Full training
Everything We Covered in Our Late-Night Marketing Session | September 25, 2025
This was a proper deep dive into the actual workflows we're using right now to run our businesses more efficiently.
We identified 11 specific ways you can use AI to work smarter, save hours every week, and make more money.
Let me break it down for you.
1. Turn Every Conversation Into Content Gold
Here's what I do, and it's stupidly simple: I record everything. We’re recording this session for example, but I also let Claude turn our WhatsApp threads into newsletters.
My exact process:
Record the conversation (this PLAUD device I'm wearing does it automatically - records, transcribes, summarizes, even creates mind maps)
For iPhone users: go to your Voice Memos, hit "generate transcript," copy it, and Bluetooth it to your MacBook
For WhatsApp conversations: export the chat as a .txt file from your phone
Upload to Claude (my go-to for writing) or Gemini
Tomorrow, I'm going to upload tonight's transcript and ask it to recap the key conclusions. Then I'll share that in our newsletter, the inner circle, and drop it in Notion as training material.
Pro tip: Claude is much better than GPT for actual writing. GPT gets too "bullet-pointy" and robotic. It's brilliant for outlines and reasoning, but when you want something that sounds human, Claude wins every time.
2. Scale Your Knowledge Into Training Programs
This one's a game-changer, especially if you're in the education space. Olly and I have cracked the code on turning expert knowledge into comprehensive training programs for his mentorship programme.
The workflow that's working:
Olly sends me 2-10 minute voice messages explaining his vision for a training (Dream 100, welcome sequences, high-ticket coaching, whatever)
I transcribe it and feed it into GPT with massive, detailed prompts
GPT creates the structure: why this training matters, mindset shifts needed, step-by-step implementation
Olly reviews and approves the outline
I take that approved outline to Gemini and have it write the actual modules, one by one
We add real case studies and examples from our world - this is crucial because generic content is useless
We have a detailed SOP for this process, with well-thought-out prompts.
The magic happens because GPT is incredible at creating logical structures, while Gemini excels at turning those structures into actual readable content. But the human touch - adding our specific examples and voice - that's what makes it valuable.
3. Email Marketing That Actually Converts
Forget manually writing every email. I've created custom GPTs that know our voice, our templates, our entire system.
How it works:
- I've fed these GPTs all our training materials on welcome sequences, flash sales, nurture campaigns
- When I need an email sequence, I give it the sales page, customer survey insights, and avatar details
It summarizes everything, asks for confirmation, then writes the entire sequence following our proven templates
Are they perfect? No. But they're getting better every time we feed them more context. And they save hours of work.
4. Analytics That Actually Make Sense
Last week Arieh did a big shout-out for Teacher AI. We needed to know if it was actually working. Instead of spending hours in Stripe's terrible reporting system, I just exported all the data since his shout-out and uploaded it to GPT.
What I asked:
Our usual baseline is about 140 emails per day - how much have leads increased?
What are our conversion rates looking like?
How many are on free trials vs. actually converting?
Based on our patterns, how many will likely convert next week?
How do we stack up against industry standards?
Five minutes later, I had a comprehensive report that would have taken me hours to compile manually, and we know the ROI on the campaign.
For Teachable users: Export your sales reports (the transactions one with timestamps). You can ask GPT things like "How many people who buy Course A end up buying Course B within a year?" The insights are incredible.
5. Building Software
With AI, we can now finally solve the ‘ not enough speaking practice problem’. That’s why we’re building Teacher AI.
Our developers are using this tool called Cursor, and their productivity has jumped 30-40%. For $50-100 a month, that's a no-brainer ROI.
But here's what's really interesting - you can "vibe-code" with ChatGPT even if you can't code at all. Can built these interactive lead magnets for his language courses just by describing what he wanted.
The process:
- Describe exactly what you want the game or quiz to do
- Keep refining: "not this, not that, make it mobile-optimized"
- Preview it in ChatGPT (though always test on your actual site)
- Drop the code into a WordPress block or iframe
It's not perfect, but it's functional. And it beats paying a developer thousands for simple interactive tools.
