I've used ChatGPT for 1.5 years, creating over 200 pieces of content and having 2000+ chats.
3 tips I always follow for a better output:
Yes. I know. It’s been quite a ride.
I have created so much stuff with ChatGPT. Bad. Good. Amazing. I still explore ChatGPT along with many other AI tools and share what I learn on LinkedIn daily.
There’s so much I can talk about to help you create better content with them. But for now, take these 3 snacks I usually use for writing better with AI:
1: Add context.
More than half of you just ask ChatGPT. No context. No specificity. very less key details.
So what happens?
With a non-specific question comes a non-specific response. A response that is just some fluff you probably won’t use.
For the quality of the output, work on the prompt. Don’t just write what comes to your mind. Spare some 10 minutes and think of what context you can provide about what you want.
Think of the Whats, Whys and Hows of the context.
Imagine this:
“How can I write a ghostwriting cold email?”
The output for this will be too general. Because there’s no target audience and no known goal of what to achieve. There are many ways it can be improved.
For example, check out this prompt:
“I am a ghostwriter with over 5 years of experience. I am now looking for new clients to offer my service to. I can help gym owners write technical copies for social media to reach more fitness-passionate people. I want you to write a cold DM for my prospect in a professional but friendly tone. I also want you to include that I have already helped 5 clients reach +500k audience each in the last 1 month.”
Noticed anything? A lot probably.
The second prompt has enough context. The ICP is clear. Your goal is clear. The tone and the voice are specified.
If you look at the prompt's beginning, you’ll see the writer is identified. The industry he is in and his background experience.
Why did we even specify the amount of experience?
To write from an experienced person’s perspective. It builds trust and authority.
2: Tell ChatGPT to assume an identity.
ChatGPT can write. Be it from a doctor’s point of view or a farmer’s.
And assuming an identity can be done in many ways. One clear way of doing it is using the “Point of view” technique.
Let’s try it with a prompt:
“From the point of view of a marketing manager, explain how a clothing brand can be developed.”
“From the point of view of a human resources manager, explain how a clothing brand can be developed.”
Now check out the outputs for both.
From the marketing manager’s point of view:
“A clothing brand can be developed by identifying a unique market niche, creating a strong brand identity, and consistently delivering high-quality, stylish products that resonate with the target audience.”
From the human resources manager’s point of view:
“A clothing brand can be developed by building a talented and motivated team, fostering a positive and inclusive company culture, and investing in continuous employee development to drive innovation and excellence.”
3: Make sure the output isn’t made up.
Yes. ChatGPT can make stuff up. It may also lose track of the main idea and go another direction.
In such times, try to push it for justification.
“How to do that?”
Ask it questions like “Why do you think this is true?” or “Can you provide any evidence for what you just generated?”
It may provide a proper reasoning or come up with an even better response. In both scenarios, the output will be better than the previous one.
But giving up too soon will make it harder for you to get the best from ChatGPT.
Many times, the first response won’t be the best one. Imagine, they put the Regenerate button at the end of the chat for a reason.
PS: Do you prompt ChatGPT or just ask it questions?
What difference do you see between the two?