EU Ruling Looms, AI Mode Ads on 1 in 3 Queries

Estimated reading time: 28 minutes

Regulators, ad surfaces and measurement gaps dominate this week’s search marketing news. The European Commission is expected to rule that Google favoured its own services in search, SE Ranking found AI Mode showing ads on nearly a third of commercial queries, and new research ties wider ChatGPT access to a 9% drop in traditional search. Add Google’s new AI disclosure labels for ads, a Performance Max control advertisers have requested for years, and a GA4 quirk that’s quietly miscounting your AI traffic, and there’s plenty to act on.

EU Set to Rule Google Favoured Its Own Services in Search

The European Commission is expected to rule next week that Google illegally favoured its own shopping, travel and other specialised services over rivals in search results, according to a Financial Times report citing internal Commission documents. The decision would come under the Digital Markets Act, with fines expected to run to hundreds of millions of euros across two decisions, plus daily penalties if Google fails to comply with parts of the orders within 60 days. The Commission is also expected to decide on search data access obligations at the same time.

Google isn’t waiting quietly. On the day the story broke, it published a blog post arguing the DMA decisions “risk undermining vital privacy and security guardrails” for European users, which tells you how much it expects to lose.

The stakes for marketers are practical, not abstract. Google’s vertical results occupy some of the most valuable commercial real estate in search: flights modules, hotel blocks, Shopping units. If the Commission forces changes to how those surfaces treat rivals, visibility could reshuffle across exactly the high-intent queries where comparison sites, travel platforms and retailers fight hardest. The UK sits outside the DMA, but remedies of this size rarely stay neatly inside EU borders; rebuilding a results page per jurisdiction is expensive, and the CMA is watching the same behaviour. If your clients compete with a Google vertical, start benchmarking those SERPs now so you can spot the changes when they land.

Google Images Turns 25 With a New Homepage and AI Generation in AI Overviews

Google Images is 25, and Google marked the birthday with two changes. The clean search box on the Images homepage is being replaced by a browseable, real-time gallery, personalised to your interests when signed in, with saved collections appearing as tabs. That’s rolling out on US desktop in English over the coming weeks. I’ll admit a pang of nostalgia for the empty search box, but the direction is clear: Google wants Images to be a discovery feed you browse, not just a query tool you visit.

The second change matters more for search marketers. AI Overviews can now generate images from text prompts using Google’s Nano Banana model, so a search asking to visualise a nautical-style bedroom returns a generated picture, with follow-up prompts to refine it, and side-by-side comparison images are supported too. It rolls out in English across all regions that support image creation in AI Mode.

Every generated image is a query that no longer needs an image from the web. Interiors, fashion inspiration, product mock-ups and how-would-this-look searches have been reliable image traffic for retailers and content sites; some of that demand will now be satisfied by an image Google invents on the spot. Notable too: when visual results came to AI Mode last year, Google told SEJ its systems didn’t explicitly distinguish real photos from AI-generated ones. Image SEO isn’t dead, but the surface it plays on is shrinking at the edges.

Google: Unindexed Pages Are Often a Quality Verdict, Not a Bug

Why won’t Google index your pages? On the latest Search Off the Record podcast, John Mueller and Martin Splitt gave the answer many site owners don’t want to hear: when systems have serious concerns about a site’s overall quality, Google deliberately crawls less and indexes less. A pile-up of “crawled – currently not indexed” pages is sometimes the visible symptom of that judgement, not a technical fault you can patch.

Mueller’s advice was to stop treating it as a bug to fix and step back to assess quality with an outsider’s eyes, which is hard, he acknowledged, “because a lot of times it’s your website and it’s your baby”. He was careful not to condemn AI-generated content outright, but his description of the failure mode is blunt:

“Anyone could have written this. This tells me nothing.”
– John Mueller, Search Advocate, Google, on how low-value AI content reads

This connects to a pattern we’ve seen all year: indexing has become a quality filter, and generic AI output is exactly what it filters. If a client’s page indexing report shows a swelling not-indexed pile with no technical cause, audit what’s unique on those pages before touching the sitemap. Pages that tell Google nothing new don’t earn a slot, however clean the markup.

Mueller Untangles the A/B Testing Rules

Two Mueller clarifications on A/B testing landed this week, and together they draw a clearer boundary than Google’s documentation manages. First: running an A/B test for a long time, even a year, won’t earn an SEO demotion, despite official guidance implying prolonged tests might be treated as deceptive. Second, the caveat: if your variants are “significantly different” from each other, the differences can show up in search results, and depending on setup, one version or the other may be what Google indexes.

