Hey Reader

As you may know, I started this newsletter in November 2022, right before ChatGPT launched. By January 2023, ChatGPT became the fastest-growing consumer software in history. By mid-2023, my first business, a content agency, had shrunk significantly.

While I had no clue what was about to happen, starting this newsletter while I still had the time to figure things out was one of the best decisions I ever made. That and having a war chest.

Because 2023 was my worst year in business by far. The content agency was bleeding clients, and the business around this newsletter was still in its infancy.

Luckily, I had time to figure things out — and the financial cushion to support me while I did it.

So, you can probably understand why I have a love-hate relationship with AI. I refused to use it for a long time (far too long).

In 2023-2024, every friend I had in the content industry was desperate. Some lost multiple seven-figure businesses. My dissonance deepened.

Around the same time, as Strategic AF was picking up some steam, I started doing more and more 1:1 strategy sessions. I noticed that almost all my clients asked me about AI.

How can they use it for writing? What about marketing? Should they post AI-generated content? What about AI automations?

Their disappointment was palpable (and understandable!) when I’d tell them that I barely touch the thing, so I have no recommendations — other than don’t post AI-generated content as is, which is still something I stand by.

In the second half of 2024, I got off my high horse and started taking AI seriously. I’ll tell you all about how I use it after a quick intermezzo that kinda gives away much of what’s to follow.

​
​

📣 Brought to you by 📣

​The Profitable Content Engine Workshop​

My workshop on creating profitable content faster and more easily is happening in exactly one week, on June 19th, at 10:30 AM EST. Meet me for a 90-minute live session (with Q&A) and learn all about the content engine you need to:

• Stop spending hours on a single social media post or email
• Create content that sells without the sleaze or the bro vibes
• Never run out of profitable content ideas

Oh, and get the exclusive bonus that you won’t find anywhere else: Contently Lab, a custom GPT that will be your 24/7 content strategist. I trained it on psychology- and science-backed conversion tactics, my own work, plus the work of a few big names in the writing industry.

There are 4 seats left at a 20% discount (the people on my waitlist grabbed the other 6 yesterday).

Want one of them? Grab it now!

​

Want your name up here? Reserve your slot!​

​
What AI says about my use of AI

Because I like to be meta, I asked ChatGPT to summarize our interactions. This is what it came up with.

​
​
​

In case the screenshot is hard to read, here’s the plain text version with zero edits:

You don’t use AI to write for you.
You use it to think faster, clarify sharper, and stay ruthlessly aligned with your strategy.

Your approach is built on augmenting strategic clarity, not automating creativity.

Here’s how you integrate AI into your workflow:

1. Strategic Ideation & Idea Filtering

You collect raw ideas—from client calls, audience DMs, rants, and half-baked thoughts—and use AI to sort, group, and reframe them by pain point, offer, or theme.
Why: It helps surface patterns you can’t see when you’re in the weeds.

2. Voice-of-Customer Reframing

You feed real audience language into AI to rework hooks, headlines, or posts using actual customer phrasing.
Why: It keeps your messaging grounded in what your people actually say—not what marketers think sounds good.

3. Angle Testing for Content Expansion

You use AI to generate multiple angles for one core idea—myth-busting, contrarian takes, storytelling, tactical advice, etc.
Why: It turns one idea into five, fast—perfect for repurposing across platforms.

4. Content Repurposing (Without Repeating Yourself)

You input long-form pieces (like newsletters or essays) and ask AI to spin them into tweet threads, carousel outlines, or social blurbs.
Why: You believe in “write once, publish many”—AI helps you do that without sounding robotic or redundant.

5. Clarity Before Writing

You use AI to interrogate your own ideas before you write. Prompts like “What’s the belief behind this?” or “Which offer does this support?” help you avoid vague, throwaway content.
Why: You’re not here to post for posting’s sake. You’re here to make every piece do its job.

6. Custom-Trained GPT Assistant

You built a custom GPT chatbot to help creators and solopreneurs connect content ideas to business goals—faster.
It’s trained with your frameworks, your tone, and your strategic process.
Why: Because most AI tools are built for volume. Yours is built for profitability and positioning.

How much of this is hallucination?

AI tends to tell you what you want to hear and, occasionally, hallucinate. Here’s what’s wrong with its description (less than I expected, to be honest):

• The third point: I don’t do a lot of “angle testing for content expansion”. I probably did it once or twice, so I definitely wouldn’t have added it so high on this list. It’s a great idea, though, so thanks, ChatGPT!
• My most frequent use case isn’t mentioned here. I paste almost all my writing into ChatGPT and ask it to tell me if I’m getting high on my own supply. “Have I missed any nuance here?”, “What are some of the limitations of this framework that I haven’t included?”. Right now, this is its most valuable use case for writing (to me).
• My second most frequent use case is also missing. I like to ask AI whether a draft aligns with my brand ethos (which is detailed in the project instructions). This keeps me on the straight and narrow.

