SEO Recap covering July 16, 2026
11 min read·16 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)
How to run a local GEO baseline audit before spending on AI visibility
Source: Search Engine Land
- ChatGPT recommended only 1.2% of nearly 350,000 locations in SOCi's 2026 Local Visibility Index, versus a 35.9% appearance rate in Google's local 3-pack; Gemini recommended 11% and Perplexity 7.4%.
- Business profile data was about 68% accurate on ChatGPT and Perplexity but 100% on Gemini, which relies entirely on Google Maps data.
- The author recommends testing discovery, comparison, trust, and logistics queries across ChatGPT, Perplexity, Gemini, and AI Overviews, logging mention, position, sentiment, accuracy, and cited sources from a fixed location and date.
- Fix in order: eligibility first (crawler access, NAP consistency, schema), then trust signals (reviews, engagement), then relevance (location-specific content).
- Cloudflare's default blocking of AI crawlers can quietly make a site invisible to answer engines; the author says to check robots.txt and Cloudflare settings before anything else.
Reddit leads AI search citations for multi-location brands
Source: Search Engine Journal
- Uberall data cited in the session shows one in every five off-page citations in AI answers comes from Reddit, a share growing 30% year over year.
- AI models read 5 to 16 sources before answering, with a brand's own website accounting for only about 15% of what they find; roughly 75% of businesses are absent from AI conversations about their category.
- Google has begun surfacing Reddit threads directly on business profiles next to address, phone, and hours, so unanswered local questions appear to every searcher.
- Speakers recommend auditing NAP consistency across Google, Apple, Yelp, and vertical directories first, since conflicting data makes AI pick a wrong answer or skip the business.
- Carl's Jr. drove a 176% behavioral lift and 85% lower cost per visit versus benchmark through Reddit engagement measured with Foursquare foot-traffic data.
Does a domain's search fingerprint carry into answer engines?
Source: Search Engine Journal
- The author frames a site's accumulated per-domain profile (links, Core Web Vitals, E-E-A-T signals, freshness, schema) as a "fingerprint" and asks whether it persists into AI answer surfaces.
- Google says AI Mode and AI Overviews run on its core ranking systems, and its site reputation abuse guidance shows authority is assessed and retained at a granular, per-section level.
- Microsoft is similarly integrated and unusually public, documenting IndexNow (a pipe for pushing freshness signals) and Web IQ for scoring the web underneath its answers.
- ChatGPT's search runs substantially on Bing's index, but whether a persistent per-domain reputation survives on OpenAI's side is genuinely opaque and undocumented.
- The author says there is no clean public evidence that models accumulate a native domain-level fingerprint in their weights, and warns against confidently claiming they do.
Specific, insightful writing may earn more AI citations
Source: Search Engine Journal
- A Bluesky post by @danabra.mov reported that long, specific articles began getting summarized and occasionally cited by Claude within about a year of publication.
- Other writers echoed that even not-especially-insightful but specific content got pulled into LLM outputs within about six months.
- The author argues keyword-first SEO is outdated and that focused, on-topic writing (ruthlessly cutting tangents) matches how both readers and LLMs process content.
- Google's John Mueller reposted the thread with the comment "Make more insightful & useful stuff."
AI in Search / AI Overviews
Google AI Mode adds Instacart, Canva, and YouTube Music integrations
Source: Search Engine Land
- Google is rolling out connected app integrations in AI Mode for U.S. users, letting them link services like Instacart, Canva, and YouTube Music to act on AI-generated responses.
- Google's examples: adding barbecue ingredients to an Instacart cart, pulling Canva flyer templates, and saving a curated playlist to YouTube Music, all without leaving Search.
- Letting users complete tasks inside AI Mode could reduce the clicks, visibility, and control brands get after a search.
- Google says more app partners are coming.
Google starts rolling out connected apps in AI Mode Search
Source: Search Engine Journal
- Google is bringing connected apps to AI Mode in U.S. Search this week, letting users send searches into Instacart, Canva, and YouTube Music; the feature previously lived only in the Gemini app.
- The integrations are built on Model Context Protocol, the open standard for linking AI systems to outside tools.
- Google's post doesn't tie the app set to search rankings or explain how a company becomes eligible, so being on the list becomes a second visibility question alongside whether a page ranks.
- SEJ notes connected apps differ from Personal Intelligence: the new feature sends requests out to third-party services, while Personal Intelligence uses your connected Google apps as context.
Technical SEO
Google's Job Indexing API confirms receipt, not indexing
Source: Search Engine Land
- The author says the Indexing API only supports pages with JobPosting structured data or livestream BroadcastEvent markup, not general URLs, and a 200 response confirms only that Google received the notification.
- Google's documentation says it "may" recrawl or drop a URL after a request; submitted is not indexed, and getMetadata can return a 404 when an account is still operating in Google's sandbox.
- The default quota is 200 requests for onboarding, with additional access requiring an approval form Google says takes 2 to 3 weeks to review.
- The author and Alexander Chukovski report that legitimate job boards submitting quota-increase requests received no response for six months to a year, suggesting approvals have effectively stalled.
- The author built a free Job Indexing Health Check on SEOJobs.com to validate schema, API responses, and Search Console status.
Google: low-quality AI content can trigger "crawled - currently not indexed"
Source: Search Engine Roundtable
- On the Search Off the Record podcast, John Mueller said that when Google's systems have strong concerns about a site's overall quality, they crawl and index fewer pages, producing "crawled - currently not indexed" or "discovered - not indexed" statuses.
