In partnership with
The Exact Tools & Prompts I Use to Write Every Newsletter Issue
I used to spend 4–5 hours on a single newsletter issue.
Research, writing, editing, second-guessing everything — it was exhausting. Now I'm done in under 90 minutes. Not because I got faster. Because I finally figured out how to use AI the right way.
Here's my exact process. Steal it.
HubSpot AEO
Picture this. A buyer opens ChatGPT and asks for a recommendation in your category. Your competitor's name comes up. Yours doesn't. And that buyer never makes it to your website.
That's happening right now in markets everywhere. And most teams don't know it's happening because it never shows up in their analytics.
HubSpot AEO shows you exactly where your brand stands in AI search, where competitors are getting recommended instead of you, and tells you specifically what to fix. No expertise needed.
Try it free for 28 days. Just $50 a month after.
Step 1: Find a topic angle that isn't boring
Generic topics kill newsletters. "How to use ChatGPT" — nobody needs another one of those.
Before I write anything, I open Claude or Perplexity and run this prompt:
❝
"I write an AI productivity newsletter. Give me 10 specific newsletter angles about [TOPIC] that are counterintuitive, practical, and haven't been overdone. Focus on daily workflow use cases, not theory."
I pick ONE angle. Not ten. One. The more specific, the better. "5 ChatGPT prompts for writing cold emails" beats "AI for communication" every single time.
Step 2: Build the outline before writing a single word
This is the step most people skip — and then wonder why their drafts feel all over the place.
I Use This Prompt To Get a Skeleton
❝
"Create a newsletter outline on this topic: [YOUR TOPIC + ANGLE]. Structure: 1) Hook that creates curiosity, 2) The problem my reader actually faces, 3) The solution with 3–4 actionable steps, 4) A real example, 5) One clear CTA. Keep it skimmable."
I don't write from the outline directly. I just use it to make sure the piece has a clear arc before I start drafting.
Step 3: Write section by section — never the whole thing at once
This is where most people go wrong. They ask AI to "write my newsletter" and get back something that sounds like a corporate press release.
I set this as my writing persona first, every single time:
❝
"You're writing for an AI productivity newsletter. Voice: conversational, direct, like a smart friend explaining things over coffee. Short paragraphs — 2 to 3 lines max. No fluff, no corporate phrases. Sound human. If it sounds like AI wrote it, rewrite it."
Then I Write Each Section Separately
❝
"Write just the HOOK for my newsletter. Topic: [X]. 1–2 sentences. Creates curiosity or names a real pain. Don't start with 'Are you...' or 'Have you ever...'"
One prompt per section. Hook, problem, solution, example, CTA — each gets its own request. The output quality is night and day compared to asking for everything at once.
Stop Fine-Tuning Models You Don’t Need
Fine-tuning sounds like the answer until you factor in the cost, the data pipeline, and the six months before a bigger model makes yours obsolete. Most of the time, prompt engineering or better context gets you there. But sometimes it doesn't — and that's where things get interesting.
In this free night session, Aaron Gallant covers the real tradeoffs behind fine-tuning LLMs, from synthesizing training data with frontier models to running PEFT and QLoRA on constrained hardware. You'll learn when smaller, specialized models actually beat throwing money at a bigger one — and why data curation is the work nobody wants to talk about. Built for engineers who want to make the right call, not just the cool one.
Live and remote. Wednesday, June 3 at 5 PM CT. Register now.
Register Free
Step 4: Run the humanizing pass
Raw AI draft still has tells. Too balanced. Too polished. Zero personality.
Before I do anything else, I paste my full draft and run this:
❝
"Rewrite this to sound less like AI wrote it. Vary sentence lengths — some very short, some longer. Remove filler phrases like 'it's important to note' or 'in today's world.' Add one slightly imperfect or opinionated line. Make it feel like one person talking to another. Don't change the meaning or structure. Here's the draft: [PASTE DRAFT]"
After this, I read the whole thing out loud. If I stumble anywhere, I rewrite that sentence by hand. That's it.
Step 5: Generate subject lines — never go with your first idea
The best newsletter in the world still dies if nobody opens it.
I generate 10 subject lines every time, using this prompt:
❝
"Write 10 subject lines for a newsletter about [TOPIC]. Mix styles: curiosity gap, specific result, personal/relatable, counterintuitive. Under 50 characters each. No clickbait. The newsletter covers: [ONE LINE SUMMARY]."
I pick the top 3, sit with them for a minute, and ask myself: would I actually open this? The one that makes me a little nervous — that's usually the right one.
❝
💡 Quick reminder: In Beehiiv or Substack, use H2 for section headings, Callout blocks for key tips, and a divider before your CTA. Looks cleaner on mobile.
Your accounting should keep pace with your business. BELAY's Financial Experts handle books, payroll, and reporting so you can lead with confidence. Download the Free Guide
The honest truth
AI is a first-draft machine. Not a publish button.
The writers winning right now use it to kill the hardest part — starting. Then they spend real time on voice, editing, and making it sound like something only they could write.
Save these prompts. Run them in order next time. Blank page to publish-ready in 90 minutes.
If this was useful, forward it to one person who's still spending 5 hours on a single newsletter.