Plus: get a sneak peek into a Copy School Pro training
If I’m known for anything, it’s swiping.
…I taught this Copyhackers tutorial on swiping emails from your inbox to inspire better copy.
…My podcast is literally named Email Swipes.
…And I don’t let my VA delete a single email from our shared inbox, lest I lose potential inspiration.
So even though LinkedIn serves me AI tips with every scroll and I’m subscribed to every trendy AI newsletter (including an AI lab WhatsApp group!), the ideas that actually make it into my workflow don’t come from feeds or lists.
I experiment with AI the same way I experiment with anything:
By swiping ideas from peers and mentors.
Here are the top 10 ways my network is using AI right now – plus takeaways from Copy School Pro’s month of AI
- Get smoother first drafts
- Play devil’s advocate with your copy
- Coach better client revision requests
- Automate e-v-e-r-y-t-h-i-n-g (sans consultant or subscription fees!)
- Engage students at scale
- Tackle whatever challenge you’re facing
- Bribe your way through procrastination
- Find meatier content ideas
- Actually get your content out there
- Capture your ideas anywhere – and access them in seconds
Eden, Jo, Abi, and Sabine have been using AI while writing copy to:
1. Get smoother first drafts
Garbage in, garbage out applies to your AI copy outputs, just as much as it applies to your fruit pies – and it’s why my 2022 AI drafts nearly had me throwing in the towel on the whole “AI as a copywriter” thing.
It’s also why Eden Bidani’s prompting hack is genius.
Instead of overloading a first prompt with context and docs — or dumping everything into a Claude project and hoping it gets it — Eden flips the process:
I give Claude a prompt – and then I ask it to ask me any clarifying questions it needs to get its job done.
The depth of questions it comes back with is often surprising, and after answering them, the first draft it produces ends up stronger than when I front-load the prompt with every scrap of info.
The back-and-forth takes a couple more minutes of work, of course, but having tried it again and again, this way ends up producing much stronger first drafts.
How I’m using this use case:
My number one takeaway from Copyschool Pro’s AI-focused month?
Don’t expect perfection from your bots.
Sure, they are computer programs and should be infallible – but they’re not.
I’ve seen myself – and heard from others – that a few prompts in, they’ll ignore all the docs you painstakingly uploaded and need you to nudge them back on track.
Starting the chats off by asking clarifying questions not only corrects for this – it also mentally puts me in a training space versus an order-giving space.
2. Play devil’s advocate with your copy
The Seven Sweeps is my favorite Tutorial Tuesday series – and it’s one that I still use for every project, eight years later.
And before custom GPTs were a thing, Copyhackers published a set of prompts to help AI run your sweeps.
Now Jo is taking AI sweeps even further:
She’s using AI to challenge her own assumptions — specifically in the manuscript for her upcoming book, The Copyselling System.
“I fed the first chapter of my manuscript into a Claude project, along with my back-cover copy, a description of my ideal reader and a few other parts of my book project. Then I asked Claude to read the files I’d added, assume the persona of my ideal reader and read my first chapter. After that, I asked Claude to play devil’s advocate and, as my ideal reader, tear apart the first chapter, one point at a time.
“It. Was. Amazing. Claude was so good at challenging my assumptions and pointing out flaws in my argument that I rewrote almost half of Chapter One. There actually came a point when I had to stop Claude because “he” was pinpointing so many little things that I was about to throw the whole book out.”
– Joanna Wiebe, during our Using AI to Challenge Assumptions CSP training
How I’m using this use case:
My checklist of sweeps officially has an eighth checkbox.
3. Coach better client revision requests
Back in October 2022, the worst revision requester was your client’s wife. Now? It’s their favorite LLM.
See, ChatGPT never runs out of suggestions.
There’s always something to nitpick — and it often sounds convincingly smart.
After too many AI-driven revision requests from clients who used to respect her process, Abi Prendergast brought this challenge to a Copyschool Pro group call.
Now she does this:
As part of our onboarding process, I share how we’re going to be using AI, that it’s part of our internal feedback process – plugging in the messaging guide, uploading documents for tone, and matching prompts against our own internal best practices.
We request feedback based on consistency with all of this context. We’re not just asking it to give us feedback based on everything that was on the internet ever.
