Are you using ChatGPT in your writing workflow? If so, how are you using it?

I’m asked questions like those pretty frequently. And I get it—the robots are all the rage these days. It feels like everybody is talking about them. 

Admittedly, I resisted AI for quite a while. Partly because I’m a creature of habit who steers clear of any type of change and partly because I thought AI’s output was inconsistent (or, even less generously, consistently crappy).

For the record, I still think that a lot of the time. And if you want to hear about some of my broader thoughts on AI and content quality, you can find those in this newsletter I wrote a few months ago. 

But I’ll admit that I can only see, “Adopt AI or watch your career go down in flames!” fear tactics in my LinkedIn feed so many times before I talk myself into seeing what all the fuss is about.

So, let’s answer the first question: Am I using ChatGPT? Yep. I am.

Even typing that makes me feel a little disloyal to my craft, but the reality is that I have found some ways to use it that make me more efficient without keeping me up at night worrying about plagiarism or factual inaccuracies. 

That brings us to the second (and meatier question): How am I using ChatGPT?

Here’s the simplest way I can sum it up for you: I use it to help me with thinking, not writing. 

What the heck does that mean? I’ll give you a few examples of some ways this friendly yet sometimes dim-witted little robot has snuck its way into my processes: 

1. Simplifying complex concepts
Example prompt: “Explain [concept] to me like I’m in kindergarten.”

I recently wrote an article about flowchart loops. As someone with very little (read: no) coding or programming knowledge, all of the existing research and resources I read on the topic made my head spin.

So, I asked ChatGPT to break it down for me. And surprisingly, it spit out a simple, real-world example that made me go, “Ohhh, so that’s how it works!” Even better, I was also able to adapt and improve that example to use it in my final piece. 

2. Uncovering unique angles

Example prompt: “Create an outline for a blog post about [topic] written for [audience].” 

I’ll be the first to admit that ChatGPT’s outlines can feel a little choppy and clunky. However, I usually ask it to generate one anyway.

Why? It often includes a section or two that I wouldn’t have thought of myself. So, it’s still a worthy exercise for coming up with a more comprehensive approach or unique direction for my own post—even if I ditch a lot of the rest of the outline. 

3. Brainstorming interview questions

Example prompt: “Give me [#] questions I should ask [type of expert] about [topic].” 

Similar to the above, ChatGPT has also helped me come up with some interview questions to ask subject matter experts. 

Again, some of them are good, some of them are bad, and nearly all of them need some tweaking. But it’s a helpful starting point (especially when I’m struggling to distill all of my thoughts into actual questions to ask someone). 

4. Creating examples

I don’t have an example prompt for this one because it will vary widely depending on what you need it to spit out for you.

But, to give you an anecdote from my own experience, I had to write a blog post about LinkedIn summaries. I sourced plenty of real-world examples but also had some instances when I needed to generate generic, example summaries for various jobs.

In those cases, I told ChatGPT something like, “Write a professional summary for a [job title] with [X] years of experience in [industry]. Use first-person language and write at least three paragraphs.” 

5. Generating headlines and meta descriptions

Example prompt: “Give me 10 headline ideas for a blog post about [topic].”

Example prompt: “Write a meta description for a blog post about [topic]. Use the key phrase [key phrase] in the description and make it less than 155 characters.”

Okay, I’ll just say it: I think ChatGPT’s headlines are bad. This is probably a me problem, but I have yet to find a way to get it to give me a headline that doesn’t start with some bogus phrase and a colon (they’re all some version of “Unlocking the Future: Blah Blah Blah”). 

But even so, sometimes it spits out a word or idea that I want to use. And as someone who admittedly hates writing meta descriptions, it helps get the wheels turning for those too.

Listen, I’m by no means an AI expert. I’m sure most of my prompts are laughably bad and I’m still learning a lot about how to be better at getting it to give me what I need. I run into plenty of instances when I feel like I need to feed it so much information that it would’ve been quicker to just do the dang thing myself.

I also know I’ve mentioned this several other places, but it feels worth reiterating: I never copy something out of ChatGPT and paste it into my work. I use it as a starting point and tweak, tailor, and fact check (!!!) stuff pretty relentlessly. 

I’m too afraid to use exactly what it gives me, honestly. Plus, I know my clients would see right through it because, no matter how much I beg it to sound like a normal, conversational human, it just can’t pull it off. And it sure as heck doesn’t sound like me.

With all of that said, I’m doing my best to embrace it as a tool that’s more than likely going to be around for a while and am taking inspiration from the clever ways other writers are using it (I loved this post from Bani Kaur earlier this week). 

In short, am I totally pumped about AI and all-in on its value? Not yet. For now, I’m thinking of it like a semi-obnoxious coworker. Do I love it? No. But it’s here, it’s doing some decent work, and I figure we might as well learn to play nice together. 😉