Published: June 27, 2026 Series: Search Labs AI Intelligence Briefing – published weekly by Search Labs (searchlabs.com)
A note on naming: Search Labs is an AI search consultancy. Google Search Labs is Google’s experimental search programme. This briefing covers both, because understanding one without the other leaves you optimising blind.
What Happened This Week
The week of June 18–25 had one governing signal: Google completed the handover from publishing rules to enforcing them on AI search manipulation.
Forty days after Google quietly revised its spam policies to name “attempting to manipulate generative AI responses in Google Search” as spam (May 15, 2026), it deployed a SpamBrain update that puts active detection behind those rules (June 24, confirmed complete June 26). This is not a core update. It is not targeting link spam. It is the first enforcement mechanism specifically trained against the generative search optimization (GEO) and AI answer manipulation tactics the industry spent most of 2025 building.
At the same time, Google expanded access to its Generative AI performance reports in Search Console beyond the initial UK pilot (June 23), giving practitioners their first native measurement tool for AI visibility in LLMs and generative search surfaces. The timing is deliberate or coincidental, but the effect is the same: Google is simultaneously closing ungoverned AI optimisation and opening a measurement layer for governed AI visibility.
The through-line this week is not individual feature launches. It is the shift from “AI search is coming” to “AI search has rules and now it has enforcement.”
This issue covers:
The June 2026 spam update: what it targeted, what it excluded, what recovery looks like
The May 15 AI-manipulation clause and which GEO tactics are now formally inside the spam boundary
The GSC Generative AI performance reports: what you can measure and what you cannot
The three AI visibility controls and which one actually does what you think it does
Google’s AI Mode: Gemini 3.5 Flash, Managed Agents, Antigravity, and UCP checkout expansion
What the 2026 CTR studies actually say, with methodology attached
Competitor moves from OpenAI and Perplexity that shift the strategic picture
The Findings
What did the June 2026 spam update actually target?
The June 2026 spam update began rolling out on June 24, 2026 at 09:00 PT and was confirmed complete on the Google Search Status Dashboard on June 26 at 10:00 PT. It applies globally across all languages. This is the second confirmed spam update of 2026, following the March 2026 update that completed in under 20 hours.
Google confirmed directly that this update does not target link spam, site-reputation abuse, or several other specific policy categories. It is a SpamBrain improvement: automated, binary, enforcement-focused rather than broad content-quality focused. Barry Schwartz at Search Engine Roundtable reported “heated” ranking volatility in the 24 hours post-launch and noted the update felt “a bit bigger” than March’s.
Community reports of 10–15% organic traffic drops on compliant sites circulated on Webmaster World and Search Engine Roundtable during the rollout window. These are unconfirmed signals, not verified averages. Do not use them to diagnose your situation.
Practitioner implication: If you lost visibility between June 24 and June 26, check Search Console for manual actions first. If there are none, wait for SERPs to stabilise before drawing conclusions. If your losses do not align with what spam updates target, you may be looking at residual volatility from the May 2026 core update or unrelated SERP displacement. Spam and core updates interact differently. Treat them as separate diagnostic tracks.
Which AI optimisation tactics did Google formally ban on May 15?
On May 15, 2026, Google revised the opening definition of spam in its policies documentation to read: “In the context of Google Search, spam refers to techniques used to deceive users or manipulate our Search systems into featuring content prominently, such as attempting to manipulate Search systems into ranking content highly or attempting to manipulate generative AI responses in Google Search.”
This was the first time AI answer manipulation was named directly in Google’s written rules. The transition from the Search Generative Experience (SGE) era to today’s AI Overviews and AI Mode has brought with it a new category of manipulation: tactics designed to influence large language model (LLM) retrieval and citation rather than traditional link-based ranking algorithms. These are what the May 15 clause targets.
The specific tactics now inside the spam boundary include recommendation poisoning (engineering entity associations in page text to force model citation), biased “best of” listicles built without authentic evaluation, and cloaking optimised schemas for AI crawlers while showing different content to users. These are LLM optimization strategies that treat generative AI citation as a new form of PageRank manipulation. Google is treating them the same way.
