Don't let AI turn you into a marketing monster

_{⏩ Forward this email to: a marketer who’s being told to do more with AI}

👋 Hi, I’m Amanda. I’m a fractional head of growth with over a decade of experience working with rapidly-growing B2C startups. I work with lean teams in a few ways:

✅ Fractional: You have an expertise gap on your team you need to fill with someone who can flex into both strategy and execution, for an open-ended time frame.

✅ Advisory: You’re either just starting out or need to validate your existing approach, and need someone with growth expertise to guide you.

If you’d like to explore working together, start here to tell me a bit about what you’re looking for. If you’re not quite ready yet and have growth questions you want to chat about, you can book office hours here.

Is the AI directive kind of a trap?

We’re all feeling the pressure of needing to “use AI.” Maybe it’s been a clear directive at your org, or maybe it’s that unrelenting feeling that everyone else is doing it, way better than you are.

Being told to “do more with AI” today is about as useful as being told to “do more with the internet” in the 90s. You can do incredible work with it, or you can waste equally incredible amounts of time with it, or you can stare at a blank screen with no idea where to start, scratching your head as to how everyone around you seems to be spinning gold with this thing.

To be clear - this is not an anti-AI post. I use it daily, and it’s made parts of my workflows drastically more efficient and some of my thinking meaningfully clearer and sharper. But there’s a trap I keep seeing really smart people fall into that could have some pretty serious implications for marketers in particular.

First of all, we’re not saving any time

The promise of efficiency may be true in pockets, but on the whole, is wildly exaggerated, as HBR found. A workflow that used to take four hours can now be done in 20 minutes, or a research synthesis maybe takes one well-versed person now instead of a team (the reality, even if we see the domino effect there).

But what most AI productivity evangelists aren’t talking about is the dynamic where as soon as you get one thing on your list running more smoothly, there are fifteen more things that could also benefit from the same treatment. The upfront work of building, connecting, prompting, testing, and tweaking these systems is pretty substantial, and the work of automating and improving more and more feels never ending.

If history is any guide, new technologies tend to free us up to…do more work, not give us lots of free time back. Productivity gains are real, but typically translate to fewer people needed, not free time granted to employees.

The marketing problem specifically

Here’s where I’m starting to see this get dicey for marketers.

When you talk to AI, there are no consequences for being sloppy. “Vibe”-whatever-ing, talking out your thoughts, iterating, and tweaking as you go are encouraged, in fact. You can forget something mid conversation and ask to be reminded. You can get your ask wrong four times and work through iterations until the output is what you wanted. You can give bits and pieces of all the context you should give, and still get something workable back. These tools are endlessly patient, and don’t acknowledge or care if you’re wasting their time.

Human communication obviously doesn’t work this way. Your stakeholders notice when you haven’t thought through your ask, your boss notices when you show up unprepared.

But most critically for marketers, your customers will notice - or, fatally, stop caring what you have to say - when it feels like your messaging was produced by someone who doesn’t know them.

If you’re spending hours a day in a frictionless communication environment with AI, your tolerance for the friction that real communication provides will quietly erode. Prompts and loose thoughts that eventually produce output implicitly train you that vague or half-baked is fine, but that approach doesn’t work when you’re marketing to humans.

Marketers know that messaging is the output of a deep understanding of who we’re talking to - what they care about, how they think, and what makes them trust us enough to act. You can’t really iterate your way to that understanding, because you’ll have lost that trust before you even have a chance to earn it. Marketers need to make sure the behaviors we’re learning by doing more with AI (which has significant benefits, to be clear) never extend into our relationship with the people we’re marketing to.

We all know AI is excellent

I’m not suggesting you use AI less. It’s an incredible tool with infinite use cases (which is part of why it can seem so overwhelming at first). Your job is to be thoughtful about what you bring to the start of an AI conversation or task, and where AI’s judgment ends and where yours begins.

The best applications of AI for marketers are the ones downstream of human judgment. It’s excellent at analysis and synthesis of information. Give it a large, messy data set, or a bunch of qualitative customer feedback, or six months of campaign performance, and ask it to surface patterns. It will do this instantly and probably do it better than a human.

It’s also a discerning thought partner (as long as you avoid simply confirming your own beliefs) for working through a strategic problem, thinking creatively when you’re stuck, pressure-testing an angle, or providing fresh eyes when you’ve been staring at the same doc for too long.

There are thousands of other great use cases I could insert here, but none of them should enable marketers to get lazy in their judgment.

It can’t replace your judgment

Your strategic brain, human judgment, and customer understanding shouldn’t be outsourced to even the best AI model.

You could write the world’s most detailed prompt and still not capture every contextual strategic nuance that lives in your head. My argument is that it should stay that way, no matter how advanced these models get, while acknowledging we’re really at the tip of the iceberg with this technology.

I don’t have a crystal ball, but I believe the marketers who are going to be irreplaceable in a few years are the ones who kept doing this level of work themselves, even though it could technically be outsourced to AI. The contextualizing of information to feed to AI, and the filtering of its outputs to determine what gets used, are already emerging as important skills. Don’t let a flawed promise of efficiency be the reason to stop building that irreplaceable judgment.

Was this message forwarded to you? Need a little Growth Therapy? Sign up below 👇

Questions? Comments? Topic requests? Just hit reply ↩