The practical reading for CRO and SEO teams sharing a site: test length isn’t the risk, variant divergence is. Keep test variants recognisably the same page (same core content and intent), don’t serve Googlebot a special version, and accept that whichever variant Google encounters may be the one that ranks while the test runs. If a test radically restructures a commercially important page, agree upfront which version you’d be happy having indexed, because you may not get to choose.

Wrong Lastmod Dates Are Worse Than None

A short, sharp one from Google’s Gary Illyes: if the lastmod dates in your XML sitemap are wrong, you’re better off removing them entirely. “At least you save a few bytes,” he added on Bluesky, with characteristic warmth.

The offender he’s describing is common: CMS and plugin setups that stamp every URL with today’s date on each sitemap regeneration, or that update lastmod when a footer or sidebar changes rather than the content itself. Google uses lastmod as a crawl-scheduling hint, and a sitemap that cries wolf on every URL teaches Google to ignore the signal on your site. Worth two minutes this week: open a client sitemap and check whether the dates vary sensibly. If every page claims it changed today, fix the generation logic or strip the field.

Google Calendar Plugs Into AI Mode's Personal Intelligence

Google Calendar is now connected to Personal Intelligence in AI Mode, announced by Search VP Robby Stein. It’s the third connected app after Gmail and Photos, and the first that AI Mode can write to as well as read: it can add invites and meetings directly, and its answers now account for what’s already on your schedule. US only for now, with more countries promised.

For search marketers, the interesting part isn’t the diary admin, it’s the personalisation compounding. iPullRank’s research in May found that simply connecting Gmail changed which brands appeared in AI Mode responses to identical prompts. Calendar adds a time-based variable on top of an interests-based one: ask for restaurant suggestions and the answer can now depend on whether tonight is booked.

Two people typing the same words, getting different brands. That’s the trajectory, and it quietly breaks the idea of a single verifiable results page. Rank tracking and AI visibility tools sample from clean accounts; your actual buyers search from accounts loaded with personal context. The gap between what your tools see and what customers see just widened again, and it will keep widening with every app Google connects.

Gemini in Chrome Rolls Out Across the UK

One for the home market: Gemini in Chrome is now rolling out to UK desktop users, with iOS following next month. The browsing assistant summarises long pages, compares information across open tabs, remembers context from past conversations, and hooks into Calendar, Maps, Gmail and YouTube without leaving the page. It can even rework images on the web via Nano Banana 2 prompts.

UK marketers should treat this as a new consumption layer between your content and your audience. When a visitor can summarise your 2,000-word guide into four bullet points without scrolling past your CTA, page structure starts doing commercial work: clear headings, self-contained sections and explicit product facts are what survive summarisation. It’s the same lesson AI Overviews taught, now sitting inside the browser on every site you own. Try it on your own key pages before your clients’ customers do; seeing what Gemini keeps and discards from a landing page is a usefully humbling exercise.

AI Mode Now Shows Ads on Nearly 1 in 3 Commercial Queries

AI Mode went from answer engine to ad surface in under a year, and SE Ranking has now measured how far that’s gone. Across 50,032 commercial keywords in 20 niches, 29.45% returned a text ad in AI Mode, and the researchers call that a floor rather than a ceiling given how volatile the surface is. When ads appeared, 71.1% of the time they came as a pair, meaning you’re usually placed directly beside a competitor.

CPC predicts placement better than anything else tested. Ad presence climbed from 24.33% on sub-$2 keywords to 53.56% at $10 and above, while search volume and keyword difficulty showed no direct correlation. Niche matters enormously too: Pets returned ads on 72.38% of keywords, Healthcare on just 2.64%.

The finding that should reframe budget conversations: buying an ad won’t get your site cited. In 88% of ad keywords, the advertiser’s domain wasn’t among the sources AI Mode cited, and for roughly 85% of ad keywords the advertiser didn’t appear in organic results at all. Paid presence and organic AI visibility are separate games on the same pitch, and one doesn’t purchase the other. Sort your commercial keyword set by CPC to map where AI Mode ads will hit your clients first, and set expectations early that citations have to be earned through the organic route.

ChatGPT Access Linked to a 9% Drop in Traditional Search

Researchers at Bocconi University analysed US desktop clickstream data from Comscore and put numbers on two questions the industry keeps arguing about. How often does ChatGPT send people to websites? In 5.2% of sessions, against 31.1% for Google searches. And what did wider ChatGPT Search access do to search behaviour? Reduced weekly traditional search queries by 9.4%, deepening to 17% after 20 weeks.