The June 24 spam update landed 40 days after this policy revision. The sequence is deliberate: publish the rulebook, then improve the detection engine. That gap will likely compress with subsequent cycles.
Practitioner implication: Any tactic that exists purely because AI Overviews exist is now formally inside the spam boundary. If you cannot defend a content decision to a human editor, it probably will not survive SpamBrain either. Audit affiliate structures, sponsored comparative content, and guest post networks before the next enforcement cycle, not after. Recovery from a SpamBrain action is measured in months, not weeks.
What are the GSC Generative AI performance reports actually showing you?
Google launched dedicated Generative AI performance reports inside Search Console on June 3, 2026, initially restricted to a subset of UK site owners. On June 23, 2026, access expanded to sites in the US, India, Switzerland, and other regions. Google’s John Mueller confirmed the rollout is incremental and that access is being expanded based on feedback, not on a fixed schedule.
The report tracks impressions, pages, countries, devices, and dates for AI Overviews, AI Mode, and generative Discover feeds. Data starts from May 18, 2026. There is no historical backfill.
Current limitations you need to understand before drawing conclusions:
There are no clicks, no CTR figures, and no average position metrics. You are looking at impression volume only. The report does not separate AI Overviews impressions from AI Mode impressions — they are grouped as “Generative AI.” An impression is recorded only when a direct, clickable link to your domain is rendered within the viewport of the generative summary. And there are known discrepancies between chart totals and table totals due to canonical URL redirect handling.
The May 18 data start date overlaps with the May 2026 core update window (May 21 to June 2), which complicates any trend analysis in the early data.
Practitioner implication: Access the report the day it appears in your Search Console. There is no historical data, so your baseline starts the moment you can see it. Screenshot the initial state, note your total Search impressions at the same point, and calculate what percentage of total impressions are coming from AI surfaces. Cross-reference against GA4 to separate zero-click AI exposure from actual referral traffic. Do not use the report in isolation to make opt-out decisions.
Which AI visibility control actually does what you think it does?
This is where most practitioners are operating on incorrect assumptions. There are three distinct controls affecting AI search visibility. They do not do the same thing.
Google-Extended (robots.txt directive): Blocks Google’s crawlers from fetching your content for raw model training. It does not prevent your content from appearing in AI Overviews or AI Mode. Blocking Google-Extended has no effect on your AI search visibility.
nosnippet / data-nosnippet (HTML meta tag): Instructs search engines not to generate text snippets or extract text from the page. This prevents extraction for AI Overviews, but it also strips snippets from standard organic listings. If you use this to opt out of AI surfaces, you damage standard organic CTR as a direct side effect.
Search Console AI Toggle: The clean, targeted option. A database-level opt-out that removes your domain from AI Overviews, AI Mode, and Discover AI features. Standard organic rankings are not affected. Active enforcement began June 17, 2026. One important distinction: this toggle does not apply to the standalone Gemini app. Opting out through Search Console does not prevent your content from being cited inside Gemini conversational responses, which are governed by separate crawling permissions.
Practitioner implication: If you are considering opting out, the GSC AI Toggle is the only control that achieves it cleanly. But measure first. The default position for most sites should be to stay in, capture your AI impressions baseline, and then make a decision based on data. Subscription publishers with high-value proprietary content where summary extraction directly cannibalises paid access are the clearest candidates for reviewing the toggle.
What is happening inside Google’s AI Mode, and why does it matter commercially?
At Google I/O on May 19, 2026, Sundar Pichai announced that AI Mode had surpassed one billion monthly active users globally in just 12 months since launch, with queries more than doubling every quarter. The default model powering AI Mode is now Gemini 3.5 Flash, also available to developers via Google AI Studio and the Gemini API.
During the week of June 18–25, two AI Mode developments emerged. On June 24, Google was spotted testing autocomplete suggestions inside the “Ask Anything” follow-up input box in AI Mode, first documented by search analyst Brodie Clark. This is the first observed integration of the enhanced autocomplete features from the I/O intelligent search box announcement.
Also expanding this week: Universal Commerce Protocol (UCP) powered checkouts, sometimes referred to as Universal Cart, within the AI Mode conversational stream. US shoppers can currently complete purchases from Etsy and Wayfair results via a native “Buy” button inside the conversational workspace, drawing directly from the Google Shopping Graph to surface product listings. Walmart, Target, and Shopify integrations are in the pipeline.
