Hey, it's Martin.
A few months back I wrote about the AEO opportunity. Traffic from ChatGPT converts 6x better than Google. The long tail is wide open.
Cool. But then what?
Most people read that stuff, nod along, and go right back to their old ways. Because the tactical path forward isn't obvious.
So let's fix that. Here's the playbook:
Step 1: Find the questions that matter
Forget broad keywords. AEO runs on hyper-specific queries. The kind of 25-word questions people type into ChatGPT like they're texting a friend.
Start with a simple grid. Your buyer personas on one axis. Their journey stage on the other:
awareness
consideration
decision.
Now fill each cell with questions. Not keywords. Questions.
Where do you find them? Sales calls. Support tickets. Reddit threads in your niche. The stuff people ask before they buy.
Here's a prompt to to accelerate this:
"Act as a [your ICP]. You're evaluating solutions for [problem]. What specific questions would you ask an AI assistant before making a decision? Focus on integration details, pricing edge cases, and comparison questions."
You'll get 20 questions in 30 seconds. Most will be garbage. Three or four will be gold. Dig deeper on these.
Step 2: Structure content for lazy readers
LLMs don't read like humans. They grab chunks. Middle paragraphs. Random sections.
This means your content needs to pass this test:
If the AI only eats the middle of your article, does it still make sense?
Tactical shifts:
Answer the question in your first sentence. Not your third paragraph. First sentence.
Make every section standalone. Each H2 should contain enough context that it works even if nothing else gets read.
Mention your product in context. Every other paragraph or so. Not salesy. Just present. "When we built this workflow for a client using [your tool]..." The AI needs to connect the topic to your brand.
And here's the unsexy one: original data wins. LLMs are hunting for information that doesn't exist anywhere else. That case study with actual numbers? That internal benchmark you've never published? That's citation bait.
Step 3: Measure visibility (automatically)
This is where it gets fun.
You can build an AI agent that asks your target questions to ChatGPT every month, tracks whether you're getting cited, and identifies exactly which sources the AI is pulling from.
I built one in Relay. The workflow is simple:
Create a table with your high-intent questions. Loop through each one with a custom prompt that has web search enabled. After the loop, feed all the outputs to an AI analyst step that produces a gap report. Email it to yourself monthly.
Now you have a recurring strategy doc landing in your inbox. Which questions cite you. Which don't. And specifically which Reddit threads, blog posts, or review sites the AI trusts enough to reference.
That's your hit list for off-page work.
The freshness problem
One thing we're still figuring out: content decay hits harder in AEO.
LLMs seem to weight recency. A blog post from 2023 loses ground to something published last month, even if the older content is better.
The optimal solution here is to have content agents that updates and refreshes content regularly.
For our consulting business we're currently running the site on Framer. We’ve started looking at moving our blog and playbook content to some other storage that lets us hook in agents to refresh and improve content over time without manual rewrites.
Haven't cracked it yet. But the direction is clear: static content is a liability now.
The AEO game rewards operators who show up consistently with fresh, specific, citation-worthy content.
That's not a content calendar. That's a system.
That’s it for today. If you have ideas you want me to write about, hit reply!
Until next week,
Martin
PS: feel free to connect with me on LinkedIn and say hi! It’s always fun to chat with new readers.
PS2: want to accelerate AI adoption in your business? Book a free 30-min disovery call here.