SEO Recap covering June 23, 2026

6 min read·8 stories

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)

The piece frames a measurement gap: teams can confirm rankings but struggle to track which content gets indexed and cited across ChatGPT, Claude, Perplexity, AI Overviews, and AI Mode, with supporting data spread across six to twelve non-integrating tools.
Writesonic's Sam Garg will present a system of AI agents that identifies citation gaps, prioritizes fixes, drafts updates, and verifies whether changes held after publication.
Garg outlines a four-layer agent framework (identity, knowledge, skills, loops) and notes most AI tools stop at the knowledge layer.
The session covers a citation-outreach system that surfaces opportunities and drafts outreach by 7 AM, with open-source components teams can build from.

Mark Williams-Cook argues AI assistants run traditional web searches in the background via Retrieval Augmented Generation, so the pages that rank for those hidden queries get cited; QueryFan captures those exact queries for free.
The tool generates persona-specific prompts, enriches them with AlsoAsked People Also Ask data, then sends them to GPT-5 (OpenAI Responses API) and Gemini (Google Search grounding), harvesting the webSearchQueries and web_search_call logs each returns.
The author cites PromptWatch data showing Reddit's ChatGPT citation rate fell from as high as 15% to below 2% around September 10, 2026, tracking Google's removal of the num=100 search parameter rather than anything Reddit did.
Not every prompt triggers a search: stable, in-model answers ("What do red blood cells do?") are not actionable through SEO, while fresh queries that ground are; QueryFan flags which prompts triggered searches.
The recommended workflow is a gap analysis with two outputs: content to create on your own site and placements to earn on high-authority third-party sites, since LLMs synthesize across the top results.

Similarweb found brands appearing in ChatGPT recommendations were 2.5x more likely to get a site visit within seven days than non-recommended brands.
Among users who visited a recommended brand's site, 55.9% of traffic came from branded searches, showing AI recommendations push people into traditional search.
ChatGPT-influenced visitors viewed about 12 pages and stayed 11.8 minutes on average, versus 6.5 pages and 5.6 minutes for other visitors (correlation only).
The report aligns with SparkToro's finding (Rand Fishkin) that AI tools return different recommended brands across repeated identical queries, making AI visibility unstable to plan around.
Data covers US desktop activity in finance, travel, and beauty; Similarweb plans to expand to mobile and other categories.

Mike Maynard argues good SEO informs but won't maximize GEO visibility, and a current lack of competition creates a window for proactive brands to dominate AI answers.
He separates white-hat GEO (high-quality content, FAQs, TL;DR summaries, question-based headlines, schema) from black-hat GEO (AI-generated spam, fake reviews, misleading claims, cloaking content for AI crawlers).
The piece predicts AI vendors will penalize manipulative tactics much as Google did, citing the 2006 removal of BMW's site for serving a keyword-stuffed cloaked page.
Maynard expects FAQ and TL;DR signposts to persist because they add value and won't attract penalties even if models stop needing them.
The recommendation: choose GEO techniques that deliver sustainable advantage rather than short-term boosts likely to be penalized.

Google is rolling out the Search Console generative AI (AI Overviews) performance report beyond the UK, with users reporting access in India, Switzerland, and US properties.
The report shows impressions, pages, countries, devices, and dates for URLs appearing in generative AI features in Search and Discover, but does not include clicks.
John Mueller confirmed the incremental rollout on Bluesky, saying Google is reviewing feedback along the way and asking for detailed input from users.

Google's Martin Splitt and John Mueller advised against publishing separate markdown versions for AI, saying it duplicates work and recreates HTML parsing that has been tried and tested for decades.
Splitt noted markdown lacks layouts, colors, and images, making it a poorer user experience than HTML for human visitors.
Mueller warned parallel content is hard to maintain: a broken AI-facing version can linger unnoticed because no user will report it, and automated systems may index broken text anyway.
Google compared the approach to dynamic rendering, a past stopgap that caused more debugging problems than it solved.
The takeaway: improve existing HTML pages rather than building parallel AI-specific versions.

Adam Heitzman recommends reporting organic-sourced and organic-assisted pipeline, organic CAC versus paid CAC, SEO ROI, and sales cycle/win rate rather than sessions, impressions, and average position.
A cited randomized field experiment found a 38% drop in organic clicks where AI Overviews appear, with zero-click searches rising from 54% to 72%.
Advanced Web Ranking research found informational-query CTRs fell as AI Overviews expanded, with the top four desktop positions losing a combined seven points in a single quarter.
The author suggests a three-layer report: financial outcomes on top, drivers (branded vs non-branded traffic, conversion, visibility) in the middle, and operational metrics at the bottom.
A cited example: one ecommerce client's organic traffic dropped 12% after an algorithm update while organic revenue grew 9% due to high-intent keyword focus and product page conversion work.

Google posted a new Subscription Linking policy in the Publisher Center, part of the News Reader Revenue Manager, requiring all publishers using the service to comply.
Subscription Linking lets paying readers connect publisher subscriptions to their Google Accounts so paid content surfaces in Search, Discover, and other Google products.
The document covers policy violations (with possible suspension for abuse or manipulation), eligible surfaces, abuse reporting, and adherence to Google Search content policies.

Published on June 24, 2026