- Mueller said these aren't technical problems to fix directly; a broad pattern of non-indexing with no technical cause should prompt a step back to assess site quality.
- He clarified that not all AI-generated content is bad, but content where "anyone could have written this" and nothing is unique or valuable gives Google little reason to index it.
Google: drop lastmod dates if they're incorrect
Source: Search Engine Roundtable
- Gary Illyes said on Bluesky that sites are "probably better off without the lastmods" when the dates are wrong, adding you at least save a few bytes.
- The advice followed a question about a bug that produced unintentionally incorrect lastmod dates, given Google's largely binary trust of the field.
- Google has long said it trusts lastmod only when used correctly, meaning updated only for significant content changes, not copyright or trivial edits.
Organic Search & Algorithm Updates
EU orders Google to share search data with rivals and AI search tools
Source: Search Engine Land
- The European Commission ordered Google to share anonymized search data with eligible third-party search providers starting January 2027 under the Digital Markets Act.
- Google must share the same data it uses to optimize its own search, subject to a multilayer anonymization process, with a fair pricing formula and transparent access.
- AI chatbots offering search functionality qualify to receive the shared data, not just traditional search engines.
- Google must also give rival AI assistants better access to Android features available to Gemini, with those changes expected to begin July 2027.
- The specification proceedings clarify compliance and impose no fines, and the measures can be amended based on market and privacy assessments.
The four types of content decay and how to diagnose each
Source: Search Engine Land
- The author argues content decay now has four causes, not three: ranking decay, zero-click capture, intent drift, and demand decay, each with a distinct signature in clicks, impressions, and average position.
- In 2026, about 68% of Google searches end without a click (up from roughly 60% two years ago), and on queries with an AI Overview the top organic result loses around 58% of its clicks.
- Diagnose with Google Search Console and a spreadsheet using a three-month year-over-year comparison of clicks, impressions, and position; only ranking decay is reliably fixed by a content refresh.
- The author warns 2025 impression data is unreliable after Google removed the &num=100 parameter and disclosed a logging error that inflated impressions since May 2025; clicks were unaffected.
- For zero-click capture, the recommendation is to add original data, tools, or a point of view the answer box can't replace, and reallocate effort to nearby conversion pages when the click is gone.
Bloomberg reports Gemini 3.5 Pro is delayed over coding performance
Source: Search Engine Journal
- Bloomberg reports Gemini 3.5 Pro is months behind the June rollout Google signaled at I/O on May 19, with coding performance cited as one holdup.
- Google updated Gemini's training data late last month to improve coding, and one source called the results disappointing; a spokesperson said the company is testing 3.5 Pro with partners.
- Ten current and former employees described internal frustration that Google is losing ground to Anthropic and OpenAI on model quality.
- Google made 3.5 Flash the default model in AI Mode globally, so a delayed flagship says nothing about the answers Search generates today; no new release date has been announced.
Search Console social tracking as a shield for AI click loss
Source: Search Engine Journal
- The author argues Google's Search Console update tracking social and video post performance reframes success around impressions to mask organic click loss from AI Overviews.
- Verifying social accounts in Search Console, the author says, hands Google clean entity-resolution data linking a website, X profile, and TikTok account to one author.
- The author notes Search Console shows clicks on X posts from October 2024, and found LLMs surfacing pricing from a July 2022 student offer, arguing for a post-deprecation strategy to avoid stale misinformation.
- The recommendation: treat these Google metrics as one piece of the puzzle and use social channels to build direct audience relationships off Google's platform.
Balancing conversion pages with brand-authority content
Source: Search Engine Journal
- The author recommends splitting page purpose in a shared document: blog and informational pages build audience, product pages convert, and neither should be forced to do the other's job.
- CRO tactics like deleting copy, burying video, or split-testing without checking canonicals and meta robots can quietly damage SEO and GEO performance.
- The author suggests giving CRO and brand teams a reference guide covering page-type must-haves, off-limits folders and pages, and a plain-English tech SEO checklist.
- Practical guardrails: canonical split-test URLs back to the main page, block temporary landing pages via robots.txt, and test head scripts in Search Console to confirm rendering.
Google warns DMA rulings undercut privacy and security for Europeans
Source: Google Search Product News
- Kent Walker, Google's President of Global Affairs, said the Commission's DMA decisions risk undermining privacy and security guardrails for millions of Europeans.
- Google argues the Android ruling would grant external apps sensitive device permissions without existing vetting safeguards, and that sharing private searches with unfamiliar companies lacks adequate anonymization and user consent.
- Google says it will keep advocating for a balanced, evidence-based approach and notes the Commission's decision allows measures to be adjusted for identified harms.
AI martech roundup: OpenAI ad forecast questioned, AEO tools multiply
Source: MarTech
- Emarketer estimates the entire U.S. standalone chatbot ad market (ChatGPT, Copilot, Google AI Mode, Amazon Alexa) will total under $1 billion this year and $5.41 billion by 2030, versus OpenAI's projected $100 billion annual ad target for 2030.
- OpenAI has roughly $25 billion annualized revenue against $27 billion cash burn and is committed to $600 billion in infrastructure spending by 2030, down from an earlier $1.4 trillion plan.
- Several vendors launched AI-search visibility and answer-engine tools, including Profound FactCheck for monitoring brand representation in model outputs, SeeResponse's Answer Engine Optimization service, and Cision AI search visibility metrics in CisionOne.
- A wave of releases leaned on Model Context Protocol servers (6sense, Omneky, OneSignal, Wistia) to feed data and requests to external AI agents.
Published on