AI and copywriting: everyone has an opinion, but most advice is either overly technical or uselessly vague. After two years of trying out different AI tools and using it as an assistant for early drafts, here's what we've learned.
The internet is full of "ultimate prompts" with seventeen parameters and specific formatting instructions. In theory, prompt engineering can produce impressive results. In practice, it's overkill.
Unless you're running the same prompt hundreds of times, you're better off having a conversation with your AI. Start simple, iterate on the output, and refine as you go. It's faster and produces better results than trying to create the perfect prompt on your first attempt.
We recently made a video covering this, along with some other practical tips. The core message is simple: stop overthinking the setup and start having better conversations.
The best way to have those better conversations is to write a briefing like you would for a colleague. Not someone who knows everything about the project, but someone smart enough to work with the information you give them. Include:
Think of it as explaining the task to a talented junior copywriter. Give them context, not commands.
Here's a trick that makes a real difference: use different conversations for different types of copy or audiences. One conversation for LinkedIn posts aimed at CTOs, another for blog content explaining technical concepts, and yet another for internal documentation.
Why? The AI learns from your corrections and preferences within each conversation. Keep them focused, and you'll notice the tone of voice becoming more consistent over time without you having to explain it every single time.
The best prompt is a conversation, not a command. Start simple and iterate!
This is probably the most important skill for any copywriter using AI: recognising AI's favourite words. Fostering, empowering, leveraging, innovative solutions, seamless integration, ... It might sound professional, but it's a red flag to anyone who reads a lot of AI-generated content.
AI has default patterns. It loves certain phrases, certain structures, certain rhythms. Your job as a copywriter is to recognise these patterns and edit them out. The best AI-assisted copy doesn't sound like AI wrote it. It sounds like you wrote it, just faster.
No matter how good AI gets, it can't replace subject matter expertise. If you don't understand what you're writing about, the AI will happily generate plausible-sounding nonsense.
This is especially true for B2B tech content. If you're writing about database architecture or cloud infrastructure, you need to know enough to spot when the AI is making things up or being superficial.
The unbeatable combination is a copywriter who understands language, knows the subject matter, and can use AI to accelerate execution. That's what clients are actually paying for.
We sometimes feel guilty about how demanding we are with AI. We ask for five variations, discard 90% of the output, and iterate endlessly. But it's important to remember that it's a tool, not a teammate. You wouldn't hesitate to rewrite your own work multiple times to get it right. The same principle applies here.
Start by asking for 3-5 variations with different tones and structures. Pick the best elements from each. Combine them. Iterate again. Be demanding. The AI doesn't have feelings, but your clients do. So you better make sure that the result matches their expectations.
Most people know ChatGPT. A few have tried Google's Gemini or Anthropic's Claude. Even fewer compare them regularly.
But different AI models have different strengths. One might be better at technical accuracy. Another might nail conversational tone. A third might handle complex instructions better.
They all have free tiers, some more generous than others. Try them for your specific use case. You might be surprised which one works best for your style of copywriting.
On our experience, here's what actually works in practice:
Notice what's missing? Complex prompts, rigid templates, or trying to get it perfect on the first try. These kinds of things just slow you down.
AI doesn't make you a better copywriter. It makes you a faster copywriter, but only if you know what good copy looks like in the first place.
Your taste, your subject knowledge, your ability to recognise when something sounds generic or fake, that's what matters. AI just helps you get there quicker.
Want more tips on using AI to create better copy?
.avif)
