How I Use AI to Cover the Skills I Don't Have Time to Master

After 16 years as a marketing director, I know how marketing works. I know what each channel needs from us, what should be where, and how to keep messaging and visuals cohesive across every piece of a campaign when something is released or announced. I see the big picture and I know how to get there piece by piece.

What I don't have is every specialist skill required to execute every piece myself at the highest level. I can write, but I'm not a copywriter. I know basic HTML and I've hand-coded enough emails over the years to do it again if I had to, but I'm not a developer. Design isn't where I struggle, because I'm a graphic designer and I either do it myself or describe exactly what I want. The places I actually lean on AI are smaller and more specific than people sometimes assume when they hear "AI in marketing."

What I have less of than I'd like is time. I'm spread across teams and projects, and the cost of every "I'd do this myself if I had a free hour" decision is something else not getting done that week. AI fills the gap between knowing exactly what something should be and having the time to make it happen.

The Before

Before AI was useful enough for this kind of work, the gap between knowing what something should be and being able to make it myself was a real cost. If I needed a small marketing email and the writer on my team was tied up with three other beats, my options were to write it myself in something between "okay" and "not quite right," wait until the writer was free, which usually meant the email shipped late or didn't ship at all, or drop it from scope and pretend it wasn't important.

The same thing happened with social copy, banner text, dev journal scaffolds, and any of the dozens of small writing tasks that crop up every week. They don't all need a specialist, but they do need to get done right. I'd end up doing too many of them myself in mediocre shape, or push them out of scope so they didn't happen.

Why This Approach

There's a natural pushback on AI across the marketing industry right now, and I understand it. The concern that AI is going to replace specialist work is real, and the people raising it are raising a fair question. The way I've come to think about it, and the way I talk about it with the people I work with, is that AI is most useful when it fills small gaps, not when it tries to replace whole roles. That distinction is what makes it land or not land for a team.

Part of my job has been helping the people on my team find the small places where AI can save them time without touching the parts of their work they don't want it touching. AI on a campaign isn't going to replace what my designer or my writer does for our hero pieces, and I would not want it to. What it can do is pick up the small one-off pieces that would otherwise either land on my desk in a less-skilled version or fall off the calendar entirely. Those small pieces add up, and recovering them was where this approach paid off most.

There's a second thing that's true and underrated. Every conversation I have with the AI to fill a gap is also a conversation where I'm learning the craft of that gap a little better. The next time a similar task comes up, my knowledge base is a little larger. AI hasn't replaced what I know, it's grown it.

How I Use It

I'm not the best copywriter, but I know effective writing and marketing, what should be said where, and what's important to put where. So I can feed the AI what I want to say and have it assist me with the actual writing, and I'm able to dictate and lead it the same way I would a copywriter, making sure it's written in the way I need it to be written.

The prompting matters a lot. AI is only ever as good as you prompt it to be. You have to give it all the information it needs to be the effective agent you're looking for, the same way you'd brief a contractor or a freelancer.

Here's what I mean. Take an email campaign as the example. You can't just say "write a promotional email using this press release for our gaming title's announcement list." That's the version that produces output you can't use. What you have to say instead is something more like this:

Create an email that will result in high open rates and high click-through rates. The goal is to get people to click the wishlist button first and foremost, and to get them excited about the DLC we're announcing so it creates organic buzz. We don't know who owns the base game and who doesn't on this list, so the copy can't assume the reader knows what the base game is about. Make sure the copy doesn't trip spam filters and uses terminology that helps the email land in their inbox. This is a PC gaming title, so reference the brand voice doc and the MD file for this specific game for any additional context you need.

That's a brief. It's the same brief I'd give a copywriter, just typed into a chat instead of a Slack message. Even with that level of detail you'll go back and forth a couple of times before it's right, but the difference between a one-line prompt and a brief like that is the difference between unusable output and the first draft of something I can finish.

The other place AI has helped is platform-specific social copy. Algorithms on every platform are constantly shifting, and AI is decent at catching the patterns of what's working on a given platform that week and shaping copy for it. That's not something I'd have the time to track manually, and it's not really any one person's job, so it sat in the gaps until the AI started filling it.

The Outcome

I'm shipping more and shipping faster, and the small one-off pieces are landing closer to specialist quality than they would if I were doing them alone with no help. The pieces that used to fall off the calendar entirely because nobody had the bandwidth to take them are getting done.

What's also true is that I haven't replaced anyone. AI doesn't do my writer's job. AI makes me good enough at the small one-off pieces of writing that would otherwise pile up on the writer's plate. The writer still does the work that needs a writer, and the AI lets me handle the rest without dropping any of it or bottlenecking the team.

The other thing that's happened, the thing I didn't expect, is that my own skills have widened. After enough back-and-forth with AI on the kinds of writing I used to outsource entirely, I prompt better, I direct better, and I notice things in drafts I wouldn't have noticed before. The AI conversations have functioned as a kind of ongoing tutorial in the crafts I sit adjacent to.

The honest summary is this: I know what I want things to say or be or look like, and after 16 years in this work I have most of the skills I need to make most of it happen. What I don't always have is the time, and for the very specific pieces that aren't quite in my wheelhouse, AI fills the gap so the work still ships and ships well.

What's Still Messy

The industry pushback on AI isn't all unreasonable.

Specialists are right to be cautious about a tool that some companies are clearly trying to use as a replacement rather than an enabler. The framing that helps most, in my experience, is being clear and consistent that the AI is filling gaps, not taking jobs, and that the standard for what ships is the same whether a specialist did it or I did it with AI's help.

The other messy part is that this only works if you already know enough about the work to direct it. If I didn't already understand effective email marketing, I couldn't tell whether the AI's draft was good or not. AI is a multiplier on the skill and judgment you already have, not a substitute for not having any, which is part of why I think specialists in marketing aren't going anywhere. The people who can direct AI well are the people who already know what good looks like.

The last messy part is prompt discipline. When I'm tired or rushed I write a short prompt and end up unhappy with the output, and the failure is mine, not the AI's. The brief is the work, and if I write the brief well the rest goes fast.