Google Must Share Search Data, LSAs Join Google Ads

Estimated reading time: 29 minutes

Courtrooms and commissions set the agenda in this week’s search marketing news. The EU’s data-sharing order against Google went from expected to binding, a US federal judge threw out Google’s DMCA claims against SerpApi, and Google announced that Local Services Ads is folding into Google Ads as Performance Max pay-per-lead campaigns. Around the edges: Reddit became AI’s second favourite source, a 50,000-brand study found most ChatGPT topics still have no owner, and Meta’s crawlers took nine billion pages while sending nothing back.

Last week it was expected; now it’s binding. The European Commission has adopted two decisions under the Digital Markets Act: Google must share anonymised Search data with eligible rivals, including AI chatbots with search functions, and must open parts of Android to competing AI assistants. The Commission said Google’s voluntary data-sharing approach had failed, so it set the terms itself.

What Google has to hand over, on fair and non-discriminatory terms priced at cost recovery:

Anonymised queries, clicks, views and result positions, from both free and paid results
Metadata such as language and device type
Not included: ranking algorithms, FastSearch, account details, search histories, or rare and lengthy queries (suppressed for privacy)

Recipients can use the data to build their own retrieval and ranking systems, with grounding AI answers an approved use, but not to train general models or replicate Google’s results wholesale. There are no fines attached to this decision, though the separate self-preferencing case still carries them.

The grounding clause is the part search marketers should sit with. The quality of an AI assistant’s answers depends heavily on the search data underneath it, and until now Google’s interaction data was a moat nobody could cross. A rival chatbot that qualifies now starts from a different baseline. If Perplexity, Claude or a European challenger noticeably improves on fresh and commercial queries over the next year, this decision will be part of the reason, and brand visibility work will need to cover those surfaces properly rather than treating them as rounding errors.

Court Dismisses Google's DMCA Claims Against SerpApi

Google’s other legal setback came in a California courtroom. Chief US District Judge Yvonne Gonzalez Rogers dismissed both of Google’s DMCA anti-circumvention claims against SerpApi, the company whose scraped SERP data powers a good chunk of the rank-tracking industry. The core holding: bypassing SearchGuard, Google’s anti-scraping technology, isn’t copyright circumvention when the results being accessed contain no copyrighted content.

For plain search results, that claim is dismissed for good. For results containing licensed Knowledge Panel images, Google has 21 days to amend, and it will need to show the copyright owners authorised SearchGuard’s deployment, a fact the court found missing. Discovery is paused meanwhile.

“We’re pleased that the court rejected Google’s attempts to expand the DMCA to assert control over access to public pages.”
– Julien Khaleghy, CEO, SerpApi

This matters well beyond one vendor. Most rank tracking, SERP monitoring and a fair amount of AI grounding runs on scraped results, and Google has been fighting that on two fronts: legal (this case) and technical (the goto redirect URLs we covered a fortnight ago). The legal front just narrowed considerably, which makes the technical arms race the one to watch. Note also who else is in this fight: Reddit has its own DMCA suit against SerpApi still running, so the law here is far from settled.

Nick Fox: AI Search Sends Billions of Clicks Every Week

Nick Fox, Google’s head of Search, said on Friday that AI Mode and AI Overviews send billions of clicks to websites every week, that Search overall still sends billions daily, and that the numbers show only growth. Coming seven days after we covered the Bocconi research tying wider ChatGPT access to a 9% drop in traditional search, the timing of the pushback isn’t subtle.

Here’s the thing: both claims can be true at once. Aggregate click volume can grow while the informational middle of the query mix migrates to AI answers and specific publishers lose badly. “Billions of clicks” is an average; nobody experiences an average.

What would settle the argument is per-site data, and Google holds it. Search Console still doesn’t separate AI Overviews or AI Mode clicks from classic results, a gap SEJ catalogued in detail this week. Until that changes, treat platform-level reassurance and site-level pain as two different measurements, and keep your own first-party numbers as the tiebreaker.