The composition of the loss matters as much as the size. Informational searching took the hit: academic referrals fell 32.8% and reference 26.5%, while transactional and recreational searches barely moved. And the clicks ChatGPT does send go to a different web than Google’s: more reference, academic, developer and SaaS destinations, with ad-supported sites receiving 27.6 percentage points less of the referral mix. The authors are careful to claim only a change in traffic allocation, not publisher revenue or welfare effects.

Put this beside Google’s record World Cup usage from last week and the picture sharpens rather than contradicts: the informational middle of search is migrating to assistants, while real-time, transactional and navigational demand stays put. For content strategy, that’s an argument for weighting commercial and experience-led pages over encyclopaedic explainers that an assistant can answer without you.

Google Is Now AI Mode's Second Most-Cited Domain

Profound’s tracking found AI Mode increased citations to google.com by 8.4x in about two months, making Google itself the second most-cited domain in AI Mode answers. The growth came almost entirely from Google Business Profiles and Product Knowledge Panels appearing as inline cards, especially on local and shopping queries.

For local searches, that means a business’s hours, photos, reviews and location can now render inside the AI answer before anyone reaches the website. Product queries about comparisons, compatibility or specifications increasingly surface Google-hosted product panels instead of links to the brand or retailer.

It’s hard to miss the irony of this landing the same week the EU prepares a self-preferencing ruling; AI Mode citing Google as an authority on businesses is exactly the pattern regulators are circling. But while that plays out, the practical response is unglamorous: your Google-hosted presence is now part of your AI visibility. Stale opening hours, thin photos or unanswered reviews on a Business Profile aren’t just Map Pack problems any more; they’re what AI Mode shows instead of your site. Fold GBP hygiene and product feed completeness into AI visibility work, because Google is increasingly answering with its own copy of your data.

ChatGPT Callers Qualify as Leads More Often Than Any Other Channel

Invoca’s new benchmark report, built on over 70 million tracked calls across 10 industries, is the first to break out phone calls referred by generative AI. The headline: calls from ChatGPT qualify as sales leads 49% of the time, roughly 10 percentage points above the cross-channel average and 6 points above Google Business Profiles. Once qualified, though, they convert at 40%, fractionally below the 42% average.

The plausible story is that people who research with an assistant first arrive on the phone already filtered: they’ve had their comparison questions answered and are calling to act. The assistant does the qualifying; the human closes at a normal rate.

I want to believe the 49%, but the report earns some caution. Invoca doesn’t publish how many ChatGPT-referred calls sit behind the figure, saying only that generative AI volume remains very low, and rates from small bases wobble. Nor does it explain the attribution trail connecting a call to ChatGPT, and Gemini, Claude and Perplexity aren’t measured at all. Treat it as an early, directionally interesting signal: AI referrals are small but unusually qualified, which matches the Similarweb finding last month that ChatGPT recommendations made a site visit 2.5x more likely. If you run call-heavy clients, get AI referral sources labelled in your call tracking now so you have your own base rates by the time volume arrives.

Pricing Pages Are Where AI Agents Give Up on B2B Sites

A study on Search Engine Land tested how AI agents handle real buyer tasks across 100 B2B products, running each task five times to account for LLM randomness. Agents were asked to find pricing and features, integrations, and security and compliance information, starting from a web search with no links provided.

Most sites passed most tasks. The consistent breaking point was pricing: when a vendor’s pricing wasn’t findable or fetchable on its own site, agents went and got it from third-party sources instead, then cited those sources in their answers. The authors’ framing is memorable: agents turn websites from showrooms into barcodes. An agent doesn’t admire your design; it fetches facts, and a page can persuade every human who reads it while failing the machine that increasingly reads first.

That’s a problem when third-party pricing data is stale or wrong, which it often is for B2B SaaS. The fix list is refreshingly concrete: publish pricing (even ranges or starting-at figures) in crawlable HTML text, keep integrations and security pages equally plain, and check what an agent actually retrieves by asking one. With Salesforce reporting a fifth of sales involving agents and bot visits now outnumbering human ones on the web, “we don’t publish pricing” is turning from a sales strategy into a visibility leak.