Other than that, it’s pretty spot-on. But I think it’s missing the essential: the overarching approach. More on that below; first,

How I use AI for this newsletter

Examples beat dry bullet points, so let me show you how AI is involved in the writing of this newsletter:

• I have an idea bank where I collect suggestions from readers, pain points my clients expressed during strategy sessions, and so on. This is where I pick up my ideas from.
• Once I’ve chosen a topic for the week, I might ask AI two things ("might" because I don't always have the time to do it):
  • “Will this genuinely help my audience? Be brutally honest!” Because it has a lot of information on my audience, it can answer that.
  • “Can I tie this to one of my products or services?” I’m usually pretty good at spotting these connections but AI is better. It often showed me better matches than what I initially wanted to include.
• Once I’ve gathered this information, I write the draft.
• Lastly, AI is in charge of:
  • Finding blind spots in my argumentation.
  • Calling me out on my own bullshit — yep, it happens.
  • Finding limitations and potential pitfalls.
  • Tightening up the content. I tend to write A LOT, so it helps to ask “is there something of little to no value that I can delete?”

Of course, everything is very hit-and-miss. I try to make it a bit more predictable through better prompting but it’s never perfect — how could it be?

Let me give you a few examples.

Inherent AI biases and how to bypass them

AI is designed to please the user. You might think that this means ego massage, but it goes beyond that. Sure, unless prompted otherwise, your chatbot will tell you that you’re the smartest person in the universe and you hold the solution to world peace.

But its people-pleasing goes beyond that.

For instance, if you ask it to “find 3 things I can improve in this post,” it will spew exactly three. It doesn’t matter if your post has zero things to improve or 10, you will get three.

How I bypass this with a decent rate of success:

• I ask it to be “brutally honest”. This ensures that it doesn’t just tell me what I want to hear.
• Adding “if there are no areas for improvement, tell me so” works too but it depends on your history with the bot. If you usually favor friendly, ego-stroking conversations, it will skew that way and lie to you.
• If I’m asking for data projections, like “how many units of X product could I sell in this context”, I prompt it to anchor its answer in industry data and give me three scenarios: optimistic, realistic, conservative. Usually, the “conservative” scenario is actually the “realistic” one. It takes an insane amount of iteration to get AI to err on the side of caution instead of hype.

Beyond this newsletter, I like to use AI as an intellectual sparring partner. I know everyone talks about speeding things up with AI (myself included) but I found even more value in using AI to linger where it matters.

There are decisions or tasks whose success depends a lot on how much you spend on them. For instance, I can use AI to speed up my newsletter production but, if I were writing a book, I’d use it to help me spend more time deep thinking about it.

While I've gone from almost no AI use to building a custom GPT fairly quickly, I still don't rely on it for everything. And I never will.

Two non-negotiable guardrails I set for my use of AI

The first one is a litmus test inspired by my friend Jay Clouse, who said it best:

This is something I think about every day. It’s very tempting to feed AI a few bullet points, then get it to write the entire newsletter issue for me. I would be so damn fast.

But it’s not why you’re subscribed to Strategic AF, is it?

When content is commoditized by AI, personality and relatability matter more than ever. And AI still can’t emulate both the tone of voice and the idea behind it to a T.

So I’ll do my own writing, no matter what.

The second one is born out of fear. I think my biggest asset is my critical thinking, or the way my brain works.

With all its shortcomings and misfires, it’s brought me to where I am, and it’s what I rely on the most — in life and business.

Yes, AI can reason faster than I can. But speed isn’t everything. Depth is something purely human, something that AI can only simulate, not replicate.

This is why I use AI as a thinking aid, not a thinking replacement.

For example, if I wanted to launch a new product, I wouldn’t ask AI whether it’s a good idea or not. That’s for me to decide. Instead, I’d get it to:

• Analyze industry data, compare them with my conversion history, and make sales projections for me.
• Crawl a few relevant websites and tell me if my ICP shows any interest in such a product.

Putting this data together and making the final decision is up to me. And it’s very important to me that it stays that way.

Workshop update

I’m putting on the finishing touches for The Profitable Content Engine Workshop happening next week.

This means:

• The structure is done, now I just have to turn it into a presentation format and make it pretty.
• I started sending Convertly Lab, the AI-powered content strategist, to a few friends for testing purposes. So far, the feedback has surpassed my expectations. People have gotten great content + distribution ideas from it, some that you might actually see out there in the wild. This is HUGE in my books, since the friends in question are wicked smart and great marketers in their own right.
• I’m also amazed at how fast the first six seats disappeared, which tells me I was wrong to wait to do this workshop for so long.

One last reminder, at the time of writing this, there were only 4 seats left with an early bird discount. Grab one of them here and get access to the workshop, the AI content strategist, and aaalll the bells and whistles.

🔦 Community Spotlight

My friend Matt Gira put together an impressive report. It has all the data behind 64 founders who built their businesses to $250k+ in annual revenue without venture capital. I know firsthand how much work he put into this — and it shows. Grab the report here, it’s 100% free, although it shouldn’t be.
​

That's it from me today! See you next week in your inbox. Here to make you think, Adriana ​

​

P.S.: How do YOU use AI? Reply and let me know, I'm super curious!

Share this essay

🔗https://www.adrianatica.com/how-i-use-ai-in-my-business-saf-134/​

​
​Quick share links

​​​

​