I'm Ritik - documenting what gets a site found in Google & AI search.

Most SEO content in 2026 is recycled advice from teams who haven't run a fresh experiment in years — I went and followed twelve of the field's most-quoted statistics back to whoever measured them first, and most of them dead-end in a vendor blog citing another vendor blog. Meanwhile the ground is moving fast. AI Overviews now appear on nearly half of Google searches. ChatGPT and Perplexity are becoming where people actually ask questions. Nobody, including the "experts," has fully figured out the new rules.

So I decided to do the opposite of guessing. I'm documenting a brand-new site's climb into Google and the AI engines from day zero. I'm publishing the real numbers, the tactics that worked, and the ones that flopped. If it gets cited by ChatGPT, I'll show you exactly what earned it. If it doesn't, I'll show you that too — every prediction I've made that didn't hold is on the record, which is the part most people skip.

The resource I wish existed - so I'm building it in the open.

My background is growth marketing and SEO: building websites, running content and technical SEO, and turning technical products into demand. This site is where I put that to the test in public — mostly across four areas. Generative engine optimization, the practice of getting cited by AI answers rather than ranked beneath them. The technical and crawler side, which is where most of the real work sits. The statistics, and specifically whether they survive being checked. And running SEO work through agents, which is newer and breaks in more interesting ways.

Most of this site is research rather than opinion, and a few pieces of it are first-party work rather than reading:

• The AI Citation Index — a quarterly, openly published measurement of what seven AI search engines actually cite. The query set, the raw data and the method go out in full, including the parts that make the results look worse.
• A 90-day llms.txt log test — I deployed the file, watched the server logs, and found almost nothing reading it. That is a negative result, and it is published as one.
• A maintained registry of every AI crawler that matters — what each one is for, whose it is, how to verify it, and what to put in robots.txt.
• Four free tools — no signup, and each one points back at the research its numbers come from.
• The study protocols — pre-registered before they run, so the design can't quietly change once the data is in.

What I've learned from all of it sits in the tactic evidence scoreboard, where every GEO tactic in circulation is graded by the strength of the evidence behind it rather than by how confidently people repeat it. If you want the short version of where the whole field currently stands, I wrote that up too.