At every copywriting conference I’ve ever been to, there’s been at least one session on AI copywriting. Often, we’ll see examples of passable and good AI copywriting. We’ll also often see examples of when AI gets it hilariously wrong so we can reassure ourselves we’re safe. But predictive text always used to be hilariously bad – until one day it wasn’t.
So what’s the deal? Do copywriters need to panic? Are the robots coming to take our jobs? It’s been something I’ve been thinking about a lot recently. Here’s my take on it.
AI copywriting at work
I’ve been looking at some of the AI copywriting tools out there. Lots of them are impressively good – and certainly better than I was expecting.
For example, here are some of the outputs I got after I entered writing prompts into Rytr. The tool gives you 300 characters to outline your topic and lets you choose the tone of voice you want and the framework you want it to work to.
I asked it to write to a problem-agitation-solution framework.
The first topic I entered was: “Copywriters write content that help you sell your product or service.” Here’s what it came up with:
Problem: Writing can be a challenging task. A writer can get caught up on how to approach a problem, how to answer a question, or where to start. When writing for marketing purposes, the writer doesn’t want to lose customers or be misunderstood by the customer.
Agitate: Marketing is an essential component of any business’s success. You need creative and compelling content for your marketing campaign in order to stand out among your competitors and engage your target audience. Without this content, you’ll lack good visibility and will miss out on new potential customers.
Solve: A copywriter is what you need in order to have successful marketing campaigns. A copywriter will make sure that the content you produce is engaging and reflects
Not bad. It does outline common problems with the writing process. It does agitate effectively. It does give a copywriter as the solution. It’s a bit bland, but it’s not bad.
Then I entered: “Plumbers fix burst pipes and fit new bathrooms.” Here’s what it came up with:
Problem: It’s no secret that plumbers are expensive, and when pipes burst and you need one ASAP, it can be as expensive as hiring a lawyer.
Agitate: How many times have you been in need of a plumber and had to wait days for them to arrive? If you’re like most people, the answer is far too many.
Solve: Unlike other plumbing companies, we offer an on-demand service. For just $30 an hour, we’ll send the nearest plumber to your house or office within 30 minutes. That’s right — when it comes to plumbing emergencies, we’ll get you taken care of in no time!
I was interested in the angle it took on this until I realised that it had picked up on ‘fix burst pipes’ – it was running with the idea of fixing plumbing emergencies. However, where did the $30 an hour sentence come from? Has it been plagiarised?
I then made the plumbing example more general: “Plumbers fit new bathrooms.” Here’s what it came up with:
Problem: New home buyers are hesitant to hire a plumber to put in their new bathroom, as they are expensive, don’t do the job in a timely manner, and unprofessional
Agitate: The reality is that new bathrooms can be time consuming and if a plumber is rushed or isn’t qualified for the job, it can lead to a disaster. It’s important to find a reliable professional like Gary’s Plumbing, who is dedicated to providing an honest service with quality workmanship.
Solve: Gary’s Plumbing was started by Gary C., an expert in plumbing for more than 30 years. Gary guarantees that he will never overcharge for his services and will always complete the job on time. He has
Here, the grammar isn’t great. The thinking is a little bit all over the place. And again, I’d be really worried the references to Gary’s Plumbing are directly plagiarising someone else’s website. But it broadly makes sense – we’ve all certainly seen worse on live sites.
Then, I asked it to come up with a blog section based on “the robots aren’t coming, AI can’t really replace the job of the copywriter”. Here’s what it came up with:
“AI is not here to replace the job of the copywriter. It is just here to help them do their job better. We should not think of these AI writers as a replacement for human copywriters. They just provide assistance to the content writers by getting rid of writer’s block and generating content ideas at scale.”
I think this is a very interesting take and is undoubtedly inspired by FAQ pages on an AI writing sites somewhere aimed at allaying concerns about robots taking people’s jobs.
Interestingly, when I ask it to generate a blog section for the opposite: “The robots are coming and they will take over the job of the copywriter”, it simply restates the outline in more words:
“It is not hard to imagine a future where AI will be able to generate content at scale, without any human input. It is just a matter of time until this happens.
I can see an AI writing tool as a replacement for human copywriters in the future.”
What do these AI copywriting examples tell us?
I think these examples start to get to the heart of the issue with AI copywriters. Possible plagiarism aside, the examples above are basically perfectly fine, even if they won’t set the world on fire. But how many copywriters can confidently say they’ve never done work that simply dialled it in?
AI copywriters do what they do very well. They take inspiration from words that are already out there and rework them into something that sounds different.
There’s a market for AI copywriting
There are two distinct opinions of copy and content.
On the one hand, you have platforms like Upwork and Fiverr. Here, copy and content is a commodity. You want 500 words on a topic for $10 (or less)? Someone on one of these sites will write it, no problem.
People using platforms like this want words to fill pages or fool an algorithm. It’s content for content’s sake.
For people who think like this, AI copywriting is perfect. It doesn’t say anything new, but that doesn’t matter. It gives them words that fill up pages, which is the goal.
There’s also a market for human copywriting
On the other hand, you have people who understand that good copy helps you make money. It’s why the average copywriter earns £387 a day. And it’s why some businesses are prepared to pay five-figure sums for a landing page – it’s because they know that when it’s good it has the capacity to generate six figures in revenue year after year.
Copywriters charging these rates bring something new to the party – and to their output. They know that good copy is a mix of science and art. For them, the writing element of their work is just one small part. Before they write they’ll research the topic. They’ll assess the competition and their client’s strengths and weaknesses in relation to them. They’ll interview subject matter experts and customers. They’ll take all this information and work out what they need to say and how they need to say it.
As a result, their output offers something genuinely new, something genuinely useful. This is copy written with purpose, copy that’s designed to be read by real people who will find it really valuable. It’s copy that does more than just fill up a space.
What’s the conclusion?
In short, I think there’s a place for AI writing tools in the same way I think there’s a place for content mills. (I don’t think content mills are a good place, but that’s another story.) For people who value copy and content, real copywriters will be the only resource in town – for the time being at least.