AI Citation Audit Query: “Best product analytics tools for SaaS” I ran this query across multiple AI engines and analyzed the citation sources behind the answers. Vendors appearing included: - Mixpanel - Amplitude - Heap - PostHog But the interesting part wasn’t the vendors.

A 17-day-old domain got cited for “Best feature flag tools in 2026.” We don’t sell feature flag tools. We’re not in that category. No authority. No backlinks. Still got cited by Perplexity. This wasn’t luck. It was a controlled test. We generated a page using a structural

Your SEO team is building a bridge that doesn't reach the other side. You’ve spent $50k on "high-quality" pillar pages. You’ve optimized for every keyword in the book. You have the backlinks. Yet, when a buyer asks Perplexity or Gemini for a vendor recommendation, your name

Structurally similar pages tend to get cited by AI 4.2x more than the mismatched ones. We ran a controlled experiment testing this hypothesis. We took 100 different B2B SaaS buyer intent queries and made 5 variations for each - total 500 queries. We automated the analysis to

Most SaaS pages don’t fail because of content. They fail because of structure. We analyzed a large set of SaaS pages trying to rank in AI answers. What we found was uncomfortable: The majority weren’t even being *considered*. Not because they were bad. Because they didn’t pass

TL;DR Same brand. Same query. Went from cited → not cited in 2 weeks. No penalty. No update. Just replaced. --- I ran this query: “best e-signature tool for small B2B teams that is not DocuSign” Clear BOFU intent — the model is trying to finalize a buying decision. --- Week

x.com/i/article/2039…

𝐀𝐈 𝐬𝐞𝐚𝐫𝐜𝐡 𝐢𝐬 𝐧𝐨𝐭 𝐫𝐚𝐧𝐤𝐢𝐧𝐠 𝐲𝐨𝐮. 𝐈𝐭’𝐬 𝐟𝐢𝐥𝐭𝐞𝐫𝐢𝐧𝐠 𝐲𝐨𝐮 𝐨𝐮𝐭. 🚫 Most brands won’t show up in ChatGPT, Gemini, or Perplexity. Not because of bad SEO. Because they don’t qualify. AI doesn’t “rank pages.” It selects answers using 3 filters: ⚙️

𝐖𝐞𝐛𝐌𝐂𝐏 𝐰𝐨𝐫𝐤𝐬 𝟏 / 𝟐𝟓 𝐭𝐢𝐦𝐞𝐬 😡 Been fiddling around with integrating with #WebMCP onto latticeocean to facilitate quick demos and audit requests. Even after doing everything fine in terms of WebMCP APIs (#Declarative and #Imperative), I got it to work with

𝗔𝗜 𝗲𝗻𝗴𝗶𝗻𝗲𝘀 𝗱𝗼𝗻'𝘁 𝗰𝗶𝘁𝗲 𝗰𝗼𝗻𝘁𝗲𝗻𝘁 𝘁𝗵𝗲 𝘄𝗮𝘆 𝗚𝗼𝗼𝗴𝗹𝗲 𝗿𝗮𝗻𝗸𝘀 𝗶𝘁. Here are the patterns that show up, you would have seen them too. After analyzing dozens of pages across multiple buyer-intent queries, here are the patterns that keep showing up.

It felt rewarding!