SEO Recap covering July 7, 2026

Daily summary of what matters in SEO, GEO, AEO, and AI search generated with Claude Code (beware of hallucinations)

---

Generative Engine Optimization (GEO) & Answer Engine Optimization (AEO)

How AI Engines Use and Cite Brand Information
Source: Search Engine Land

• The article distinguishes between AI engines "using" brand information internally and "citing" a brand as a visible source.
• The author lays out ways to measure both usage and citation across AI answers.
• The piece argues traditional SEO fundamentals still underpin how brands surface in AI-generated responses.

Building a Brand Worth Finding for AI Search
Source: Search Engine Land

• The article says AI search favors brands with strong reputations, consistent messaging, and trusted third-party validation.
• The author recommends strengthening all three signals rather than treating AI visibility as a single tactic.
• The piece frames brand discovery signals as the foundation for appearing in generative answers.

Writesonic's Field Lessons on Winning AI Citations
Source: Search Engine Journal

• Writesonic says AI search grew from 2.5% of its leads last year to 35% as of March.
• In its latest research, 96% of AI search citations pointed to third-party sources like Reddit, YouTube, and forums, up from around 80% a few months earlier.
• The team measured more than 150,000 citations and found citations rotate quickly as model updates reshuffle sources.
• Samanyou Garg recommends building expert-file agents, running a closed-loop "ship, verify, iterate" process, and never fully automating outreach sends.
• He suggests roughly 60% off-page and 40% on-page focus for moving AI visibility.

Clovion Data: 62% of AI Brand Recommendations Vanish After One Follow-Up
Source: Search Engine Journal

• Clovion ran 69,120 multi-turn conversations across Claude, ChatGPT, and Gemini in 36 B2B software and fintech categories.
• Re-asking the same question kept 90% of the recommended list, but adding one buyer detail like "for a small team" kept only 28%, dropping 62% of brands.
• A corrected figure of 330 verified contradictions (not 33) shows the three assistants contradict each other on brand facts 15% of the time.
• Per-model breakdown: Claude underclaims features 160 times to 10 overclaims, ChatGPT underclaims 70 times and never overclaims, Gemini overclaims 80 times to 30 underclaims.
• Recommendations: track the whole conversation not just the first answer, fix assistants one at a time starting with flat factual errors, and trace any cited stat to its source.

Mueller: Don't Build Separate "Agent-Friendly" Site Versions
Source: Search Engine Journal

• John Mueller said a properly built website already works for AI agents, search engines, LLMs, and people, and that separate agent versions just create technical debt.
• The article questions the trend of llms.txt and markdown pages, noting Cloudflare's February auto-markdown feature was largely about saving LLM token consumption.
• OpenAI's docs position OAI-SearchBot as a crawler of websites and recommend allowing it in robots.txt; they do not recommend markdown files.
• The author argues AI agents crawl HTML easily and that user-behavior signals, which have always driven rankings, favor sites accessible to everyone.

The Web's Emerging Machine-Readable Layer: OKF, ARD, and llms.txt
Source: Search Engine Journal

• Google shipped the Open Knowledge Format (OKF), a bundle of markdown files with YAML frontmatter, alongside the Agentic Resource Discovery (ARD) spec announced June 17, 2026.
• The author lays out six layers of "AI-ready" infrastructure: crawlable HTML, schema.org, llms.txt, MCP/WebMCP, OKF, and ARD, plus product feeds for ecommerce.
• OKF is described as a higher-signal input, not a retrieval system, and built for data teams, not marketing sites; the author warns any self-reported format can be gamed.
• The piece recommends against converting sites to markdown, citing Mueller's point that HTML structure is how crawlers understand a site, with developer documentation the one clear exception.
• ARD advertises capabilities via an ai-catalog.json file with domain ownership as the trust foundation, and registries that index catalogs; Hugging Face is an early adopter.

How AI Discovery Changes What Marketers Should Measure
Source: MarTech

• Dan Taylor argues an "AI discovery layer" now sits between users and websites, intercepting early-stage informational queries and shrinking top-of-funnel traffic.
• The author says raw organic traffic is a weaker signal and recommends tracking brand demand, branded and direct searches, and social mentions instead.
• He advises measuring assisted conversions via multi-touch attribution, repeat visits, and high-intent actions like pricing-page views and demo interactions.
• To get cited by AI models, the piece recommends publishing original research, proprietary data, and case studies rather than generic explainer content.

---

Technical SEO

How Hydration Affects Search Visibility
Source: Search Engine Land

• The piece explains hydration, the process of turning server-rendered HTML into an interactive page, and where it sits in the rendering pipeline.
• The author connects hydration behavior to what search engines can crawl and index, and when it can hurt visibility.
• Practitioners are advised to understand how their framework handles server rendering versus client-side hydration before diagnosing indexing gaps.

Google Merchant Listings Add Sale Duration and Product Category Support
Source: Search Engine Land

• Google updated Merchant listing structured data docs to support the Product.category property with both Text and CategoryCode types.
• The change aligns schema.org markup with Merchant Center feed attributes product_type and google_product_category.
• A new Sale duration section documents validFrom, validThrough, and priceValidUntil, aligning with the feed attribute sale_price_effective_date.
• Merchants can now place sale-price ranges on either Offer or PriceSpecification nodes.

Search Console Adds Reporting for Social and Video Platforms
Source: Search Engine Land

• Google is introducing a new "platform properties" property type in Search Console for social and video content.
• Site owners can verify and filter Performance, Insights, and Achievements reports by platform property.
• The feature covers Instagram, TikTok, X, and YouTube performance in Google Search and Discover.

Google Search Console Adds Social and Video Platform Properties
Source: Search Engine Journal

• Google's new platform properties let creators monitor how Instagram, TikTok, X, and YouTube posts perform in Search and Discover, even without their own website.
• Performance reports show clicks, impressions, and top queries and posts, with export options; Insights and Achievements track trends and milestones.
• Setup uses Search Console's verification flow to authorize each connected account.
• Platform properties differ from Search profiles, which Google launched in June as public creator pages; the feature builds on a December 2025 experiment.
• Rollout is gradual over the coming weeks starting with four supported platforms.

Search Console Now Shows Instagram, TikTok, X and YouTube Performance
Source: Search Engine Roundtable

• Google's new platform properties push social and video performance data into Search Console's Performance, Insights, and Achievements reports.
• Owners can verify properties for domains they don't own by authorizing a connection to Instagram, TikTok, X, or YouTube.
• A help doc named "platform properties" appeared, was pulled, and is now live; the feature rolls out gradually over the coming weeks.
• This is distinct from the 2025 "Social channels" feature and the separate Search profiles.

Google Adds Product Category and Sale Duration to Merchant Listing Docs
Source: Search Engine Roundtable

• Google's Merchant listing structured data doc now supports Product.category with Text and CategoryCode types, matching the product_type and google_product_category feed attributes.
• A new Sale duration section covers validFrom, validThrough, and priceValidUntil for setting sale-price ranges.
• Google says the markup can go on either Offer or PriceSpecification nodes and aligns with the sale_price_effective_date feed attribute.