After my last email a lot of you asked.
So, here’s how I use ChatGPT in my copywriting research process.
Keep in mind, this approach works for copywriting, but it can work for any other field where you need to take in a lot of written material and resources and “make sense” of it.
First, what was the problem before starting to use this method?
For me specifically as a perfectionist, the research often felt overwhelming.
At some point a good copywriter has to call it a day and take a stand, make a decision – based on his expertise – on what good enough looks like.
But I was often spiralling into a never-ending discovery bout of madness. Because I love it and because I think you never know enough.
Anyway, what I realized using ChatGPT is that you can use it to help you set boundaries around your thought process.
In other words, as long as you know how to put the pieces together, good solid boundaries that tell you “you have all this data but these are the 3 major themes that stand out, because they were mentioned a lot”, make easier to take that in and come up with great insights.
But it’s not so easy to make ChatGPT work that way. Let me explain…
ChatGPT is a predictive machine. You write a prompt and it returns a prediction based on the information contained it its “brain”.
Plus it still cannot go and grab text from websites on its own.
So you have to instruct it.
This is where most of the work lies.
My 3 step approach works something like this [comments in square brackets are mine for you]:
- Collect the raw data [can be copy from competitors site, customer reviews, transcript sections from interviews etc.]. Here you need to have a clear idea of what data you need. Nobody can do it for you.
- Feed it to ChatGPT. Be careful here to provide context on what that data is too. Treat ChatGPT as a junior copywriter or helper. You might say something like “Here’s a few snippets from an interview I did with the Founder at XYZ startup -> Paste snippet -> Let me know if you got it [you want to acknowledge it before having it spit out stuff]
- Ask for insights or a recap. This is where the magic happens. Once ChatGPT has gotten all the data in, asking something like “Based on this information from the interview with the founder [don’t take these references for granted as they help providing context around what ChatGPT needs to consider, remember it’s still dumb], how would you define their brand voice? How do they stand out against the competition?”
That’s it…
Next step is building on top of each insight you get. The knowledge compounds when you keep prompting for learnings.
For example in the last point, when I asked how the company stands out against the competition, if previously I had fed CHatGPT with information on competitors’ value proposition, its answer could have been even more relevant and insightful.
In the end you only need to know a couple of things:
- For how smart it seems, ChatGPT is still dumb, being redundant often gives you better results (for now)
- Always feed it with real-world data unless you want simplified guesses that might derail your research
- Consider the sequence of prompts. As each one builds on what came previously, it’s important go in the right order. Think of it as your junior helper. They don’t know what they don’t know.
Hope these are helpful.
P.S. I’ve been working with a communication coach for the past 3 months and it’s going great. Lots of learnings and new perspectives. I’d love to share them with you and I think a lot of it can come in handy when “communicating” with ChatGPT and AI tools. It’s the human stuff these bots still can’t understand.
I’ll share more this week.
In the meantime if you need help with your messaging, just get in touch.
Quote and reflection of the day:
“Perfectionism is like a straitjacket that you need to break out of if you want to take risks, go bigger, and succeed on the highest levels.”
- Ryan Serhant, Big Money Energy
Achieving your goals, completing a project successfully, reaching a big milestone… none of these is about getting it perfect. It’s about risking not to get it perfect in favor of getting it done. Intention, not attachment.