Generative AI is a powerful tool for creating content, including PPC ad copy, but it should be used wisely.
It’s not about generating content to copy and paste but assisting in the creative process.
We’ve all seen enough AI-generated social media posts, blogs, and website content to know the web doesn’t need more of that.
When used properly, AI can help spark ideas, refine messages, and combine data with creativity to better connect with audiences.
This article covers practical ways to use generative AI and large language models (LLMs) for ad copywriting without sacrificing originality or quality.
What does AI say it’s good at?
When I began writing about using AI, my first instinct was, naturally, to ask AI itself.
So, I turned to ChatGPT and posed the same question I’m exploring here:
“What are some good ways to use AI to write ad copy for Google Ads?”
Here’s what ChatGPT suggested:
Dynamic headline generation.
Keyword-optimized copy.
Personalized ad copy.
Seasonal analysis (e.g., incorporating “holidays” or “snow” into descriptions during winter).
These initial responses were neither surprising nor particularly helpful. I don’t want AI to write my copy outright, even though it seems eager to do so.
However, this exercise got me thinking – not just about what I don’t want AI to do, but about what it can and should do. Which brings us to the first point.
1. Ideation
My top use for AI in content creation, whether for ad copy or blog posts, is ideation. And few tools are better for this than LLMs.
While their responses are inherently generic, they’re great for sparking ideas and getting your creative juices flowing.
For example, plugging in a prompt like “Give me ideas for how to use LLMs and generative AI for ad copy” can provide a good starting point, helping you figure out what to write – and just as importantly, what not to write.
LLMs compile the most common answers on the web, so copying them directly won’t add anything unique or valuable.
My process is to ask an LLM a few topic-related questions, then use its responses to kickstart my own thoughts and begin writing from there.
2. A/B testing suggestions
Another way to use AI is for iterating on current ads through A/B testing.
If you’re looking for small tweaks to test in headlines or descriptions, you can input your headlines into an LLM and have it recommend different variations.
You can further refine the results by asking it to highlight specific selling points for your business.
In my experience, this method has been more about ideation than direct copying.
For example, when I give ChatGPT a headline or description from my ads and ask for “10 A/B testing suggestions that comply with Google Ads’ character requirements,” I get decent ideas.
However, I still need to tweak them to fit the rest of the copy.
This isn’t perfect, but it’s much quicker than starting from scratch and can provide ideas you might not have thought of otherwise.