GA4's New AI Assistant Channel Undercounts Your AI Traffic

In May, GA4 added a native AI Assistant channel that automatically tags recognised assistant referrers, and it reached most properties by early June. Genuinely useful; no more homemade regex. But Itamar Blauer’s investigation shows why reading that channel on its own will undercount your AI traffic every time: the same source, chatgpt.com, can appear in three channels at once.

chatgpt.com / ai-assistant lands in the AI Assistant channel (the tagged slice)
chatgpt.com / referral holds untagged sessions, plus everything from before the channel switched on for your property
chatgpt.com / (not set) drops into Unassigned, usually because ChatGPT’s in-app browser strips the referrer medium

On top of the fragmentation, Google’s recognised-platform list keeps shifting. Claude was quietly dropped between launch and June, and Perplexity has never been on it, so its traffic sits in Referral indefinitely. Month-on-month comparisons that reach back before the rollout are comparing tagged data against untagged data, which is an artefact, not a trend.

The fix is unglamorous: report AI traffic by source, not by the native channel. Aggregate across all three buckets per assistant, check Unassigned, and re-read Google’s channel definitions on the day you publish anything. Given Similarweb pegs ChatGPT referral conversion around 7%, ahead of organic search, this is a small, high-intent channel that deserves accurate counting.

TikTok Goes After AI Spam Accounts in Health, Finance and Politics

TikTok will start testing improved detection for accounts dedicated to AI-generated spam, focusing on politics and current events, financial advice, and medical content, the categories it says carry the most risk to public trust. Detection runs at the account level, and the company claims over 86 million fake accounts dismantled in the first three months of the year. TikTok also took a steering committee seat at the C2PA, the coalition behind the content provenance standard, having already tagged over 3 billion videos as AI-generated.

The pattern across platforms is now unmistakable: YouTube tightened monetisation on repetitive inauthentic video last summer, Meta followed, Google published its account-clustering spam research (covered here last week), and now TikTok. Detection everywhere is moving from judging individual assets to judging publishing patterns. For brands producing AI-assisted video in health or finance, the safe side of the line is the same on every platform: on-camera humans, genuine expertise and varied production. Accounts that look like content farms will increasingly be treated as such, whatever the quality of any single clip.

Google Ads Now Requires AI Disclosure Labels

Google Ads is introducing AI transparency labels across Search, YouTube and Discover. A new “How this ad was made” section in the My Ad Center panel will state whether generative AI was used to create or edit an ad, and the disclosure requirement extends to advertisers using third-party AI tools, not just Google’s own.

The mechanics split by tooling. Ads built with Google’s generative AI advertising features get the disclosure automatically; creative made or edited with outside AI tools needs the advertiser to flag it through a new control. The rollout runs through July across Google Ads, Display & Video 360, Campaign Manager 360, Merchant Center and Ads Editor. In jurisdictions where law requires it, including the EU, India and New York State, labels may appear directly on the ad itself rather than tucked behind the menu.

Google puts responsibility for determining when AI use needs disclosing squarely on advertisers. For agencies, that creates a genuinely new workflow question: do you know whether the creative your client supplied last month touched an AI tool? Most production pipelines don’t record that today. Add an AI-usage field to creative briefs and approval checklists now, and for accounts running in on-ad label territory, start watching whether labelled creative performs differently. Nobody knows yet if disclosure moves click-through rates; the advertisers who instrument for it first will.

Performance Max Tests Opt-Outs for Search Partners and Display

Some advertisers are seeing a new Partners (Alpha) setting in Performance Max that lets them choose whether campaigns run on the Search Partners network and the Google Display Network. Until now both were automatically included, with no off switch.

That’s a control PMax has never offered.

The alpha label means limited availability and no promise of a wider rollout, but the direction is consistent with what we’ve seen from Google all summer: channel-level reporting arrived first, then the Smart Bidding predictability change, now network opt-outs in testing. The black box keeps growing windows. If the setting appears in your accounts, the obvious test writes itself: benchmark ROAS and CPA with partners and display excluded against the all-networks baseline, especially for clients who’ve long suspected junk placements were padding PMax volume. Document what you find; alpha features have a habit of vanishing and returning changed.

ChatGPT Ads Gains Location and Audience Exclusions

Third week running with ChatGPT Ads news: OpenAI has added the ability to exclude ads from serving by geo-location and by audience list in Ads Manager. Exclusions sound minor, but they’re the feature that separates a broadcast toy from a targetable platform; being able to suppress existing customers or out-of-territory impressions is basic hygiene for lead-gen and retail budgets alike.