Two Unconfirmed Ranking Shakes in Ten Days

Rank trackers and SEO forums lit up twice recently: once around the 11th (Barry Schwartz, noting the date, christened it the “7-Eleven update”) and again over the weekend of the 18th and 19th. Google has confirmed neither.

Unconfirmed volatility is routine, but two spikes in ten days with no core update announced is worth logging properly. Annotate both windows in your analytics now, so that if a client’s traffic steps up or down this month you can separate algorithm movement from everything else. Resist the urge to change anything on the strength of a volatility chart alone; if a genuine update is running, the picture usually clarifies within a fortnight, and Google has a habit of confirming updates only after the chatter peaks.

Crawled but Not Indexed: The Field Guide Arrives

Last week Google’s John Mueller framed “crawled – currently not indexed” as a quality verdict rather than a bug. This week Marie Haynes published the practitioner’s companion piece, drawn from a run of site owners bringing her exactly this problem. Her finding matches Google’s framing: in almost every case she examined, the stuck pages had quality issues.

Her working method is usefully blunt. Read the stuck pages next to the pages that ARE indexed and ask what the unindexed ones add that the web didn’t already have. Thin variations, near-duplicate templates and generic AI-assisted copy dominate the pile. The remedies are the unglamorous ones: consolidate overlapping pages, add genuinely unique information and evidence, strengthen internal links to pages you want indexed, and accept that some pages should simply be removed.

Two independent voices converging on the same diagnosis in a fortnight is about as clear as SEO guidance gets. If your indexing reports are swelling, the audit question isn’t “what’s broken?” but “what would make Google want these pages?”

Bot-Check Screens Can Knock Your Pages Out of Google

An ironic hazard of the anti-bot era: John Mueller explained that if your site serves “are you a bot” challenge screens to Googlebot, Google may drop those pages, and in some cases pick a different site as the canonical version of your content. From Googlebot’s perspective, every challenged page looks like the same empty interstitial, which reads as duplicate or valueless content.

The risk concentrates in CDN and WAF configurations that tightened bot rules recently, which describes half the web given how much AI crawler traffic has grown. The check is quick: run key templates through URL Inspection in Search Console and confirm Google sees your content rather than a challenge page. With bot management about to get more aggressive everywhere (see the Meta crawler story below), keep your verified-crawler allowlists current every time security rules change. Blocking scrapers is a choice; blocking Googlebot is an accident you find out about in the rankings.

Search Console Errors Aren't a To-Do List

Two related Mueller clarifications this week amount to a plea for calmer Search Console reading. First: many “error” reports aren’t real problems; the tool was never meant as a checklist where every red item needs fixing. A noindexed page reported as “excluded by noindex” is the system working, not failing. Second: the Validate Fix button simply asks Google to re-check the affected URLs; it’s useful when you’ve actually changed something, and does nothing magical when you haven’t.

I’ll confess a professional weakness here: a dashboard full of red badges triggers the tidy-up instinct in all of us, and I’ve watched hours disappear into “fixing” exclusions that were behaving exactly as intended. The better discipline is to sort GSC issues into three buckets: genuinely broken (fix), intentional (document and ignore), and informational (glance monthly). Client reporting improves too; a report that explains why 4,000 excluded URLs are fine builds more trust than one that promises to make the number zero.

Gemini Flash-Lite Starts Powering More of Search

Google is rolling Gemini 3.5 Flash-Lite into Search, alongside launching Gemini 3.6 Flash for developers. Flash-Lite is built for low-latency, high-throughput work, which Google explicitly connects to agentic search tasks.

