(Week 3 of 10) ChatGPT read my website 579 times last month. Here's how to make it read yours.

Day 84. $125,000 ARR on the product. Over $100K/month across the whole business. 416 days left.

I pulled a report this week that stopped me cold.

In the last 4 weeks, AI crawlers hit my website 30,508 times. Fifteen different bots. Google's, OpenAI's, Anthropic's, Meta's, Apple's, Perplexity's, all quietly reading my pages.

But one number in that report matters more than all the others: ChatGPT-User, 579 requests.

That bot isn't a crawler building an index for later. It fires in real time, when an actual human asks ChatGPT a question and ChatGPT goes to read a website before answering. Which means 579 times last month, a real person asked an AI something, and the AI walked over and read MY content to form its answer.

Every one of those 579 moments was a potential recommendation. And a few weeks ago one of them became a paying customer who told me, word for word, "I found you on Gemini."

This is Week 3 of the roadmap, and it's the channel I'd bet the next decade on: getting AI to recommend your product. Here's the entire playbook.

First, understand what makes this channel strange.

When your customer asks ChatGPT or Gemini "what's the best tool for [your problem]," the AI answers with 2 or 3 names. If you're one of them, you get a warm lead for $0. If you're not, you lost a customer and there's no dashboard anywhere that shows it. No impression counter. No missed click. Silence.

So most businesses are losing this game without knowing it's being played. Which is exactly why the founders who move now win big: the shelf space in an AI's answer is tiny, and it's being allocated RIGHT NOW based on what's readable on the internet.

The good news: what AI reads is content, and content is the one asset where a solo founder can outwork a giant.

The workhorse. For every major competitor in your space, publish a page: "[Competitor] vs [You]."

Why it works: when someone asks an AI "what's the best alternative to [Competitor]," the AI goes looking for content that directly answers that. A fair, specific comparison page is exactly the shape of answer it wants to cite. I wrote these pages thinking I was doing boring Google SEO. Turns out I was feeding every AI on earth the exact document they needed to recommend me.

The format that works:

Open with one honest paragraph on what the competitor does well. Then a plain-language comparison table: pricing, channels, setup, what's included, what costs extra. Then who should pick THEM (yes, really), and who should pick you. Real numbers everywhere. No adjectives where a number could live.

The fairness isn't charity. It's strategy. AI models are trained to distrust one-sided sales pages and cite balanced ones. Being fair to your competitor is what gets you cited.

Different search, different page: "[Competitor] review 2026."

Here's a secret about the person typing that: half of them are already unhappy. Nobody researches reviews of a product they love. A review page catches people at the exact moment they're looking for a reason to leave.

I lived this one from the customer side. A friend who runs a diamond store needed an AI chat solution, so I went shopping like a normal buyer. Found the big name in the category. Setup was genuinely quick, entry price looked great at $40 a month. Then I read the fine print: removing their branding cost serious money, replies burned through credits, per-agent fees stacked up, and their Trustpilot was sitting at 3.8 with one-star reviews about surprise charges.

That shopping trip became a review page and a YouTube video. Real experience, real numbers, real screenshots. That's content no agency can fake, because I actually went through it.

The lesson inside the lesson: your best review content comes from genuinely shopping your competitors. Sign up for their trials. Feel their onboarding. Read their fine print. You'll either learn what to steal or find what to publish. Usually both.

This one I learned filming the video version this week, and it applies to every piece of comparison content you make.

My instinct was to tease: hook them with mystery, reveal my product at the end. Wrong, and here's why. The person reading a "[Competitor] review" isn't a scroller you need to trick into staying. They SEARCHED. They already want the answer. What they fear is exactly you: a fake review that's secretly an ad.

So you do the opposite of ambush. You disclose the itinerary in the first breath: "This is my honest review: the pricing, what it does well, the fine print that surprised me, and what I ended up using instead." That last clause isn't bait. It's honesty about where this is going, and it's what earns the reader's trust for the two-thirds of the page where you genuinely review the competitor.

The proportions matter: give most of the content to honestly reviewing THEM. Make the page useful to someone who never clicks your link. The recommendation at the end then reads as a conclusion, not a switch. And always say the quiet part: "Full disclosure, I build the alternative." AI models and humans both reward it.

Remember: 30,508 of my last month's visitors were robots. Format for them.

Clear headings that match real questions ("How much does X actually cost?"). Tables for anything comparable. Specific numbers with dates ("as of August 2026, the entry plan is $40/month"). Plain sentences over clever ones. Your pricing ON your website, not hidden behind "book a call," because an AI can't recommend a price it can't read.

And put it on a real website you own. AI doesn't crawl your Instagram. It crawls URLs.

Every comparison page gets a YouTube version. Same content, spoken, with your screen visible.

Why double the work? Because YouTube results rank in Google, video transcripts get indexed, and the person who won't read 1,500 words will watch 5 minutes. One research session becomes two assets on two surfaces.

One targeting note from this week's filming: a search viewer wants the answer immediately, so open the video by serving the search ("here's my honest review of X"), not with a flashy teaser. Save the teaser cuts for Shorts, where you're catching scrollers instead of searchers. Same footage, two front doors.

Two free rituals:

Ask the AIs yourself. Open ChatGPT, Gemini, and Perplexity. Ask "what's the best [your category] for [your customer]?" and "[your competitor] alternatives." Record whether you show up. Do it monthly. That's your ranking report.

Look for the bots in your traffic. Most modern hosts (and Cloudflare) show crawler analytics. Look for GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, PerplexityBot. If they're reading you, you're in the pipeline. If they're not, your content isn't reachable, and that's the first thing to fix.

Pick your ONE biggest competitor. This week:

1. Shop them like a customer. Trial, onboarding, pricing page, fine print. Take screenshots and notes.

2. Write the comparison page: honest opening, comparison table with real numbers, who should pick them, who should pick you, disclosure that you're the maker.

3. Ask ChatGPT, Gemini, and Perplexity your category question and record whether you appear. That's your baseline. We'll check it again in a month and you'll see this machine move.

Next Wednesday, Week 4: Cold Email. How to book calls with total strangers without burning your domain or your reputation.

Day: 84 of 500

Product ARR: $125,000

Combined business: $100K+/month run rate

AI crawler visits (last 4 weeks): 30,508 across 15 bots, including 579 real-time ChatGPT fetches

Filmed this week: The competitor review video, with a twist at the end involving Preeya that I'm not spoiling here. YouTube: https://www.youtube.com/@raj.thareja

Biggest realization this week: The most valuable reader of your website isn't a person anymore. Write for both.

Hit reply and tell me which competitor you're writing your first comparison page against. I read every one, and if you send me a draft, I'll tell you whether an AI would cite it or skip it.

And if you know a founder still spending everything on ads while their competitors quietly colonize the AI recommendation shelf, send them this issue. In a year, this channel won't be a secret anymore.

See you next Wednesday.

Dhiraj