The cadence is the real story. Audiences arrived a fortnight ago, exclusions this week, and the self-serve Ads Manager only opened in beta in May. OpenAI is speed-running a decade of ad platform development, and the gaps are closing fast enough that “we’ll wait until it matures” arguments may not survive many more weekly updates. Early testing at modest spend still looks like the sensible posture for accounts whose audiences skew toward assistant-first research.

Local Inventory Ads Will Switch On by Default

Google has emailed advertisers to say local inventory ads will be turned on by default in Shopping campaigns. Advertisers who don’t want local inventory serving have until 31 August to adjust, using the inventory filter rather than the old Local products setting.

Diary this one for any retail client with physical stores, and for pure-play ecommerce accounts too: defaults have a way of flipping quietly during holiday build-up and surfacing in performance reviews later. If local inventory feeds aren’t accurate, in-store availability promises in ads become a customer experience problem, not just a wasted click. Check which accounts have local feeds connected, decide deliberately whether the default suits each one, and set the filter where it doesn’t. Fifteen minutes now beats an awkward September conversation.

Shopping Ads and Free Listings Policies Are Merging

Google will consolidate its Shopping ads and free listings policies into a single Shopping policy document in September. Some clauses will still apply only to ads and others only to free listings, but they’ll live in one place. The same week brought an expansion of Shopping ads and free listings eligibility to more countries.

Housekeeping on Google’s side, mildly useful on ours: one policy document means one review when a merchant gets a disapproval, instead of cross-referencing two near-identical texts to work out which rule fired. Make a note in your Merchant Center troubleshooting playbook for the autumn, and remember that policy consolidations sometimes smuggle in small wording changes; skim the unified document when it lands rather than assuming it’s a pure merge.

Ginny Marvin Spells Out AI Search Ad Eligibility

Google Ads Liaison Ginny Marvin used the latest Ads Decoded newsletter to answer the questions advertisers keep asking since Google Marketing Live. Nothing new was announced, but the clarity is useful. To be eligible for ads in AI Overviews and AI Mode, advertisers need Google’s AI-powered targeting: Broad Match, or keywordless targeting via AI Max, Performance Max, Shopping campaigns and Dynamic Search Ads as they transition to AI Max, with Smart Bidding as part of the package. Marvin noted the relevance bar is higher in AI Search, with ads matched against both the user’s query and the content of the AI response.

She also expanded on Qualified Future Conversions, the predictive metric estimating conversions up to 180 days after an ad interaction. Google’s stated rationale: standard attribution windows capture roughly 70% of conversions for standard campaigns, about 50% for PMax and only 40% for Demand Gen, and QFC exists to fill that gap with forecasts built from early signals like branded searches.

Read plainly, the message is that access to Google’s newest ad inventory is conditional on adopting its automation stack, and its newest measurement asks you to trust a prediction. Both may well be worth it; neither should go unexamined. Keep incrementality testing alongside QFC rather than letting a forecast metric mark its own homework.

Video Campaign Groups Manage Reach and Frequency Across Campaigns

Google has introduced video campaign groups, letting advertisers optimise reach and frequency across multiple video campaigns while each keeps its own settings. Until now, frequency management worked per campaign, so a viewer targeted by three campaigns could see far more impressions than any single frequency cap implied.

It’s a quality-of-life feature with a real audience: brand and YouTube-heavy accounts running simultaneous flights for different products or regions. If that’s you, grouping campaigns should reduce audience fatigue and reclaim the impressions currently wasted on over-exposed viewers. Smaller accounts running one video campaign at a time can file this and move on.

Apple Maps Ads Bans Home Service Businesses

Apple has updated its Apple Maps Ads policies to prohibit home services outright: plumbing, electrical, locksmiths, HVAC, pest control, roofing and general contracting are all banned, alongside bail bonds, cryptocurrency ATMs and some medical services.

Banning an entire high-demand category from a young ad platform is a striking choice, and the likeliest explanation is the one local marketers will recognise immediately: home services is where lead-gen fraud, fake listings and licence ambiguity concentrate, and Apple has seemingly decided policing it isn’t worth the revenue. Google Ads still runs Local Services Ads with verification for exactly these businesses, which now looks like a moat rather than an overhead. If you market home service brands, Apple Maps just left the media plan; if you compete with the cowboys, a verification-gated channel remains your friend.