Model plumbing rarely makes marketing headlines, but the economics deserve a note. Every time Google makes AI inference cheaper and faster, the set of queries where an AI answer is affordable expands. AI Overviews on more queries, AI Mode responding faster, agentic features running longer chains of retrieval: all of that scales with cost per call. Cheaper models are how experimental AI surfaces become default ones, so treat announcements like this as a leading indicator that AI treatment of the SERP will keep spreading, including onto query types that today still show ten blue links.

AI Mode Now Connects to Instacart, Canva and YouTube Music

The Personal Intelligence story rolls on. A week after Calendar joined Gmail and Photos as context sources, Google is now letting US users connect third-party apps to AI Mode: add barbecue ingredients straight to an Instacart cart, pull Canva templates for a flyer, or build and save a YouTube Music playlist without leaving Search. More partners are promised.

Notice the direction: these connections aren’t about informing answers, they’re about completing tasks. Search is becoming a place where things get done, with the click-through to a website increasingly optional. For brands, the strategic question shifts from “do we rank?” to “are we the connected service, the ingredient in its cart, or the citation left behind?” Grocery brands live inside Instacart’s catalogue data now; that’s a feed and partnership problem, not a content problem. Expect the connected-app roster to become a fought-over distribution channel in its own right, and expect UK availability to follow the usual few months behind the US.

Only 15% of ChatGPT Topics Have a Clear Brand Owner

Semrush and Kevin Indig tracked 1,094 US topic categories in ChatGPT from January to June, each category built from five buyer prompts (definition, comparison, alternatives, use case, buying question). They define a topic owner as a brand named in at least four of the five prompts with a five-point lead over the runner-up.

By that bar, just 15.2% of topics have an owner. The other 85% are still contestable.

Two further findings sharpen the picture. Broad SEO strength (Authority Score, organic traffic) predicted topic ownership only about half the time, so domain-level dominance doesn’t automatically translate into ChatGPT presence. And ownership, once established with a wide margin, is sticky: owners with a five-point lead kept first place in 90% of month-over-month comparisons, while narrow leads changed hands constantly.

The strategic read: AI visibility is winnable right now in a way it won’t be later. Pick the topics that map to revenue, build coverage across the whole prompt set rather than one hero query, and aim for margin, not just presence. This connects to Clovion’s churn data from a fortnight ago: brands survive follow-up questions when the model is confident who they’re for, and confidence is exactly what a wide mention lead measures.

Reddit Is Now the Second Most-Cited Domain in AI Answers

SE Ranking’s Bogdan Krupin has quantified what many of us have watched happen in the citations: Reddit’s share of AI answer sources has roughly doubled since November, from about 2.3% to 4.5%, making it the second most-cited domain overall. The SERP side moved in step: Reddit’s share of Google top-three positions hit 10.24% after the May core update, and the keywords where Reddit ranks first jumped 54% between March and May, then rose again in June.

The more interesting finding came from tracking 482 Reddit posts across 135 subreddits over six checks in three months, logging 7,307 citations. Stability is the game: posts that stayed cited through all six checks earned 83.4 citations on average against 1.6 for one-off appearances, and just 42 posts, 9% of the sample, generated 48% of all citations. Question-framed titles were disproportionately represented.

The lesson isn’t “go spam Reddit”, which ends badly for brands every time. It’s that a small set of canonical threads effectively owns your category inside AI answers. Find them, know what they say about you, contribute genuinely where your team has standing to, and build the kind of product experience that earns unprompted mentions in the threads that matter. One durable, honest presence in the right thread now outweighs a hundred fresh posts.

Meta's Bots Crawl Billions of Pages and Send Nothing Back

DataDome’s second-quarter bot analysis puts numbers on an imbalance every site owner is paying for. Agentic traffic grew 45% quarter on quarter. Meta’s AI bots grew over 163%, with its Meta-WebIndexer crawler making around nine billion requests in the quarter, apparently building a Google-style index of the web. Referral traffic sent back to the websites footing the bandwidth bill: effectively zero.