That’s not productive, and it will always have something to say.
Pre-strategy call, we also plug the final into ChatGPT and just see what it pulls up without any prompting – and then we can preempt feedback.
We can say: if you put this into ChatGPT, it might pull up these things – and here’s why we did it differently.
So basically we’re just preempting feedback more and making sure the client really feels confident that we know what we’re doing, that we’re using AI effectively, and that them plugging into AI isn’t really adding anything new to the conversation.
That there’s nothing that their ChatGPT is going to bring up that we haven’t thought of and already integrated thoughtfully.
Sabine Harnau also suggested we give our clients coaching on how to use AI to give constructive feedback.
If they’re gonna use it, let’s make sure it’s on our terms.
In the past, when handing over copy and wireframes for feedback, I always provided instructions. Expectation setting for low-fidelity wireframes, to avoid personal taste and instead stick to accuracy, legal implications, and brand alignment, etc., etc.
With AI, I’ve updated this guidance to include a note on prompting and context engineering:
If you’d like to run this draft through AI, remember that context and prompting are key. Always add the full briefing and research report to the conversation. Ask your AI tool to review the draft in line with the briefing.
Just as with a human review, ask it to flag any issues with accuracy, legal implications and brand.
How I’m using this use case:
Just like Sabine, Jo taught me to train my clients not to tell me they “like” my copy.
Now I need to train my clients how to provide high-quality AI-driven feedback.
And I’m experimenting with a custom GPT that challenges clients’ feedback requests.
Nicely, of course.
Michal, Chava, Katie, and I are using AI to work smarter:
4. Automate e-v-e-r-y-t-h-i-n-g (sans consultant or subscription fees!)
I once booked a consult with an automation specialist to streamline operations.
Maybe she didn’t want to give away too much because she wanted me to buy her full package, but I walked away with no new ideas.
No longer.
ChatGPT – even without agents – is an automation wizard.
And Michal Eisik’s team is all in:
We recently launched a marketing personality quiz in Jotform, using ChatGPT to help map out the scoring and personas. At first, we used Zapier to send quiz results into (Convert)Kit — but with so many tasks firing, the costs were climbing fast.
So we switched gears and built the integration with Google Apps Script (GAS), a free tool inside Google Workspace that lets you automate workflows across Sheets, Gmail, Drive, and even external apps.
Now, every quiz submission flows straight into Kit, tagged with the right persona and dropped into the right sequence.
Instead of having to use an API or hire an automation consultant, ChatGPT broke it all down into clear, step-by-step instructions we could understand – turning what could’ve been a costly, complex setup into a streamlined, budget-friendly workflow.
Not running a quiz?
Her Google Apps Script + GPT hack can be adapted for anything in your workflow.
Here are a few prompt ideas she suggests:
- “Write a Google Apps Script that automatically creates a Drive folder (‘Assets,’ ‘WIP,’ ‘Finals’) when I add a new client to my spreadsheet.”
- “Make a Google Apps Script that sends a reminder email if an invoice row in my spreadsheet is still marked ‘unpaid’ after 14 days.”
- “Write a Google Apps Script that saves all Gmail attachments with the label ‘Client Feedback’ into a dedicated Drive folder.”
- “Create a Google Apps Script that logs all my client meetings from Google Calendar into a time-tracking sheet.”
Sometimes, all it takes to trigger automations like sending a personalized happy holidays email to all past clients is simply labeling an email in Gmail!
How I’m using this use case:
I’m requesting video testimonials in anticipation of my Black Friday sale.
I want to send these requests manually so they don’t land in Promotions – and so I can customize them based on whether the customer has already sent a written testimonial.
This would have been out of our no-code skillset – or I would’ve just resigned myself to sending with Kit – but now it’s just a day’s work to set it up.
No GitHub combing or workarounds necessary.
5. Engage students at scale
Chava Shapiro is my AI muse – her AI for Creatives workshop was the tipping point that turned me from an AI skeptic who used the tools so infrequently I didn’t even sign in – to someone who’s got ChatGPT and Claude open in one (or ten) tabs at any given moment.
So it’s no wonder that the use case I swiped from her is one I haven’t seen elsewhere – and solves a thorny problem that’s plagued course creators since Jeff Walker published Launch:
Low completion rates.
