In the digital age, artificial intelligence (AI) has become an essential tool for productivity, creativity and problem solvingHowever, to take full advantage of these technologies, It is crucial to know how to communicate effectively with them.
A good prompt, that is, a clear and well-defined instruction, can make a difference. the difference between a vague answer and a precise solution and useful. In this blog, we’ll explore how the quality of your prompts directly impacts the effectiveness of AI responses and share practical tips to hone your technique. Let’s dive deeper into the art of asking questions that actually drive value!
"The art of conversation is not saying the right thing, but knowing how to ask the right questions."
— Aristotle
| Simple | Correcto |
|---|---|
|
How do I add numbers in Excel? |
How can I add a total line in Excel? I want to make it automatic for the entire sheet where in the right column it says Total |
|
Who is the president? |
Who was the president of Mexico in 2021 and how many years are elections repeated? |
|
Write a code to calculate the Fibonacci sequence. |
Write a Python code to calculate the Fibonacci sequence and comment it step by step so that you can understand each component |
|
Summarize the meeting notes |
Summarize the meeting notes in a single paragraph. Then write a checklist of speakers and a checklist of key points. Finally, list any next steps or suggested actions from the speakers, if any. |
It is essential to define the terms well and avoid ambiguity to guide AI towards relevant information. In addition, A good prompt establishes the purpose and limits of the query, ensuring that the response is as closely aligned as possible with what is needed. Thus, by perfecting the way we formulate our questions, we can obtain much more valuable results in our interactions with this technology.
Let's try it!
Let's do a couple of realistic simulations, so we don't stay on the surface. While I was looking for sources to explain this topic, I came across superficial cases where trivial things are asked to AI. In this case, we are going to do the opposite, AI is a great tool to do tedious work that is boring but serves a greater purpose within the production process.
In the following cases I will test this method for cases in which I would use it.
- Let's make an Excel sheet with random data.
This can help us perform data verification tests in our software, or even learn about Excel formulas.
An example of a bad prompt would be:
This may not fit our needs since our data may not be of the same type or we might even want more rows.
An example of a more precise prompt would be:
- Let's have AI generate a JSON (JavaScript Serialized Object Notation) for us
Imagine that you are programming and you want to test your program that consumes a series of data in JSON format. You want to perform tests, but you can't let the AI freak out with random data, you need a series of data of a specific type for each attribute. In this case, AI is a very powerful tool since doing these tests by hand is slow and tedious, we are going to make the AI work for us.
A bad prompt for our purposes would be:
A prompt more suited to our needs would be:
In conclusion, the AI bases its response on the quality and level of detail in our questionThere is no point in asking generic questions for specific problems, these cases are quite specific but can be transferred to other different ones. The quality of the answers will depend on how much they fit what we needed.
It is something that seems obvious but that not many people put into practice, is that AI is not a search engine like Google, needs more information but since it is not Google, if we ask it the right question, it will give us a unique and satisfactory result.