ChatGPT ran the opposite pattern: its crawl volume fell while its referrals rose, and it now accounts for 88% of all AI referral traffic. One system is taking while it learns; the other has learned and occasionally gives back.

Every crawl costs bandwidth, CDN transactions and log noise, so this is now a commercial decision, not a philosophical one. Review what AI crawlers hit your clients’ sites and at what volume, and decide bot-by-bot what earns access. Meta’s crawler offers nothing back today, but blocking it is also a bet about whether a Meta search product matters in two years. Whatever you choose, choose it deliberately, and remember the caution from the bot-screens story above: aggressive bot rules must never catch Googlebot in the crossfire.

70% of Top Retailers Are Invisible to Agentic Commerce

SALT.agency audited product pages from leading retail brands against Google’s Universal Commerce Protocol, the open standard for AI-driven shopping that Google launched in January and extended with cart support in March. The result: 70% of the top retailers tested don’t currently meet it, making their products effectively invisible to the agents that transact through it.

The protocol race has been messy. OpenAI and Stripe’s Agentic Commerce Protocol launched Instant Checkout in September 2025, then OpenAI pulled back in March to focus on product discovery while letting retailers control checkout. Google’s UCP kept building. The uncomfortable conclusion for ecommerce teams is that a transaction can now start and finish without a website visit, which quietly unwinds three decades of traffic-first ecommerce strategy.

Last week’s stuck-agents study showed agents abandoning B2B sites over unfindable pricing. This is the retail equivalent, and the remedy rhymes: complete structured product data, accurate feeds, and machine-readable availability and pricing. The retailers who treat feeds as seriously as they treat their homepage will be the ones agents can actually buy from.

AI Gets Basic Facts Wrong About Most UK Businesses

UK-specific numbers, and they’re not flattering. Vendor tests from Searchable, reported via Retail Focus and CMOTech, found AI tools returned at least one false fact about 64% of UK high street retailers, with wrong postcodes the most common error at roughly one in ten answers, even when the prompt named the town. A companion test of 165 London businesses across 13,365 questions found 93% had at least one basic fact wrong or missing, and smaller businesses fared worse than larger ones: half received a false fact against 32% of bigger firms.

Demand for these answers is real and growing fast. BrightLocal’s consumer survey has 45% of consumers using ChatGPT or similar for local business recommendations, up from 6% a year earlier.

The thin-trail problem explains the small-business skew: when a firm’s online presence amounts to a website and a Google listing, AI systems have little to learn from and fill gaps confidently. For multi-location clients, add an AI facts audit to local SEO work: ask ChatGPT, Gemini and Perplexity the questions customers ask (address, hours, services, parking) and log what comes back. It pairs with the Profound finding from last week; Google-hosted profiles increasingly ARE the answer, so their accuracy is the ballgame.

Liz Reid Previews Search That Keeps Working After You Stop Searching

Buried in a Google I/O interview, Search chief Liz Reid described a capability that matches a Google patent published in April: autonomous, persistent search. Ask about something with no current answer (a town’s upcoming events, a stock hitting specific parameters) and the query stays live in the background, with Search notifying you when the answer finally exists. The patent, “Autonomously Providing Search Results Post-Facto”, describes six triggers for it.

That’s a standing order, not a search.

If this ships as described, it changes a foundational assumption: search sessions end. A persistent query means the moment your content becomes the answer, it can be delivered to everyone who asked before you published, which makes being first and being freshest more valuable than it has ever been. Event pages, product launch pages, and anything answering “when will X” questions should be published early and updated visibly. Speculative for now, but Reid discussed it as coming, and the patent continuation a month before I/O suggests engineering, not daydreaming.

Local Services Ads Is Moving Into Google Ads

It’s official: Local Services Ads is folding into the Google Ads platform as a specialised Performance Max campaign type built for pay-per-lead goals. Migration starts in August with a small group of US advertisers (pet care, home services, wellness, education), expands through late 2026, and reaches non-US accounts, including the UK, in 2027. The separate LSA dashboard retires after each account migrates.