Beyond commerce, Google’s Information Agents (also called Managed Agents) are part of the same I/O rollout: agentic AI features that run background research tasks inside Search 24/7, surfacing results through Antigravity, Google’s generative UI platform for building mini-apps and interactive dashboards directly inside AI results. These represent a step change in how AI search functions: not retrieval, but continuous agentic discovery.
The ads attribution implication is significant and underreported. For impressions and clicks originating from AI Mode, AI Overviews, Google Lens, and autocomplete, Google’s documentation states that the search term shown in reporting represents “the best approximation of the user’s intent” rather than the exact query entered. This applies to Smart Bidding signals as well. Ad slots in AI surfaces are now allocated on semantic intent clusters, not keyword strings. AI Max for Search, Google’s AI-powered campaign type, is designed to operate in this environment. Keyword-level transparency is compressing across both organic and paid.
Practitioner implication: If you manage ecommerce clients, get product feeds structured for UCP and the Google Shopping Graph now rather than waiting for broader rollout. The ads attribution shift is already live. If your reporting relies on exact-match keyword data from AI Mode sessions, you are already working with approximated intent data, not query data. Smart Bidding and AI Max campaigns are better positioned for this environment than manual or broad match keyword structures.
What does the 2026 CTR data actually say, without the vendor spin?
The CTR narrative is more complex than most coverage suggests. Three primary studies produce the data most practitioners reference, and they measure different things.
Ahrefs (published February 4, 2026): Analysed 300,000 keywords using aggregated Google Search Console data, comparing December 2023 with December 2025. Position-one CTR for keywords triggering AI Overviews dropped 58%, from a baseline of 0.073 to 0.016. The compression continues across the page: position two down 50.8%, position five down 32.6%, position ten down 19.4%. This study measures CTR for the top-ranking page on AIO keywords specifically. It does not measure average traffic loss across all sites. The framing matters: 58% is not a universal average, it is a position-one measurement on AIO query types.
Seer Interactive (April 2026 update): Tracked 53 brands, 5.47 million queries, and 2.43 billion impressions from January 2025 through February 2026. Organic CTR on AI Overview queries hit a floor of 1.31% in December 2025 and recovered to 2.36% by February 2026, an 80% rebound from that floor. The recovery is real but cited brands still trail non-AIO results by 38%. Being cited inside an AI Overview delivers 120% more organic clicks per impression than not being cited on the same query. Brand visibility in AI results is the single most operationally useful variable in any of the 2026 studies.
BrightEdge (February 2026): AI Overviews appear on 48% of tracked queries, up 58% year-over-year from approximately 30%. In mid-2024, 76% of AI Overview citations came from pages in the traditional top-10 organic results. By February 2026, that overlap had fallen to 17%. Five out of six citations are now pulled from sources outside the traditional first page. The shift away from link-based ranking algorithms as the primary driver of AI citation is now measurable at scale.
The vertical breakdown matters significantly for any portfolio with mixed verticals. AI Overview prevalence in ecommerce sits at approximately 4%, with 61.5% of citations sourced from outside the top 100 organic results. Healthcare sits at 88% prevalence with 22.5% of citations outside the top 100. B2B Technology at 82% prevalence with 28.1% outside the top 100. The AIO impact on your specific client portfolio depends almost entirely on vertical and query mix.
Practitioner implication: Stop treating the 58% figure as a universal benchmark. Your actual exposure depends on vertical, query mix, AI Overview prevalence for your specific keyword set, and whether your brand earns citations. The single most important variable is citation status. Cited brands earn 120% more clicks per impression than uncited brands on the same query. That is where competitive intelligence and optimisation effort should go.
What did competitors do this week that shifts the strategic picture?
OpenAI and Getty Images (June 21, 2026): OpenAI and Getty announced a multi-year display partnership integrating Getty’s licensed content into ChatGPT’s search and discovery experiences. This is a display-only agreement: rights-cleared visual content surfaces in ChatGPT responses. It is not a model training licence. Financial terms were not disclosed.
