It has only been a few weeks now since the launch of AIPRM. During this time, there have been some high-level learnings about AIPRM prompts from the first couple of weeks:

1. A prompt is more than a sentence:

So many users share a prompt that looks something like this: “behave like an e-commerce EX bird and create a homepage headline for an industry called product for this brand” That’s it, just one sentence. It’s like you going to a community and asking a question like, “can someone help me master artificial intelligence?” What do you expect with this little brief command? You will only get unsatisfying results and be frustrated and say that ChatGPT sucks. One-sentence prompts yield very poor results; the web is full of these one-liners. If you visit LinkedIn any day of any week, you’ll see these lists with hundreds or thousands of prompts out there that are all just one-liners. AIPRM takes care of one-liners, ensuring that the prompts are more than one sentence. So here’s just one example:

It doesn’t fit on the screen. It also has a lot of very specific instructions because that’s how you get the most out of a large language model like chatGPT.

Looking at this prompt inside AIPRM, you will see a tab like this:

 A little card describing that this is the keyword strategy prompt template and that it helps you create a keyword strategy and an SEO content plan from one keyword only with one click. What this does is that it takes what we’ve just seen and fills in the variable.

Prompts are the new code, and this keyword strategy card is an advanced prompt template than the one sentence. So when we type “AI prompt engineering” into the search box provided, we immediately get a table like this:

This is a table derived from the keyword strategy prompt template for the keyword cluster, the search intent, and the Meta description for all these keywords that ChatGPT generates. This can’t be done using one-sentence prompts.

2. You can ask for more output:

AIPRM, which is now like secondary support for ChatGPT thanks to its over 200,00 users, receives complaints every single hour that ChatGPT suddenly stops or is unable to do as many words as required by a user. So many people do not know that you ask for more output from ChatGPT. Consequently, AIPRM built an advanced continue button that not only allows you to ask ChatGPT to continue writing but you can also ask it to clarify, exemplify, expand, shorten- you name it.

However, a user has to instruct it to do these things else ChatGPT will stop on its own as it certainly can’t read your mind. So to get the best results, a user has to instruct ChatGTP in a very detailed way; AIPRM makes that possible. 

3. You can ask for tone and style:

Users frequently report that ChatGPT’s output is dull and lacks good writing. However, these complaints typically stem from users who provide very brief prompts. The IT principle of “garbage in garbage out” applies here, as without detailed inputs, the output will also be mediocre. AIPRM provides a solution to this issue by offering various tone and style options to improve the quality of the output.

4. You have to give assertive instructions:

If you want to get the most out of ChatGPT or any large language model, giving specific and assertive instructions is essential. ChatGPT is like a new junior team member with a limited memory of only 3,000 words, so don’t expect it to remember what you said yesterday. You need to provide clear and concise prompts to get the best results. Remember, it can’t read your mind.

5. There is extremely high traffic for top prompt templates:

Extremely high traffic is being sent from Little offer Links off the prompts in AIPRM to wherever you write about and explain your prompts. Though causing a significant increase in traffic to AIPRM’s prompt templates, this has also made AIPRM susceptible to spam attempts. Nonetheless, AIPRM takes a definitive stance by implementing strict guidelines, rigorous moderation, and automatic detection of duplicates and canonic content. In addition, AIPRM relies on user reports to facilitate takedowns and ban problematic users.