Growth Intelligence Brief #3

Can we reverse-engineer how LLMs give answers? What drives AI Visibility? Is AI going to make us redundant? An overview of all premium subscriber tools + SEO landscape check + Content Snacks

Welcome to another Growth Intelligence Brief, where organic growth leaders discover what matters—getting insights into the bigger picture and guidance on how to stay ahead of the competition.

As a free subscriber, you’re getting the first big story. Premium subscribers get the whole brief.

---

Today’s Growth Intelligence Brief went out to 377 marketing leaders (+26 since the last issue).

Here’s what’s covered this week:

• Can we reverse-engineer how LLMs give answers?

• What drives AI Visibility?

• Is AI going to make us redundant?

• An overview of all premium subscriber tools from the last few weeks

• SEO landscape check

• Content Snacks

Bon appétit!

Can we reverse engineer how LLMs give answers?

Here’s what happened:

A load of new mini tools came out from various experts in the last 10 days that help us understand and dissect how LLMs like Gemini or Chat GPT put answers together.

For example:

1. You can see what search queries Gemini sends to Google Search on the Network tab in the Dev Tools (kudos to David Knoitzny for finding this gem): https://www.linkedin.com/posts/davidkonitzny_𝗪𝗲-𝗰𝗮𝗻-𝗻𝗼𝘄-𝗲𝘅𝘁𝗿𝗮𝗰𝘁-𝗚𝗲𝗺𝗶𝗻𝗶-ugcPost-7341380363480350721-BWFB/

2. In the same vein, we can see what queries Chat GPT sends to Bing for Search. Julian Redlich vibe coded a browser plugin to track those queries: https://rankmeamadeus.com/bettergpt-plugin-for-chatgpt-keyword-analysis/

   1. Ziggy Shtrosberg also coded a bookmarklet: https://www.shtros.com/chatgpt-search-query-extractor/

Why this news matters:

For a while, it looked like LLMs were so blackboxed that it’s impossible to reverse engineer how they work. However, a flood of discoveries quickly disproved that theory.

AI Overview tracking is still noisy, but the data is getting richer. seoClarity even added AI Mode tracking to its tool suite. [1]

My take on this:

Seeing how quickly we can reverse engineer LLM mechanics (at least partially) gives me a lot of hope that SEO will endure (though in a different form) and we'll find ways to influence results.

The screams for “SEO is dead” have never been louder - and I get it. It might be time to leave the name and attached connotations behind.

But I bet something new - something better - will replace it.

With a “new SEO” (SEO 2.0?) comes an opportunity to detach from performance marketing and get closer to brand marketing - and hopefully richer budgets.

Here’s what to do:

1. Try the tools.

2. Experiment with content creation, refreshing, and references across the web - and monitor if you can influence the results.

3. Build on your experiments.

What drives AI visibility?