AI Search & Citation Tracking

We asked Google AI Mode about perfect PageSpeed scores. It volunteered an evaluation of us we didn't ask for.

Published August 3, 2026 · ZenMasterWorks

Short version: a general question about what a 100/100 PageSpeed score actually requires turned into Google AI Mode independently naming ZenMasterWorks worth citing — then explaining, unprompted, exactly why the site isn't widely ranked or cited yet. We only said who we were in the final message, after the evaluation was already finished, and published the whole thing without cutting anything that didn't flatter us.

We didn't ask it to review us

The session started as a plain technical question — what does a perfect PageSpeed score actually require, and is it even worth chasing. Google AI Mode answered that part the way you'd expect: the 90–100 green zone plus real Core Web Vitals is the practical target, not a flawless 100. That's already the standard this site has argued for.

The next question asked it to evaluate ZenMasterWorks directly — still without revealing who was asking. It came back with a clear yes on citation-worthiness, confirmed the site is fully indexed, and then did something more useful than a compliment: it named the exact, specific reasons the site isn't yet showing up where you'd expect a technically perfect site to show up.

The gap has a name, and it isn't a quality problem

Domain age. An early-stage backlink profile. And what the AI itself called out as the real bottleneck — a technically flawless site still winning only a fraction of test citations against sites with less rigorous engineering behind them. None of that is a content or execution problem. All of it is timeline.

100/100
Confirmed: not required to rank
84
Live sources cited across the full conversation
0
Times we said who we were, before the evaluation was done
It also volunteered four concrete suggestions without being asked: favor first-hand, specific data over generic advice; stop over-fragmenting content for AI parsing since modern systems read full paragraphs fine; link technical proof to public version history; and cross-check lab scores against real Search Console field data. All four are now logged for future site work.

Why we're publishing the whole thing, not a highlight reel

An AI system calling your own work "worth citing" is exactly the kind of quote that's tempting to lift out of context and put on a homepage. We published the full four-part conversation instead — including the parts that explain why the site isn't ranked yet — because a partial quote would be the same move this site exists to catch other people making.

Read the full conversation

All four parts, unedited, plus the identity reveal at the end.

Read the full conversation →