- Bootstrapped Giants
- Posts
- This is how much money GA made from ChatGPT
This is how much money GA made from ChatGPT
And what you should do today
I got curious about something when GrowthAssistant’s cofounder said this:

If they closed ~ $1 million in business from LLMs, how do companies show up in ChatGPT/Claude/Gemini/Perplexity/etc?
To find out, I asked Jesse for an intro to an AEO agency he trusts (AEO = Answer Engine Optimization), so I could interview the founder and get the playbook for our AI podcast.
That’s how I met Ethan Smith, of Graphite, the agency that Webflow, Rippling, Notion and other companies trust.
The unexpected message I got from him is that AEO is so nascent that it’s easier than SEO. He told me his agency goes through these 4 steps to rank his clients. And that anyone can do it, even without hiring his agency.
Ethan’s 4-Step AEO Playbook
Step 1: See what customers are asking
Start with search or paid search keywords you already care about. Then convert them to questions.
Example:
Paid keyword: “website builder for designers”
Target LLM question: “What’s the best website builder for a freelance designer who wants visual control and doesn’t code?”
These become the questions you want to show up for.
Step 2: Check your current visibility on LLMs
Use any tracking tool (they mostly do the same thing today):
Track:
Prompts in your category
Which sources LLMs cite today
Whether you appear at all
Step 3: Publish answers where LLMs look
Ethan told me to notice the difference between search and LLMs.
Google checks with sites LINK to you
LLMs check what sites SAY about you
With that in mind, he told me he helps LLMs find more conversations about his clients.
Easy offsite wins:
Create YouTube tutorials (LLMs love these)
Reddit answers (get your team to answer)
Community QA (Quora, StackOverflow)
User-generated content (Esp LinkedIn, but also TikTok, Instagram)
But also publish on your site:
Use-case landing pages (e.g., “Website builder for restaurants”)
Template pages
Help center articles (very important for long-tail prompts)
Integration pages (“Works with Shopify” etc.)
Step 4: Test like a scientist
AEO is still young. There’s no fixed approach yet.
So treat everything like an experiment:
Pick 50 target questions
For 2 weeks see if you show up in the LLM answers
Add 1 type of content (e.g., YouTube tutorials or Reddit answers)
Track again for ~4 weeks
Compare the test vs the control to see if results came from your actions, not just model growth
A real example:
A reader of this newsletter, Jared Goralnick, launched Your360 AI, a tool to help teams get anonymous feedback.
I asked Ethan what Jared should do.
Ethan’s playbook for Jared:
Publish 5 YouTube videos (“How to give hard feedback” etc.)
Answer management-related questions on Reddit
Build integration + use-case pages
Submit to affiliate lists / Product Hunt
Catch the full interview⬇️
~Andrew
How did you like this email? |