I was privileged to be a guest mentor for Chava’s signature course, The Creative CEO Academy – and I watched her churn out custom GPTs that basically do the homework for her students.
So instead of them dropping off halfway through module 2, they’re pulled forward by smart prompts that move them out of the weeds.
When I find people getting stuck – when I’m getting a lot of questions or students asking me for feedback, I can train a ChatGPT to be an almost stand-in.
I train it the way that I would answer the question or the way that I would critique their work.
Now, when students are stuck, they can first try seeing if the GPT will help them – and oftentimes it’s even better than me because it gives so much more detail than I would necessarily have time to give each student.
(This BTW is a testament to how well her GPTs work, because Chava gives her heart and soul to her students. If the GPTs are doing a better job than her, then that’s saying something.)
How I’m using this use case:
I’m bundling replays of my course, Email Mastery, into my Black Friday sale – and especially because there’s no live component, custom GPTs are the perfect way to support my new students async.
6. Tackle whatever challenge you’re facing
Have you ever pulled up #validation or #help-and-support in your favorite Slack group, typed in a lengthy, rant-ish explanation of whatever challenge you’re facing, and then realized you have the answer, just by hashing it all out to no one?
If that’s you, wait until you try using AI as a thought partner.
Jo introduced this idea in the Copyschool Pro training, Using AI as Your Clarifying Thought Partner — and it instantly became one of our most underrated workflows.
Here’s how Katie Peacock has been binging it:
When I’m not sure which task to focus on next, I brain dump to Claude, and then ask it to rank the tasks according to their priority based on its knowledge of my goals.
(I uploaded a doc with my big goals for the year in January, so it knows what I’m working towards.)
With that list, I have a clear starting point.
Then, I ask Claude to act as a strategic thought partner and interview me one question at a time, like we learned in the training.
From there, it doesn’t take me long to get into action – working on the task that’s most aligned with my big picture goals.
How I’m using this use case:
I’m using this one for everything from restructuring my podcast episodes, to hashing out the best content platform to focus on – and especially when working through thorny people management issues.
Having AI ask me one question at a time is the perfect mosh between coach and sounding board.
7. Bribe your way through procrastination
I snapped up Anne-Laure Le Cunff’s book, Tiny Experiments, after reading the email she sent to presell the book.
I loved the email, but the book was even better.
In one of my favorite chapters, Procrastination is not the Enemy, Anne-Laure prompts us to view procrastination as a signal that a task isn’t appropriate, exciting, or doable — and introduces her Triple Check process to help diagnose it.
My go-to reason for procrastination?
When I just don’t want to do said task.
For this diagnosis, Anne-Laure suggests:
You can use what behavioral scientists call a pairing method to help you get started.
If the task is sound but feels dreary, pair it with an enjoyable activity.
You can catch up on overdue emails from your favorite coffee shop, do your taxes while listening to your favorite band, or turn the task into a game by creating rewards for each completed chunk.
So when I notice myself drifting, I bribe myself. With these.
How I’m adapting this into an AI use case:
I’ve trained myself to use the pairing method so well that I crave those chocolates as soon as I open my accounting software.
The problem is that there are no Trader Joe’s locally. So my supply is limited to how recently we’ve been near one – and how much suitcase allowance I can beg from visiting friends and family.
Thankfully, my hands-down favorite AI use case is as my personal food scientist. And it nailed a copycat recipe for my procrastination-busting chocolates.
You’re welcome.
Gia and Claire, Ross, and Jeremy are using AI to develop better content:
8. Find meatier content ideas
A few weeks ago, Forget the Funnel sent out takeaways from an AI-moderated study they ran this summer.
An AI-moderated study?
I stopped at that too – completely intrigued – and then read on.
The findings were absolutely fascinating – almost enough to distract me from the “AI moderated study” thing.
But I followed up, fear not. I was curious about the “tradeoffs to consider.”
Gia didn’t disappoint. This is very exciting stuff.
AI-moderated studies sit in an interesting middle ground.
If a standard open-ended survey gets you to about a “2” in quality, and a 30-minute customer interview gets you to a “9,” then AI-moderated studies are around a “4.”
They’re much richer than a traditional survey, but they’ll never match the depth of a real conversation. They’re directional, and will give you a broad-based understanding, but you won’t get the depth and context to make meatier biz decisions.