What survives: keywordless targeting, Search and Maps placement only, and pay-per-lead billing for calls, messages and bookings. What changes: campaign-level Target CPA replaces per-category targets (a contractor running plumbing and HVAC in one campaign gets one blended target), BBB callouts disappear, and lead management moves inside Google Ads.

The preparation list is concrete:

Download historical performance reports before migration; lead histories transfer, performance reports do not
Decide whether multi-service accounts need splitting into separate campaigns for distinct CPA targets, weighing bidding control against thinner conversion data
Watch for the 14-day and 7-day migration notices so nothing lands mid-holiday

Put this beside Apple banning home services from Maps ads last week and the strategic picture is striking: Apple walked away from the messiest lead-gen vertical while Google is consolidating it into its main ad stack. UK agencies get a year’s notice; use it to bank the reporting history and learn the PMax-for-leads mechanics on US case studies before it arrives here.

ChatGPT Now Shows Two Ads From Different Advertisers

Week four of the ChatGPT ads build-out: OpenAI has begun showing two ads from two different advertisers in the footer of responses, where previously a single advertiser held the slot. After audiences, then exclusions, this is the inventory-density step, and it means competitive adjacency has arrived; your ad can now sit directly beside a rival’s under the same answer.

Each of these small changes moves ChatGPT ads closer to a familiar auction dynamic: more slots, more competition, eventually more sophisticated pricing. The window where inventory is cheap and competition thin doesn’t announce its own closing. If any client audience skews toward assistant-first research, the argument for a small structured test this quarter keeps strengthening; the accounts that arrive with their own performance data will out-negotiate the ones arriving on launch-day rate cards.

Google Ads Will Now Estimate the Growth You're Missing

A new beta in the Google Ads Recommendations tab estimates the clicks, conversions and revenue you’re supposedly missing because of limited budgets and low bids. It’s a forecast of foregone opportunity, rendered in your own account’s numbers.

Useful? Potentially. It’s also, structurally, a sales pitch from the party that profits when you act on it. The estimates share DNA with the Recommendations tab’s long-standing habit of equating “spend more” with “improve”, and they’ll inevitably appear in screenshots during client budget conversations. Take them as directional input, then validate independently: check impression share lost to budget, run incrementality holdouts where spend allows, and model marginal CPA rather than average. If the missed-growth number survives that scrutiny, raise the budget with confidence. If it only survives inside Google’s own maths, that tells you something too.

GA4 Linking Goes Bulk for Multi-Account Advertisers

Small but genuinely time-saving: Google Ads now supports bulk linking, letting you attach a single GA4 property to multiple Google Ads accounts in one operation instead of one link at a time. Agencies running MCC structures with many child accounts, and brands with regional account splits, are the obvious beneficiaries.

Linking errors are one of those failure modes nobody notices until conversions silently stop importing, so anything that reduces manual repetition here reduces real risk. Worth a sweep while you’re in there: audit which client accounts still have missing or stale GA4 links, fix them in bulk, and confirm conversion import settings survived the change. Ten minutes of housekeeping against a class of bug that has ruined many month-end reports.

Keep AI Labels Out of Image Corners, Says Google

A follow-up to the AI disclosure rollout we covered last week: Google has updated its AI content label documentation to specify that self-applied labels on image assets shouldn’t sit in the very corners of the image. Corners are where Google’s own overlays and interface elements land, so labels placed there risk being obscured, which defeats the disclosure.

It’s a small spec detail, but it’s exactly the kind that creative production teams need in their templates now rather than after a batch of assets bounces. If your studio is building labelled AI-assisted creative for accounts running in the EU, India or New York State, where on-ad labels carry legal weight, update the design templates this week: label inset from the corner, legible at rendered size, and consistent across the asset set.

