Can you use AI to do this?
AI has become ubiquitous very quickly. From those who use AI chatbots and agents to do everything from handling their daily tasks to offering some form of therapy (I don’t really recommend the last one), it’s everywhere. It’s everywhere even when we aren’t looking for it, hiding in the background of search engines, producing the content we see on social media, even behind the scenes in our software.
AI is used to create text. It’s used to generate huge amounts of content for websites. It can create videos, images, music, and political propaganda. And people are using it to do so. But it can make mistakes, which can be missed if you don’t know how to spot them. Just today, I saw two different posts on social media, reflecting different viewpoints about AI.
AI can do everything, just give it a landing page, and it will create your whole Google / Meta Ads campaign for you. You don’t need an agency to do this.
If your agency uses AI to do anything, then you shouldn’t pay them a penny.
Both are probably heartfelt posts (although it would be deliciously ironic if the second had been written by AI), but both miss the point of AI. And they both miss the point by a mile.
Should you use AI to do this?
Many people dislike the widespread adoption of AI; some are scared of it, and yet many more people have embraced it wholeheartedly. But several facts remain, AI is here and it’s here to stay (even if AI experiences it’s own version of the Dot Com Bubble, it’s not going anywhere), and, at least at the moment, AI is far from perfect. A study done earlier this month shows that even with the latest generation of AI Overviews, the answer given was incorrect 9% of the time.
Yes, you can use it to make everything from a new blog article (this one was written by hand, I assure you) to making creative assets for your new ad campaign. But, are you capable of judging the output? Do you know what makes a good blog? Do you understand the topic enough to tell if the blog is accurate? Does it add anything to your website? Is the ad campaign going to meet your goals? Is it going to actually make you any money?
AI in the hands of someone who knows what they are doing can be an incredible tool; it saves time and can massively increase productivity. But if you get AI to do something that you don’t understand, you have no way of knowing if you are just spending your time making useless junk.
Harmful AI
The potential harm that reliance on AI can have, when you can’t determine the quality of the output, is proportional to the importance of the task you’re applying it to. Unless you have the skills to backstop what you’re asking AI to do, you have to anticipate the worse case scenario.
For an online ad campaign, that means spending the entire allocated budget with zero benefit. For SEO content, that means your entire website being penalised, even a manual action from Google. Don’t even get me started on AI adoption by the US Department of Defense.
This is why, when you see a news article talking about how AI is being used to detect breast cancer more effectively in mammograms, you can be sure that those same mammograms were also reviewed by a trained doctor. Likewise, when your agency uses AI to work more efficiently, you can be sure that the creative, ad copy, SEO content, or code was appraised by a professional who knows what they are doing.
This last one is especially important. When Collins Dictionary announced that “vibe-coding” was the word of the year for 2025, I predicted that the word of the year for 2026 was going to be “catastrophic security flaw” or “massive irretrievable data loss”. Using AI to generate code is all well and good, but if you don’t understand coding and the law of unintended consequences, the potential damage to whatever you’re adding that code to is immense. And all of this is just about those using AI with good intentions and ignoring the potential damage from someone using AI maliciously.
AI for Good
Putting AI in the hands of an expert just means they can get more done. AI handles the time-consuming parts of the job, the output is evaluated by the expert, and either incorporated or discarded based on their considered opinion. In the hands of someone who doesn’t understand what they’re doing, it’s dangerous. In the hands of someone who does know what they’re doing, it’s a powerful tool.