The key is knowing when to use which research method.
If you’re making high-stakes decisions, like tackling retention issues, repositioning, or improving adoption, you need real interviews.
But if your goal is to get a broad understanding of a large audience, for say content ideas, AI-moderated studies can be wildly more valuable than the standard surveys we’ve leaned on for the last 15 years.
How I’m using this use case:
I hope Claire and Gia follow that trend of marketers-turned-SaaS-founders – because they’re developing their own AI-moderated study tool to improve on the off-the-shelf options they’ve been using, and I want in!
Even if I can’t access theirs, I’ll definitely be exploring the existing options for clients’ customer feedback sprints.
9. Actually get your content out there
One of the most action-packed Copyschool Pro trainings was by Ross Simmonds on repurposing content.
There’s no one I can think of who speaks to this more frequently (or eloquently) than Ross.
(I mean, the guy literally wrote the book on content distribution.) I love how he puts it:
Everybody forgets that content marketing is a two-word industry. Like, it’s not just content, content, content. It’s about marketing the content.
He’s been preaching this idea, but not enough people have been listening – so he built a nifty tool to remove all friction:
Put all of your energy and time into creating something amazing – and then let AI help you repurpose and distribute it. I hope that this too evaporates all of the excuses that so many people make around “I don’t have time to create and distribute my content.” Now it should be easier.
How I’m using this use case:
I know I’m creating a distribution.ai account as soon as my next podcast season is live. Prepare to see posts about Email Swipes absolutely everywhere.
10. Capture your ideas anywhere – and access them in seconds
Ask Jeremy Enns a question, and he’ll pull up a relevant Notion or web page in seconds.
Literally seconds – I’ve counted.
His organizational superpower is something to behold.
Lately, though, he’s been experimenting with longhand–writing podcast episode scripts in his notebooks, sketching out cover art, or mind-mapping new products.
I was excited to hear how he keeps this hard copy work accessible – because that’s always been my organizational sticking point.
I think better on paper – where I can highlight, draw arrows, and add callouts. I have ideas when I’m in the middle of bathing my two-year-old and don’t cherish the idea of pulling out a phone with sudsy fingers.
And so I can be more present, I trade my smartphone for a dumbphone when I leave the house.
All this leads to J.K. Rowling-style ideas:
Scribbles on spare napkins or receipts, scattered in notebooks located anywhere from my diaper bag, laptop bag, office desk, or tacked onto the various shelves of my front hallway command center.
So how does Jeremy keep his hard copy ideas organized?
Simple: he digitizes them.
I found it to be super helpful to only have one place to put ideas.
I used to have notebooks and phone notes and other places – and I realized I need one place to put everything, one place to go find them. If I have some ideas in notebooks and some in this app and some in that app, then when I want to go find an idea, there’s friction.
And then the system is less useful. And then I write down fewer ideas because it’s not useful. So making it as easy as possible, with no thinking about where to put this – removing all friction.
To digitize, I usually just dictate a voice note really quick or transcribe it into Notion and give a really blunt version of the idea. And as a backup, I take pictures of all the drafts and scribbles, and those live in Google Photos.
How I’m adapting this into an AI use case:
One of AI’s best strengths is formatting. Instead of dictating, transcribing, and then copying everything into my Notion manually, I do it in AI and tell it how I format it for that particular Notion page.
And when something is too long for dictation, a quick camera click is the next best option.
The only downside?
I need to write neatly enough for Claude to decipher my words. Not an easy task for a doctor’s daughter.
(BTW, if you think you shouldn’t follow Jeremy because he teaches podcast marketing and you don’t have a podcast, Jeremy is a brilliantly intuitive marketer – and deconstructing his business is a masterclass in its own right.)
What I want to explore next: Agents!
Agents are literally the one thing I was waiting for AI to do, and yet no one I know – even the Copyschool Pro AI wizard whose posts need repeated reading to sink in – is using them (or at least talking about using them).
Maybe it’s the security/privacy thing. Maybe they’re too expensive. Maybe it’s that everyone is already so automated they don’t need agents.
I don’t know, it sounds genius to me.
Do I need to forge my own path? Spare me, please – and tell me how you’re using them.
Conversation