5 lessons from a year of using AI writing tools

Last January I started using ChatGPT on my copywriting jobs. I was keen to test its limitations, learn how it could help me, and, if I’m honest, to see if my job was at stake.

One year on, I’ve tried ChatGPT (and other AI writing tools) on a whole host of copywriting projects. Think emails, newsletters, adcepts, social posts, concepts, blogs, case studies, web copy – you get the idea.

Here are five key lessons from 12 months with an AI writing sidekick. 

1. Beware of the wow factor

It’s not surprising that we’re wowed by this tech. We can write a prompt and AI will give us an email, blog, social media post – almost anything – in a few seconds. And it sounds pretty good.

The thing is, that wow factor can be a distraction. It’s easy to be so enchanted with the instant output that we overlook some of the issues. 

Is this actually what we wanted? Does it meet the brief? Is it on brand? Has the AI hallucinated? 

Once you’ve enjoyed the ‘magic’, it’s a good idea to look properly at your results. Check all the facts, and make sure the AI writing meets your objectives.

2. Sometimes your brain is best

AI makes a great sidekick for many jobs, but sometimes it’s more efficient to write from scratch. The key is knowing when AI will help you, and when it won’t.

For example, I’ve learnt that AI can’t help much with my conceptual copywriting. The AI results are always too generic or cliched, so I don’t waste time trying to make it work.

It’s the same when writing a short piece of precise copy. It’ll often take me longer to write a decent prompt than craft the piece myself.

3. AI is brilliant at summarising and shortening 

There are so many ways AI can help writers. Summarising and shortening writing are two of them.

Want to turn your blog into a short sentence for a social post? Need to make a long paragraph more concise? Looking to summarise a large report in a few bullet points? 

In these cases, AI is your best friend.

4. Crafted prompts make a difference

As you can imagine, I’ve tried A LOT of prompts over the last year. And, despite the hype around prompt engineering, I do think well-crafted prompts improve your results. After all, if you put generic in, you’ll get generic out. 

In my experience, it’s worth taking the time to get your prompts right. Include your own thinking, be as precise as possible, provide examples, set parameters, ask for suggestions, make it a two-way dialogue. There are lots of ways to do it.

I’d also suggest adopting a trial-and-error approach. Try making small adjustments and comparing results to see what works best for your projects. 

5. Compelling writing still takes time and graft 

AI has made writing look easier and, in a way, it is. AI means anyone can now create a piece of copy in seconds. But if you want your writing to inspire, surprise or stand out, I believe we still need a human writer in the mix. 

You might apply your brilliant creativity before and after using AI (my AI sandwich approach). Or maybe you’ll opt for 100% human writing. 

Either way, if the last year has taught me anything, it’s this: Great writing takes time and thought – with or without AI as your sidekick.

Previous
Previous

How a sandwich can help you write better with AI

Next
Next

A manifesto for writers in the era of AI