, you can see it's also very tailored to my company. It says, "LearnAIR Exchange SkillSwap Week," and then it gives precise details about how that is to be executed.
4:00 It takes into account the fact that we have groups of six to eight people. It takes into account the fact that we need to have Zoom sessions and recorded Looms because I mentioned that we have a hybrid workforce.
4:12 Because I used the DIRECT prompt, you'll notice that I have to do a lot less back-and-forth, and that output is a lot closer to what I would have wanted.
4:31 This just shows the impact of effective prompting on streamlining your work.
The other thing that I love about the DIRECT methodology is that it provides a checklist framework for me in my head.
4:43 As I'm creating a detailed prompt, I go through each of those elements: Okay, what is my D? What is the Information? What is the Role that I want it to answer from?
4:49 It provides that nice framework to ensure that I have a thorough prompt anytime that I interact with an LLM.
4:59 So hopefully you found this reminder helpful, and hopefully you continue to prompt using the DIRECT prompting methodology.
Thank you so much for watching. I appreciate it.
+ Read More
Sign in or Join the community
Don't Forget The Human Part!
Create an account
Don't Forget The Human Part!
Like
Comments (3)
Popular
Watch More
2026 model comparison video
Posted Feb 04, 2026 | Views 62
# Model comparison
# Digital employee
# Image creation
# Managed memory
# Personalization
What is AI today?
Posted Sep 10, 2026 | Views 0
# Large language models
# ChatGPT
# Claude
# Gemini
# Copilot