Merchant API Deadline Nears, Extension Form Opens

The Content API for Shopping shuts off on 18 August, replaced by the Merchant API, and Google has now opened a form for requesting a deadline extension. If any client integration, feed tool or in-house script still talks to the old API, this is the month to migrate or to file for the extension, not the month to discover the dependency.

The related warning is stranger: Google says merchants should NOT uninstall and reinstall the Google & YouTube app on Shopify during the transition. Google is handling the API cutover inside the app automatically, and a reinstall mid-transition can currently break things in ways Google and Shopify are still patching. If a client’s Shopify store misbehaves this month, resist the classic turn-it-off-and-on-again instinct for this one app; check Google’s status guidance first.

Don't Reinstall the Google & YouTube Shopify App

Expanding on that Shopify warning, because it cuts against instinct: Google’s explicit guidance is do not uninstall and reinstall the Google & YouTube app while the Content-to-Merchant API transition runs. The app handles the migration itself; a reinstall can orphan the store’s connection state, and Google says it’s still working with Shopify on fixes for stores that already did it.

Brief a note to any team managing Shopify clients, including support staff who reach for reinstalls as a default troubleshooting step. If product sync issues appear, document them, check the transition status inside Merchant Center, and raise with Google support rather than cycling the app. This has the classic shape of a temporary migration hazard: awkward for a month, forgotten by autumn, but capable of breaking product visibility mid-summer-sale if handled on autopilot.

Performance Max Now Writes Ad Text From Your Landing Pages

Google quietly updated its Performance Max documentation: the Text customization section now states that Google AI generates headlines and descriptions from your landing page content, domain, and current ads. Not “may generate”, not “if enabled”: the documented behaviour is that PMax writes copy from what it finds on your pages.

That turns landing page copy into ad copy inputs, whether or not your copywriter intended it. Claims, offers, prices and phrasing on the page can resurface as headlines, so a stale promotion on a landing page can become a live headline promising it. Two actions follow: audit what auto-generated assets PMax is actually serving (the asset report shows them), and bring landing page copy under the same compliance review as ad copy for regulated clients. This also connects to Ginny Marvin’s message from last week; Google’s AI Search ad eligibility runs through exactly these AI-generated text systems, so opting out is increasingly opting down.

Product Feeds Should Describe Problems, Not Just Products

A strong companion to the agentic commerce story, from PPC Hero. The State of PPC 2026 survey found 54% of 1,306 professionals name data errors and missing product information as their biggest feed challenge, a number that hasn’t moved in years. The author’s framing for the shift underway: performance marketing is moving from B2C to “B2R”, business to robot, where attribute completeness, feed consistency and data accuracy decide whether products surface in Shopping, PMax, Gemini and Perplexity, not creative flair.

The immediately useful detail is Google’s new Conversational Attributes in Merchant Center: six fields including Question & Answer, Document Link and Popularity Rank, built to help AI understand and surface products. Early adopters are quietly building an advantage while most feeds don’t populate them at all.

The prioritisation advice is sound too: get 100% of products listed and eligible first (roughly 80% of the available gain), then work the underperformers, products with high clicks but low conversion, where a vague title or missing attribute is usually what’s leaking the sale. Add the conversational fields to your feed roadmap this quarter; between UCP, agents and AI shopping surfaces, the feed is becoming the storefront.

Strategic Direction: Where Search Marketing Is Heading

The week’s defining theme is forced openness. Regulators are prying Google’s search data loose for rivals, a court just blunted its main legal weapon against scrapers, and Meta is busily building its own index of the web at publishers’ expense. The moat that made Google the only surface worth optimising for is leaking at several points at once, and the practical consequence for marketers is diversification: the grounding data behind rival AI answers is about to improve, so visibility strategy needs to treat ChatGPT, Perplexity, Meta and whatever the EU decision enables as first-class surfaces, measured properly rather than glanced at.

