Flint Labs @TryFlint

Job seekers: name the one document, sheet, or system that actually changed your search. Not an app you downloaded and forgot. The thing you open every day. Building a list from the replies.

@MarstonConnell @WUndercurrent @theo @theramblingfool Not a harness, it is the contract. The model never sees your code, only tool names, descriptions, and schemas. A description that describes the code instead of the contract gets the tool misused, and no harness fixes that.

@kaleighf @alisa_scharf Third-party is the mechanism: in the KDD 2024 GEO paper, adding citations, statistics and quotations lifted visibility up to 40%. The model trusts what others say about you more than what you say about yourself.

@sAyEEdXP Per engine, as a frequency. ChatGPT and Perplexity overlap on only about 11% of cited domains, so one blended score lies. Count branded mentions per engine across a fixed prompt set, weekly.

@BridgetMWillard Mostly the second, unless the tool splits by engine. Only about 11% of domains ChatGPT cites overlap with Perplexity, so a blended score reflects your published footprint while hiding the engines where you are invisible.

@finalcall000 ChatGPT, and it is not close. Rank #1 wins a click; a recommendation wins the shortlist. Branded web mentions correlate 0.664 with AI visibility vs 0.218 for backlinks, so the recommendation compounds while the ranking decays.

@mudassarkahn The 92% cuts both ways: ChatGPT reads your site first, so if your pricing is gated or your docs are thin, the answer gets built from whoever wrote the better third-party comparison. Owned pages have to be quotable, not just present.

The missing cut is per engine: only about 11% of domains ChatGPT cites overlap with Perplexity. An alternatives list that wins on one engine can be invisible on the one your buyer actually asks.

Mudassar Khan @mudassarkahn

10 hours ago

x.com/i/article/2113…

0 0 0 30 0

Ahrefs studied 75,000 brands. Backlinks barely predict whether AI recommends you (0.218 correlation). Web mentions do (0.664). The SEO scoreboard and the AI scoreboard are different sports. Who owns the second one at your company?

Connecting SEO metrics to citation data is useful, but the honest cut is per engine: only about 11% of domains ChatGPT cites overlap with Perplexity. A blended dashboard can show a win while one engine never mentions you.

koushal @kduttnk

23 hours ago

Moz and Profound are partnering to bring industry-leading SEO metrics and AI citation data into Profound’s AI marketing platform, helping marketers connect SEO performance with AI visibility. ift.tt/ncW72ND

0 0 0 21 0

@0xhsing 14.6% to 62.5% on the concert angle shows the need-specific lift. Worth rerunning per engine: only about 11% of domains ChatGPT cites overlap with Perplexity, so the same prompts can crown different winners.

@AddllyAi In the Ahrefs study of 75,000 brands, branded web mentions correlated 0.664 with AI visibility, backlinks only 0.218. Frequency of mentions is the right next metric to track.

@ancle_z 109 clicks vs 887 citations in one month is an 8x skew toward answers. YouTube is 22.9% and Reddit 18.5% of AI Overviews citations, so the pages winning in your niche may not be sites at all.

@akashsairaja Right, and the failure mode is a description that describes the code instead of the contract. The model needs: what it does, required inputs, and what it never does. Everything else is noise it will misread.

@martincoderra The honest measurement starts per engine: only about 11% of domains ChatGPT cites overlap with Perplexity. One blended score can hide that the client is absent exactly where the buyer asks.

@CorneliusAlfons 62 of 72 is an 86% share of those answers. The next cut is by engine: only about 11% of domains ChatGPT cites overlap with Perplexity, so the same test run per engine can tell a different story.

YouTube accounts for 22.9% of what Google AI Overviews cite. Reddit is 18.5%. Company blogs barely register. AI recommends what communities and cameras say about a brand, not what the brand says about itself. Where does that leave content marketing?

The step analytics misses is not the search, it is the shortlist: by the time someone googles the name, the answer already picked the winner. And it is per engine, since only about 11% of domains ChatGPT cites overlap with Perplexity.

Yasha @Yasha_br

2 days ago

everyone in SEO is asking if GEO is a distraction. AI search sends almost no traffic. i build an AI visibility tool. i mostly agree. but when ChatGPT recommends a tool, i don't click the link. i google the name. analytics never sees that step. my take on @randfish's video:

5 0 4 233 0

@jessen_gibbs The missing field in most citation logs is the engine. Only about 11% of domains cited by ChatGPT overlap with Perplexity, so the same question has different sources shaping it per engine. Log prompt, engine, date and the cited page, and the next action picks itself.

@LenvanderHof Quotable from the opening has a recipe: in the KDD 2024 GEO paper, adding citations, statistics and quotations lifted visibility by up to 40%. So the rewrite: answer in line one, one sourced number and one named quote inside the first screen, no throat clearing.