A.I. REVISITED
By Al Navarro, Co-Founder and Chief Creative Officer
In February 2025, I wrote about A.I. in Mojito. The piece was titled “Al on A.I. Or: How I learned to stop worrying and love artificial intelligence.”
It started with the following passage: “It’s no secret that I was not a big fan of A.I. (artificial intelligence) until a short time ago.
Here’s why: As recently as a year or so back, it wasn’t very good. In fact, based on my experience with it, the technology wrote at about the level of a junior copywriter.
This might sound great to you, but put yourself in my shoes. In my nearly 35 years of experience working at ad agencies of all sizes, junior copywriters get it right the first time about 50% of the time. Creative directors and departments bake this factor into schedules, staffing, etc. And it’s all part of staff development. The juniors share their work with more experienced writers and creative directors and get tips on what’s good/bad/how to improve.
But at a boutique agency like Mint, a 50% hit rate doesn’t really help. It means I have to invest as much time in revising the copy as I would if I had written the piece from scratch.”
I went on to say that I was ready to elevate A.I. to full copywriter status. But in hindsight, I still don’t think it’s quite ready for that promotion. Which is not to say that A.I. isn’t helpful. It can be.
How I use A.I. in 2026
I should start by reiterating that I still do not use ChatGPT to write headlines or taglines. As I noted in my 2025 musings “Call it personal/professional pride. I also feel like, ‘that’s what clients are paying us for’…if they could write a great headline via A.I., what value are we bringing?”
But here are four waysI use ChatGPT now:
To transcribe screenshots
This has been a mini-game-changer. Sometimes we get backgrounders that are JPGs or I’ll take a screengrab of something. Instead of having to manually key in the copy in myself, now I just upload the image to ChatGPT and presto — it provides an accurate transcription (at least in my experience so far) that I can cut and paste into a document.
To time out voiceovers or other scripts
Timing out a VO or radio script is one of those skills where there can be a lot of variability in read speed — even among experienced professionals. Part of it, IMO, is that writers WANT their scripts to time out and so read things accordingly. Using ChatGPT takes human error or inclination out of the picture.
To identify trends/themes across a lot of information
Here’s one example of this in action. We recently had a one-hour conference call with a client that we had transcribed by an A.I. tool. I fed that document into ChatGPT and asked it to identify common themes across the 50 pages of transcription. It did, picking 11 themes that emerged from the discussion. Before A.I., this is something a human would have had to read through and do manually…and that’s a lot of text to read and process.
To ensure that copy is “on voice”
This is perhaps the most interesting and “higher-level” use that I’ve started to implement. Nearly all of our clients have an established Brand Voice that maps out the tone and manner of their copy. Sometimes the Brand Voice is something Mint has developed and other times it’s been provided by the client.
But what can happen over time is that writers can lose sight of the Brand Voice. In the focus on writing for a specific project, the Brand Voice recedes into the background. I’ll admit that it’s happened to me on occasion.
Additionally, interpretation of a Brand Voice can vary. One person might think an aspect of the voice is more important than another, etc.
Using ChatGPT to assess how “on voice” a piece of copy is (you have to upload the Brand Voice, of course) has been helpful. I see it as a more objective (see commentary on VO timing above) way to ensure that I’m being faithful to the client’s Brand Voice. I don’t use it all the time, but I have on occasion and I’ve found it very helpful.
How not to use A.I. in 2026 (or like, ever)
I had started writing this piece with a mind to do a “listicle” and have three ways to use A.I. and three ways NOT to use it. But since I’m a human not and not a ‘bot, I’ve shared four “Dos” about how I use ChatGPT now — and I’m only going to share one “Don’t”. But it’s a biggie.
In fact, I’d almost consider it the Prime Directive of artificial intelligence.
And here it is:
NEVER EVER send A.I. output to a client without reviewing it first.
ChatGPT even has a disclaimer in the footer of its page to remind you: “ChatGPT can make mistakes.”
Note that it’s not just black-and-white mistakes you need to look out for. It’s also a matter of taste. Going back to my junior copywriter example from earlier…one of my jobs as a supervisor/mentor to less experienced creatives is selecting the best ideas and lines from their work — and hopefully instilling in them what makes one idea or line better than another.
This very topic was addressed in a March New York Times article, “Is Taste the One Thing A.I. Can’t Replace?”
And I think that is where the value of an experienced creative — whether myself or someone else — comes into play. THAT is what you are paying for, even if we use A.I. to help us in our work. Think of it as a new interpretation of “There’s no accounting for taste.”