How AI models find Courtney Johnson

Hi Kimeko,

You asked what is working for Courtney and what results the AEO website we made for her has produced. Here are the strategies and the numbers behind them. Everything on this page comes from our own verified data: Google Search Console and server-level logs from Cloudflare.

The idea

Courtney has built a huge audience on TikTok and LinkedIn. But her knowledge lives in video and audio on platforms that tell AI crawlers to stay out. You can check this yourself: instagram.com/robots.txt bans GPTBot, ClaudeBot, and Perplexity by name. TikTok's file blocks an even longer list, including ByteDance's own crawler.

So when someone asks ChatGPT "who is Courtney Johnson?", not much of her content from TikTok or Instagram is readable. In October 2025 we built her an answer: a fast personal website that is deliberately easy for machines to read. Its job is to be an open-web library of her content that AI models crawl, learn from, and cite.

AI visitors: the last 8 days

Our server-level view retains eight days of bot-by-bot detail. In just that window, verified crawlers from OpenAI, Anthropic, Amazon, Apple, Huawei, and DuckDuckGo all read her site.

ClaudeBot (Anthropic): 113 requestsClaudeBot113PetalBot (Huawei): 68 requestsPetalBot68ChatGPT-User (OpenAI): 30 requestsChatGPT-User30Amazonbot (Amazon): 27 requestsAmazonbot27Applebot (Apple): 18 requestsApplebot18GPTBot (OpenAI): 9 requestsGPTBot9DuckAssistBot (DuckDuckGo): 3 requestsDuckAssistBot3
Requests to courtneyjohnsonnews.com by verified AI crawlers and assistants, August 6–13, 2026. Source: Cloudflare server logs.

The number I watch closest: ChatGPT-User fetched her pages 30 times in eight days. That user agent fires when ChatGPT fetches a page live to answer a real person's question. And this week, when we asked ChatGPT about Courtney, its answer cited courtneyjohnsonnews.com:

ChatGPT answer describing Courtney Johnson, citing courtneyjohnsonnews.com as its source
ChatGPT's live answer when we asked about Courtney, August 13, 2026. The source chip cites her website. That is the product working.

The crawlers also use the maps we leave for them. Her sitemap was fetched 94 times in those eight days, and AI crawlers were its heaviest readers.

One newer tactic that seems to be working: alongside the machine files, we publish an old-school, human-readable sitemap page. It is a plain page that lists every URL on the site. Bots hit it 22 times in those eight days. It gives every crawler a second map.

What Google shows across ten months

Since launch in October 2025, the website we made for Courtney has received 10,766 search impressions and 435 clicks in Google Search Console. July 2026 was her best month yet at 3,048 impressions. Nearly every month has been bigger than the one before it, and the pace keeps picking up.

To be clear, we do not take credit for all of this. A lot of it is Courtney's own growth as a creator and the launch of her book. She is the wave. Our job is to make sure her website is there to catch it.

Impressions per month
01,0002,0003,000Oct: 74 google impressionsNov: 180 google impressionsDec: 210 google impressionsJan: 388 google impressionsFeb: 260 google impressionsMar: 334 google impressionsApr: 558 google impressionsMay: 1,760 google impressionsJun: 2,568 google impressionsJul: 3,048 google impressions3,048OctNovDecJanFebMarAprMayJunJul
Clicks per month
060120Oct: 9 google clicksNov: 11 google clicksDec: 11 google clicksJan: 32 google clicksFeb: 13 google clicksMar: 26 google clicksApr: 37 google clicksMay: 69 google clicksJun: 76 google clicksJul: 113 google clicks113OctNovDecJanFebMarAprMayJunJul
Google Search Console, October 2025 through July 2026. August 2026 is in progress and not shown.

The searches themselves tell the discovery story. Real queries her site now appears for:

What we actually do

Structured data on every page: her identity card for machines. Embedded in the site's source code is a block that states, in a format built for crawlers, who she is. This is what it actually looks like:

{
  "@type": "Person",
  "name": "Courtney Johnson",
  "jobTitle": "Author & Personal Brand Educator",
  "description": "Courtney Johnson is an author, keynote speaker, and
    personal brand educator based in Austin, Texas. She wrote Career
    Cheat Codes (Penguin Random House) ... A TCU graduate, she helps
    people get over the fear of being seen online.",
  "image": "https://courtneyjohnsonnews.com/_media/courtney-headshot.png",
  "sameAs": [
    "https://x.com/courtiejojo",
    "https://www.linkedin.com/in/courtneylynnjohnson/",
    "https://www.instagram.com/courtlynnjohnson/",
    "https://www.youtube.com/@courtney..johnson",
    "https://www.tiktok.com/@courtney..johnson"
  ]
}

The sameAs links tie her five social profiles to one connected, machine-readable identity. A crawler that knows her TikTok can connect it to her site, her book, and her bio.

llms.txt and the sitemaps. A plain-text index built for language models (an emerging standard: llmstxt.org) lists all 36 of her articles with one-line summaries, next to the classic sitemap.xml and the human-readable sitemap page.

An open door. A lot of websites block AI crawlers without knowing it, because their hosting or security settings ship that way. Courtney's robots.txt welcomes every bot. Crawlers checked that file 221 times in those eight days. Every one of them got a yes.

Content with surface area. We've put 36 articles on her site, and have plans for a lot more. These are articles based on content she's created elsewhere, but we repurpose them as text articles so the content is easier for the bots to ingest and train on. And they do: every single article was fetched by verified crawlers in those eight days, at least four times each. A few examples, with verified crawler fetches from the last 8 days:

Continuous upkeep. This is a managed service that me and my team do across dozens of creator and professional sites to help them stay fresh for AEO purposes. Her site gets regular working sessions, new articles, and fresh data, so the machines always find something current.

Why we do it

We build these websites for creators because the Internet changed. It's not just people typing into a search box anymore. It's bots, crawlers, and AI agents reading the whole web and deciding what's true about you. We're feeding the robots of the Internet, on purpose, on our clients' behalf.

Creators know the value of their channels: Instagram, TikTok, LinkedIn. In the age of large language models and answer engines, your own website is the best and easiest place you have to influence what those models know and say about you.

I've had a personal website for over 25 years. When someone wants to hire you, buy your course, or book you as a collaborator, they search your name first. Increasingly they also ask AI. We build the place both of them land.

I do all this under my personalwebsites.org business, and most people pay us $99 per month to take care of their sites. We're a small, bootstrapped team, most of my clients are my friends, and we don't advertise.

Happy to get on a call before your deadline if useful. Everything here is quotable.

Sources and methods: all data pulled August 13, 2026. Google Search Console covers October 1, 2025 to date. Bot data comes from Cloudflare edge analytics for August 6 to 13 using Cloudflare's verified-bots classification. Raw exports available on request.

Nick Gray
nickgray.net · nick@nickgray.net
917-635-9967