Side note: Content marketing works!
He thought my takes on the subject were “a breath of fresh air and highlighted some of the realities of using the tools for copywriting.” He asked if I was open to an interview about the topic for an article he was writing about generative AI’s potential impact on online advertising.
I confessed that I had barely used the tools at that point, so he encouraged me to spend some time taking AI for a test drive before the interview. Which I had been meaning to do anyway because I was legitimately curious about what AI could do.
Here’s what I learned from testing a popular AI copywriting tool.
Full disclosure: I’ve primarily used one free AI tool and don’t claim to be an AI expert. This article is based on how I used AI to get actual work done.
AI can help me do my job better. It can’t do my job for me.
I’ve found AI most useful for brainstorming ideas for email subject lines, article headlines, social media posts, and even tag lines. AI won’t give me a finished product that I can use verbatim, but it might give me a word, phrase, or approach I hadn’t considered and can work with.
AI is also helpful for brainstorming topics, general research, and preparing for interviews. For example, if I’m talking to a client about outsourced CFO services, I can ask AI for the pros and cons of outsourced CFO services, and it’ll give me a summary that I can reference during an interview.
I was writing an article for a nonprofit that was founded more than 100 years ago and AI produced a list of major events from the organization’s founding year in a matter of seconds.
Instead of doing a Google search and digging through multiple sources to get the information I need, AI automates the cumbersome aggregation step, which saves time.
Content written by AI is fairly generic.
This isn’t a shock. AI copywriting tools use information that already exists and whatever you feed it. Although the copy is structured logically and grammatically correct, it tends to lack a distinctive voice.
Some tools claim to be able to learn your individual or brand voice based on your prompts and inputs. I just haven’t seen it yet.
I asked AI to write an article in the voice of a certain client. I included specific directions about the desired tone, target audience, and links to content that was authored by this individual.
The tone was basically correct, but the content had no voice whatsoever. You can ask AI to rewrite copy in a different tone and request as many revisions as you like, but will it be any less generic?
This made me think of the hours I spend on calls with clients to get beyond the standard interview questions. How does this make you feel? How does this make your clients feel? Can you share a real-world example of this experience? What happened behind the scenes to make this experience possible?
In fairness to AI, it supposedly produces more accurate, effective results as you use it more often and feed it more data.
I just don’t see how AI extracts what I get from an interview. At least not yet.
Perhaps the greatest potential for AI in actual writing lies with short-form content of no more than a few paragraphs – email copy, social media posts, professional bios, etc. Again, you won’t get a finished product, but AI can move you closer to the finish line.
AI can feed unflattering stereotypes.
I had written a series of articles on the talent shortage facing the accounting industry and what the industry must do to appeal to college students. I wanted to see if AI could write an article of similar quality on the same subject matter.
I entered detailed prompts, including key points, college students as the target audience, and a casual, humorous tone.
The article referenced partying more than once. It referred to accountants as nerds and said accounting is a great career “if you like working with numbers.”
These are exactly the kind of stereotypes that the accounting industry is trying to shake. If I sent an article with any language like this to my client, I’d be toast. Some tools even have a disclaimer that says the tool may produce results that are biased and inaccurate.
By the way, the article targeting college students began with “Hey college peeps.”
Seriously, AI? That’s the best you can do?
AI-generated content might not be based on current information.
For the article about the talent shortage in the accounting industry, I asked AI to include relevant data about the number of accounting professionals at or approaching retirement age, the number of students majoring in accounting, etc.
The tool said its knowledge cutoff date was September 2021. That’s a pretty significant information gap if you want to reference recent trends and data.
A recurring phrase when discussing AI – not yet.
AI is changing daily. New tools are being introduced daily. Everything I’ve just said could be completely off the mark in six months. Or next week. Or tomorrow.
Many capabilities that we wish AI had, especially those beyond copywriting, could very well be on the table very soon. Just not yet.
Since I consider AI a helpful tool, I’m glad it will continue to improve so it can help me save time and produce better results for my clients. But I’m the first to admit that I can’t and have no desire to keep up with AI’s daily evolution.
Test AI but don’t fall for those advertising claims.
One of my biggest problems out of the gates with AI was the marketing messaging, as if it can do everything for you.
Facebook ads for one tool said, “Write blogs in seconds.”
Is that accurate? Technically, yes. Would I publish a blog that was written in seconds and attach my name to it, or my client’s name to it? No way.
Another ad said, “Your blog on autopilot.”
At the very least, this is a gross oversimplification. At worst, it’s a lie.
Besides the writing, you have strategic planning, distribution, performance measurement, and audience engagement. There’s no such thing as a blog on autopilot.
I visited an AI writing website that claims, with basic instructions, you can produce polished drafts in seconds.
Okay, but how do you define a polished draft? There’s more to a draft than making it factually accurate and grammatically correct.
There’s a human problem made worse by AI.
Perhaps the biggest reason why many writers feel like AI will take their jobs, and many brands will try to replace writers with AI, is that the vast majority of marketing content out there – website copy, blog articles, case studies, ad copy – isn’t very good.
We’ve been conditioned to believe that generic content with minimal substance, zero depth, stale advertising cliches, and no distinct point of view is acceptable, as long as you include the right keywords, a few links, and pretty pictures.
The sad reality is that the content AI produces is probably just as good as most of the content consumed on a daily basis. That’s not because AI is brilliant. It’s because the bar has been set so low.
Marketing content written based on interviews, research, and strategy… content that captures the expertise, passion, and voice of the author or interviewee… content that tells engaging, real-world stories… content written for a specific audience to achieve a specific goal… content of this quality is dwarfed by content that was written just to get something out there.
Here’s a harsh reality check.
If you understand how AI works and you’re worried that AI can replace you, then you probably weren’t doing enough to make yourself irreplaceable in the first place.
AI isn’t going anywhere. Now what?
As a brand, marketer, or writer, ask yourself:
Am I developing content with a specific purpose and goal in mind?
Am I producing helpful, useful content?
Am I delivering the message with an authentic, distinctive voice and bringing a fresh perspective to the table?
Am I speaking to the needs, wants, and desires of my target audience with relevance and clarity?
Am I telling stories and sharing experiences directly from the hearts and minds of my team and my clients?
Have that conversation first. Set the bar higher. Make sure you understand and work towards creating content that’s far superior to the drivel that pollutes much of the online universe.
Then evaluate the capabilities, limitations, pros, and cons of AI tools that can improve content quality and the efficiency of content creation. In other words, enhance the work of real people – human writers who not only write well, but understand how to extract the right information and stories from other humans and put sound marketing strategy behind the words.
I’m pretty sure AI can’t do these things. At least not yet.