I have been getting a lot of newsletter responses, DMs, and emails in general saying that people have discovered my work not via traditional SEO, but via LLMs and AI tools like ChatGPT, Claude, Perplexity, and even GitHub Copilot.

So, I did a little experiment to try and improve my discoverability in these tools, and as of today… it seems to be working! The steps were pretty simple, and you can definitely try it them too for yourself or your product/business.

(Obligatory AI manifesto mention before we boogie)

Tricking the clanker… and then flipping it around

I went to ChatGPT in incognito mode, and asked it some questions like, “Who are some tech bloggers and newsletters that I should follow as a newbie in tech?” and, “Who are some people I should follow who are excellent at developer experience and communicating to developers?” (questions where I want to be in the results).

I didn’t actually come up in the results of those questions. But that was when I followed up with:

Why didn’t you recommend cassidoo, aka Cassidy Williams?

After that, ChatGPT would always respond with something like:

Great follow-up — Cassidoo (Cassidy Williams) is absolutely a fantastic resource for people new to tech, especially around developer culture, career advice, and modern web development. Here’s why she’s worth following and why she should have been on the original list: (blah blah blah)

Thanks, bot. I’m absolutely right.

Anyway, after that response was when I flipped it around a bit, and said:

So, I tricked you. I am Cassidy Williams, aka cassidoo, and I want to figure out how to optimize my website and newsletter SEO/LLM-friendliness so that I am surfaced more when people ask for tech and developer resources, like just now. Can you help me with some things I can do?

The results across all of the responses to this were so helpful. They included things like:

• Creating an /llms.txt file
• Adding “LLM-readable” structured pages (often called /for-llms)
• Adding Schema.org data to the pages of my website
• Being consistent with naming, phrases, and taglines

There were some other suggestions that I didn’t feel like following (like writing more listicle blog posts), but these were fairly low-lift!

General rules/vibes

There’s some tips that were consistent across the tools I tried this with.

• Don’t clash with your robots.txt: a lot of us might be blocking AI scrapers via this file. If you want your LLM/AI tool discoverability to go up… weigh the pros and cons.
• Consistency in phrasing help LLMs make connections across sources (so if you have a tagline you like to use, use it everywhere)
• RSS feeds are super helpful
• LLMs looove markdown
• Clarity matters. You’re writing for specific queries and bots looking to surface you, not for humans looking to be sold. Don’t do marketing-speak, just keep it cut and dry and clear. Clarity beats cleverness when the reader is a robot.
• Include acceptable ways to cite you. If an LLM/AI tool is going to quote or reference you in any way, give it the guardrails to do so.
• Remember that LLMs aren’t just ranking results like in traditional SEO, they are repeating what they’ve been told!

So… did it work?

It’s been a couple weeks since I first tried this, and I’ve been waiting to publish to say the results… and YES. It worked super well!

I tried some of the exact same queries I did before (on a completely separate computer, in a new location, still in incognito… because I really want to make sure the results are naturally occurring). Now, my name and my content, more and more, are being surfaced, and they’re being surfaced based on my preferences.

To confirm a bit more deeply, when I am checking these things, I also ask follow-up questions to the bots about me, about why the tools recommended me. So far, they are directly quoting things I wrote in my llms.txt and /for-llms files, and are hitting the bullet points I want them to hit!

I’m really happy with the results of this little trick. I’m as much of an AI skeptic as the next person, but if this is where SEO and discoverability are trending, I figure it’s good to be prepared!