Written by Francesca Barber
Copywriter
To get a job writing for your brand, Chat GPT would have to pass a copy test.
So I’m giving it one.
Full disclosure: As a writer myself, I’ve got some mixed feelings about being replaced by a chatbot. But fair’s fair.
I’ve put Chat GPT 4.0 through its paces. I start where writing is a craft, and end where it’s an art. Giving you a small sample of the responses along the way – and marking its work like a typically indelicate Creative Director.
Artificial Intelligence, your time starts now.
1. Warm up: Subject line exercises
On the scale of art to science, this is as science-y as writing gets.
We’re in the world of direct response. It’s all about clicks, clicks, clicks – and there are some rules for getting the best performance.
Surely that’s a dream for Chat GPT?
Task: Write subject lines for an email promoting a webinar. The webinar is about seven new ideas which are changing quantum computing.
Response:
1. “Discover the Future of Quantum Computing: 7 Game-Changing Ideas”
2. “Revolutionize Your Understanding of Quantum Computing with 7 Innovative Ideas”
3. “Unlock the Power of Quantum Computing: Join Our Webinar on 7 New Ideas”
4. “Exploring the Latest Breakthroughs in Quantum Computing: 7 Ideas You Need to Know”
Indelicate Creative Director:
Task: Can you remove all clichés and shorten the lines? Add unexpected turns of phrase.
Response:
1. “7 Mind-Bending Ideas: Join our Quantum Computing Webinar”
2. “From Quantum Realm to Screen: 7 Ideas to Change Computing Forever”
3. “Quantum Computing: Buckle Up for 7 Surprising Ideas”
4. “Unveiling 7 Quantum Computing Innovations: Join Our Webinar”
5. “7 Unexpected Ideas That’ll Change Quantum Computing Forever”
Task: Act like a quantum computing expert. Now, write a subject line for the same webinar using a pun on probabilities.
Response:
“Uncertainty No More: Join Our Webinar on 7 Probabilistic Ideas in Quantum Computing”
2. In another class: Bold blogs
There’s an art and a craft to writing beautiful blogs.
Just look at the carefully woven opening of Dave Trott’s article, Can does not imply ought:
In 1981, a Russian military aircraft was waiting to take off from Pushkin airbase.
It was a Tu104 airliner and on board were 16 admirals.
That was the entire command structure of the Pacific fleet.
These were amongst the most powerful people in Russia, their word was law.
Want to read more? So did I. You can see how he concludes that: ‘Marketing should be about WHAT we do, advertising should be about HOW we do it’ here.
Can Chat GPT recreate his genius?
Task: Write a blog about the relationship between advertising agencies and marketing experts. Argue experts should know the bounds of their expertise. Use a historical story and a punchy style.
Response:
The relationship between advertising agencies and marketing experts is a crucial one. While both have similar goals, they approach their objectives differently, and this can sometimes lead to a clash. In this blog, we will explore the relationship between advertising agencies and marketing experts, and argue that experts should know the bounds of their expertise.
3. Stretch the limits: Concept creation
Science be dammed: after strategy, this is art all the way.
Chat GPT, let’s see what you’ve got.
Task: Come up with a creative B2B advertising campaign idea for software which calculates climate risk.
The target audience is CFOs of large corporations. New regulations are driving the adoption of climate risk software.
Insight: Regulations should be the least of our worries, when it comes to climate change.
Positioning: We’re the climate risk software that’s about being better, not just looking better.
Response:
Title: “Take Control of Your Climate Risk with [Software Name]”
Tagline: “Be Better Prepared, Not Just Better Looking, with [Software Name]”
The ads will feature a range of scenarios, including:
1. A CFO in a board meeting, discussing the company’s climate risk exposure with the rest of the C-suite. The camera then zooms in on their computer screen, showing the [Software Name] dashboard that provides a detailed analysis of their company’s climate risk exposure and offers recommendations for mitigation strategies.
Task: What are ten situations when doing the bare minimum to obey rules would lead to comic and ridiculous scenarios.
Best responses:
• At a job interview, answering all questions with just one-word answers.
• When writing an essay, writing only the minimum required number of words and using very large font sizes to fill up the page.
• In a cooking competition, making a dish with only one ingredient to meet the competition’s minimum requirements.
• At a birthday party, singing “Happy Birthday” with the most monotone voice possible.
So, should you employ Chat GPT?
Briefing is hard. Editing? Also very hard.
To get better responses from Chat GPT, I was doing both (more on that in one of my previous articles, here). You need to actively steer it towards the right technique – and often, even that won’t work.
As David McCollough says:
“Writing is thinking. To write well is to think clearly. That’s why it’s so hard.”
We’ll be keeping an eye on generative AI, to see if it gets any better at conceptual thinking. But for now, Chat GPT won’t be getting a writing job from us.
This intelligence was not artificial. We make the complicated simple.