The deal is relevant to content strategy for a specific reason: AI search platforms are building licensed, authenticated content layers. C2PA Content Credentials, the open standard for content provenance and authentication, are increasingly relevant in this environment. Visual and editorial content that carries verifiable provenance signals is gaining structural advantage in AI retrieval over unverified or AI-generated alternatives. Platforms paying for authenticated content are signalling that provenance matters to their retrieval logic, not just to their legal team.
Getty’s stock hit an all-time low of $0.58 per share on June 18, 2026, before the OpenAI announcement drove a surge back above the NYSE’s $1.00 continued-listing threshold. The deal also provides commercial momentum to Getty’s pending $3.7 billion merger with Shutterstock, which received conditional CMA approval in the UK on May 15, 2026.
Perplexity (June 19, 2026): Perplexity released its Search as Code (SaC) architecture and integrated Deep Research into its Perplexity Computer workspace. SaC allows agentic AI systems to write custom programs that orchestrate search steps natively rather than running queries serially, drawing on analyst reports and proprietary databases including CB Insights, PitchBook, and Statista directly inside research threads.
For content that wants to be cited by research-grade AI tools and LLM research workflows, the traceability layer is the strategic detail. Perplexity Computer lets users audit every claim in a generated output back to its original source. Content with clear attribution, specific sourcing, and structured factual claims performs better in this retrieval environment than content written for general audience consumption. If your clients produce competitive intelligence or original research, this is where that content earns citations beyond Google.
Practitioner implication: AI visibility is not a single-platform optimisation problem. Google, Perplexity, and ChatGPT are now distinct citation environments with different retrieval behaviours and different content preferences. Treat them as separate measurement tracks. A citation in Perplexity Deep Research and a citation in AI Overviews are not the same signal and do not come from the same content signals.
What This Means for You
This week’s events are not about individual feature launches. They are about structural shift in how AI search is governed and measured.
Google has moved from guidelines to enforcement on AI manipulation. If your content strategy includes any tactic designed specifically to game AI Overviews or AI Mode citation, that tactic now carries active spam risk. The recovery window from a SpamBrain action is measured in months. There is no shortcut back.
At the same time, Google has handed you the first native tool to measure where you actually stand. The GSC Generative AI performance reports are limited – no clicks, no CTR, no query data – but they give you something that did not exist before June 3: confirmation of whether your domain is appearing in AI surfaces at all.
The combination of enforcement and measurement creates a clear path. Earn citations through content that genuinely serves the question, structured so AI systems can extract and attribute it cleanly. Track your AI impression baseline the day access lands in your Search Console. Measure AI impressions as a percentage of total Search impressions. Stop using standard rank position as a proxy for AI visibility in LLMs. They are different signals measured in different ways, and optimising for one does not guarantee performance in the other.
The brands that get ahead of this now are the ones that will not be scrambling to explain traffic losses to clients in Q4.
What to Watch Next
A Q3 core update is the logical next event on the calendar. The May 2026 core update ended June 2. If Google holds its approximate 90-day cadence, the next core update window opens sometime before September. Prepare for content-quality evaluation signals to shift again, separately from the spam enforcement that just ran.
Watch the GSC Generative AI report for click data. The current reports show impressions only. When Google adds CTR and click data, AI visibility becomes a traffic-value KPI rather than a presence-only KPI. That is the moment the measurement conversation shifts from “are we visible” to “is that visibility worth anything.”
Watch the UCP ecommerce checkout expansion. When Walmart, Target, and Shopify go live inside AI Mode via the Google Shopping Graph, the volume of zero-click commercial sessions will increase significantly. That is the moment ecommerce SEO needs to account for AI Mode and UCP product listings as distinct conversion surfaces, not just visibility ones.
Watch Managed Agents and Antigravity. Google’s agentic AI layer is expanding. When Information Agents and Antigravity-powered generative UI move beyond US early access, the nature of what a “search result” looks like will shift again. Citation inside an agent-generated output and citation inside a static AI Overview are not the same thing and will not require the same content signals to earn.
Search Labs AI Intelligence Briefing is published weekly by Search Labs (searchlabs.com). Search Labs is an AI search consultancy. Google Search Labs is Google’s experimental search programme. This briefing covers developments across both.