Meanwhile Google is redesigning search around persistence and completion. Connected apps let AI Mode finish tasks, Liz Reid is previewing queries that outlive the session, and Local Services Ads is being absorbed into an automation stack that writes its own ad text from your landing pages. Every one of these moves the click further from the centre of the model. The work that compounds in that world is structural: complete product feeds, accurate machine-readable facts about every location, canonical presence in the community threads AI trusts, and topic-level coverage while 85% of ChatGPT categories remain unowned.

And through it all, the fundamentals keep asserting themselves: two voices in a fortnight said the same thing about indexing, and it wasn’t about tricks. Google indexes what adds something. That rule now decides AI citations too.

Key Takeaways

Download Local Services Ads historical reports before migration reaches your accounts; performance history does not transfer to Google Ads
Complete the Merchant API migration or file the extension form before 18 August, and don’t reinstall the Google & YouTube Shopify app mid-transition
Target topic ownership in ChatGPT now: 85% of categories have no owner, and wide mention leads hold 90% of the time once established
Find the canonical Reddit threads for each client’s category; 9% of tracked posts earned 48% of AI citations, and stability beats freshness
Add an AI facts audit for multi-location clients; 64% of UK high street retailers had at least one false fact in AI answers
Decide deliberately which AI crawlers earn site access; Meta’s took nine billion pages last quarter and referred nothing
Check what Googlebot sees through URL Inspection after any bot-rule tightening; challenge screens can deindex pages or flip canonicals
Audit PMax auto-generated assets and put landing page copy under ad-copy compliance review; Google now writes headlines from your pages

Frequently Asked Questions

What search data does Google actually have to share under the EU decision?

Anonymised queries, clicks, views and result positions from free and paid results, plus metadata like language and device type, priced at cost recovery on non-discriminatory terms. Ranking algorithms and FastSearch stay private, and sensitive data such as rare queries and account details is suppressed. Eligible rivals can use it to build retrieval systems and ground AI answers, not to clone Google.

Do we need to act on the Local Services Ads migration yet?

UK accounts don’t migrate until 2027, but two things are worth doing now: download historical performance reports (they don’t transfer) and follow the US rollout from August to learn how the PMax pay-per-lead mechanics behave. Accounts get 14 days’ notice before their migration.

Does the SerpApi ruling make scraping Google legal?

No. The court decided one narrow point: bypassing Google’s anti-scraping measures isn’t DMCA copyright circumvention when the scraped results contain no copyrighted content. Google can amend the Knowledge Panel part of the case, other legal theories remain available, and Reddit’s separate suit against SerpApi continues. It’s a meaningful narrowing, not a green light.

Should we block Meta’s AI crawlers?

Treat it as a per-site commercial decision. Meta’s crawlers currently return effectively no referral traffic while consuming real bandwidth, which argues for blocking; the counterargument is positioning for whatever search or assistant product that index eventually feeds. Whichever way you go, make sure bot rules never challenge Googlebot, and revisit the decision quarterly as the data changes.

How do we get a brand mentioned consistently in ChatGPT answers?

Work at topic level, not prompt level. Cover the full set of buyer questions in a category (definition, comparison, alternatives, use case, buying), aim for a wide mention lead rather than bare presence, and reinforce the association everywhere the model reads, including the canonical community threads in your category. Most topics are still unowned, so early consistent coverage compounds.

Conclusion

Weeks like this reset the board rather than move a piece. A regulator has ordered the search moat opened, a court has protected the scrapers who measure the game, and Google’s answer is to make search more persistent, more transactional and more automated than ever. The response isn’t panic; it’s scheduling. Bank the LSA reports, beat the Merchant API deadline, choose your crawler policy, claim unowned ChatGPT topics, and put feeds and location facts under the same quality discipline as your best content. Search marketing keeps rewarding the same trait through every upheaval: teams that turn news into next week’s checklist before their competitors have finished reading it.

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