6. Design Faster, Communicate Better
I needed an app icon for Vanessa's white-label project. Her website had this logo with overlapping text. Instead of asking her for a clean version or bothering our designer, I just asked GPT to cut out the overlap and add a yellow background. Thirty seconds later, done.
For bigger design projects, I describe what I want - features to highlight, exact text, overall vibe. GPT gives me a rough draft that I can send to our designer. He knows exactly what we're after, and execution is faster.
Bonus: GPT suggested adding a live word counter during recordings. We thought it was brilliant, our designer made it look good, and now it's going into our app. Sometimes the AI comes up with ideas we never would have thought of.
7. Customer Service That Scales
Leo showed us his approach: when he gets a complex email thread, he exports it as a PDF and uploads it to his custom GPT. It understands the full context and drafts a proper response.
For higher volume, Lette built an automation with Zapier and OpenAI Assistants. Emails come in, AI drafts responses, they appear as drafts in Gmail for human review before sending.
Key principle: Always keep a human in the loop. The AI handles the heavy lifting, but you maintain quality control and personal touch.
8. Content Strategy That Actually Works
When I was thinking about creating video content for Creator Empires, here's how I approached it:
Define your exact avatar - who specifically are you serving?
Make sure it aligns with your unique expertise and career
Find a niche so specific that you're automatically the best at it - "category of one"
Ask GPT to validate the fit and suggest content buckets
Generate video ideas, outlines, and even intro scripts
The strategy isn't to compete in crowded spaces. It's to own your tiny corner so completely that YouTube's algorithm brings you the right people. AI can help you with video ideas as long as you tell it what avatar you’re creating content for, what you’re unique skills and experiences are.
9. Audio and Video Production Made Simple
Voice generation:
11 Labs has voices from real people in their library - high quality, natural breathing and intonation. Great for language education channels that need voice samples.
Google AI Studio can create dialogues with emotional controls (more upbeat, more sad)
Audio assets:
- Sound effects: 11 Labs can generate anything (moving chairs, animals, whatever)
- Music: Suno (S-U-N-O) for background music without licensing headaches
- B-roll: Various tools for visual content
Strategy: Use AI for internal course materials and shorter clips. For longer, important content, consider mixing AI with real voices to future-proof your content.
10. Multilingual Content Without the Headache
The workflow for subtitles that actually makes sense:
- Match your original language audio to timestamps first
- Translate the text using GPT
- Upload the translated subtitles
GPT is incredibly good at translation, and you can even ask it to handle the technical subtitle formatting. Much faster than doing it manually or hiring translators for every language.
11. AI as Your Business Coach and Decision Partner
This is where AI becomes your strategic thinking partner for complex situations:
Business coaching and decisions: The more you feed AI information about your business, the better it gets at helping you think through strategic decisions. It knows your context, challenges, and goals. Use it to work through tough choices, evaluate opportunities, or think through problems.
Difficult conversations: Whether it's managing staff, handling customer disputes, or navigating personal situations, AI can help you structure these conversations. Gareth mentioned using it to script difficult conversations and think through different approaches before having them.
Legal and contracts: Upload existing contracts, ask GPT to summarize key terms, identify risks, suggest what to negotiate. (Always have a lawyer review, but this gives you a starting point.)
Data mining: I had a court case where the other party made claims. I uploaded 10 years of WhatsApp conversations and asked GPT to find evidence contradicting their claims. One minute later, I had everything I needed with translations and reasoning.
Health and wellness: Input your age, gender, diet restrictions, and goals. GPT can suggest supplement plans, immunity strategies for events, even meal planning from whatever's in your kitchen.
Productivity: Describe your weekly tasks and energy patterns. AI can suggest the optimal schedule (though sticking to it is still on you).
The key insight here: AI becomes more valuable as a thinking partner the more context you give it about your specific situation. It's not just about automating tasks - it's about having a sounding board that knows your business inside and out.
The Tools That Actually Matter
AI Models and When to Use Them
Claude: Writing, newsletters, anything that needs to sound human
GPT-4: Analysis, outlines, reasoning, structured thinking
Gemini: Good for writing, handles large files well