How to Use Ai Gemini Like Expert – ChatGPT Ala Freddie Kashawan
Many people feel disappointed because the results of AI such as ChatGPT or Google Gemini often do not meet expectations. However, the real problem lies not in the AI itself, but in the way we communicate with the tool. The output produced by AI is highly dependent on the input given. If the instructions are not clear, then there is no need to expect sharp, relevant, and quality results.
Freddie Kashawan, a content creator and business practitioner, has the right approach in giving instructions to AI. The key is only one: the instruction (prompt) must be clear and structured. According to him, there is a simple formula that is easy to remember before promoting. Just remember T-C-R-E-I.
This formula is an abbreviation of task (task), context (supporting story), reference (reference), evaluate (evaluation), and iterate (repeat). These are simple steps so that AI understands our intentions without the need to repeat prompting.
t = task (task)
The first step is to explain clearly what you want AI to do. For example, writing lists, articles, or making video scripts. In addition, also add who and the desired result format. For example, “You are a technology journalist, write a three-paragraph article about AI trends in Indonesia.”
C = Context (supporting story)
The more complete the context you provide, the better the AI in adjusting the answer style. Freddie stated that AI is not a psychic who can guess what’s in your mind. AI needs information about the situation and purpose. Without context, the results given can be inaccurate or even “inconsequential.”
R = Reference (reference)
This is the fastest way to make the AI “Catch Vibe” you want, such as a specific text, style, or tone example. Give a clear example and state that it is a reference, not just a patch. That way, AI can imitate the pattern without going off the path.
E = Evaluate (Evaluation)
Do not immediately accept the first result—check first whether the output is in accordance with the tone, accurate, and free from the word hallucination. AI is not human, so it needs continuous validation. If the result is still missing, proceed to the last step.
I = iterate (repeat)
According to Freddie, the prompt was like a recipe—not always perfect in the first experiment. However, the more often you change and retest, the more precise the results will be and as expected.
Don’t expect AI to understand what you want without detailed direction. Use the T-C-R-E-I formula so that the output is not only cool, but also on target and can be used in the real world.
This method can be applied in all AI tools—from CHATGPT to AI Image Generator. “Ai is not magical, but he will be magical if you teach him the right way,” said Freddie.