Strategic Direction: Where Search Marketing Is Heading

The week’s stories keep colliding in one place: who controls the surface where buyers decide. The EU is about to rule on Google preferring its own services just as Profound’s data shows AI Mode citing google.com more than almost anyone, and just as SE Ranking shows the same surface selling ads on a third of commercial queries where paid presence doesn’t buy citations. Google is simultaneously the referee, a player and the stadium, and regulators have noticed. For marketers the near-term consequence is unheroic: keep your Google-hosted assets (Business Profiles, product feeds) as accurate as your website, because they’re increasingly what the machine answers with.

The second thread is trust infrastructure hardening around AI content: disclosure labels in ads, account-level spam detection at TikTok and Google, and Mueller spelling out that generic AI pages simply don’t earn indexing. The era of consequence-free AI scale is closing from both the platform and the regulatory side, which rewards the brands that kept humans and genuine expertise in the loop.

And underneath it all sits a measurement gap widening faster than most reporting stacks are adapting: GA4 fragmenting AI referrals, call attribution guessing at assistant influence, personalised AI Mode answers that no rank tracker can see. The agencies that win the next year will be the ones whose numbers can be trusted while everyone else’s wobble.

Key Takeaways

Benchmark SERPs where clients compete with Google verticals (shopping, travel, local) before the EU ruling lands; you’ll want a before picture
Sort commercial keywords by CPC to predict AI Mode ad exposure, and set expectations that ads there don’t earn citations (advertisers weren’t cited on 88% of ad keywords)
Rebuild AI referral reporting by source rather than GA4’s native AI Assistant channel, which splits chatgpt.com across three channels and misses Perplexity entirely
Treat Google Business Profile and product feed accuracy as AI visibility work; google.com is now AI Mode’s second most-cited domain, driven by those cards
Add an AI-usage field to creative briefs and approvals now; Google’s disclosure controls roll out through July and agencies are on the hook for client-supplied creative
Set the local inventory ads filter before 31 August for any Shopping account where the new default doesn’t suit
Publish B2B pricing in crawlable text; agents that can’t fetch it cite third-party sources instead
Audit “crawled – currently not indexed” pile-ups as quality verdicts, not technical bugs, especially on AI-assisted content

Frequently Asked Questions

Will the EU ruling change how my site appears in Google?

Not immediately, and the UK isn’t directly covered by the DMA. But if the Commission forces changes to how Google’s shopping, travel and local surfaces treat rivals, visibility on high-intent commercial queries could reshuffle, and Google rarely maintains fully separate results layouts per jurisdiction. Benchmark your competitive SERPs now so changes are measurable.

Do ads in AI Mode help my brand get cited in AI answers?

No. SE Ranking found the advertiser’s domain absent from AI Mode’s cited sources on 88% of keywords showing that advertiser’s ad, and about 85% of advertisers didn’t rank organically for the keywords they bid on. Paid placement and organic citation are separate systems; citations still have to be earned with content.

Why doesn’t my GA4 AI Assistant channel match my actual ChatGPT traffic?

Because GA4 assigns channels using source and medium together, chatgpt.com sessions split across AI Assistant, Referral and Unassigned, the last usually caused by ChatGPT’s in-app browser stripping the referrer. Report by source instead, aggregate the three buckets, and remember Perplexity isn’t on Google’s recognised list at all.

What do I need to do about Google’s new AI ad disclosure rules?

Ads made with Google’s own AI tools are labelled automatically. If creative was made or edited with third-party AI tools, you must flag it via the new control, and Google places responsibility on the advertiser. Agencies should start recording AI usage on all client-supplied creative, since on-ad labels already apply by law in the EU, India and New York State.

Is AI-generated content stopping my pages from being indexed?

It can be. Mueller says Google deliberately crawls and indexes less when it has quality concerns about a site overall, and generic AI content that “tells me nothing” is a common trigger. The fix isn’t technical; add genuinely unique information, experience and evidence to the pages, or consolidate them.

Conclusion

A regulator preparing to redraw Google’s results page, an AI surface selling ads it won’t let you buy citations on, and a measurement stack that needs rebuilding underneath it all: this was a week about the plumbing of search, and plumbing rewards early attention. The actions are all datable. Benchmark competitive SERPs before the EU decision, fix the GA4 AI channel reporting before the next client review, set the local inventory default before 31 August, and get AI disclosure into creative workflows this month. None of it is glamorous; all of it compounds. That’s most weeks in search, honestly, and it’s why the teams that do the unglamorous work keep winning.

Need help adapting your search strategy for the AI era? Contact the Anicca team for expert SEO and PPC guidance.

Table of Contents

SEO & Google Search Updates

AI Search & Visibility

PPC & Paid Media

Wrap-Up