AI Traffic Gets Its Own GA Channel & Ahrefs Proves Schema Doesn’t Move AI Citations
Estimated reading time: 25 minutes
A heavy news week for search marketing. Google Analytics finally treats AI assistant traffic as a first-class channel. Ahrefs published a 1,885-page study finding schema markup doesn’t move AI citations. Conde Nast’s CEO told staff to plan as if search traffic will be zero. Add Kevin Indig’s research showing AI engines disagree on which sources to cite 97% of the time, Google quietly changing search terms reporting for AI queries, and Adobe data showing AI-referred retail conversions are now up 393% year on year, and there’s plenty of search marketing news to work through.
Table of Contents
SEO & Algorithm Updates
Google Analytics Adds an “AI Assistant” Default Channel
Ahrefs Tests 1,885 Pages: Schema Markup Doesn’t Move AI Citations
Google Search Ranking Volatility Spikes Across Mid-May
Marie Haynes: “Crawled, Currently Not Indexed” Numbers Are Climbing
Google Discover Publisher Pages Get Links and Featured Posts
AI Search, GEO & Measurement
Inside ChatGPT Search: Web.run, Fan-Out Queries, and a 20% Drop in Cited Domains
Conde Nast CEO: “Plan as if Search Is Zero”
AI Visibility Isn’t One Problem. It’s Three Layers
Adobe Q2 2026: AI-Referred Retail Conversions Up 393% Year on Year
Google’s ALDRIFT: AI Answers That Sound Right and Actually Are
The Consensus Gap: Only 2.37% of AI Citations Overlap Across Engines
AI Overviews Are Surfacing Negative Reviews Users Didn’t Ask For
iPullRank’s “Cloaking for LLMs”: A Sharp Provocation
PPC & Paid Media
Google Quietly Changes Search Terms Reporting for AI Queries
Microsoft Advertising Extends LinkedIn Profile Targeting to CTV
Google Ads Teases Gemini-Powered Dashboards
Industry & Tooling
Yoast Launches an AI Content Planner Inside Premium
Wrap-Up
Strategic Direction: Where Search Marketing Is Heading
Google Analytics Adds an “AI Assistant” Default Channel
A single new line in Google Analytics that took most teams two years longer than it should have. On 14 May, Google rolled out an “AI Assistant” channel group to GA4. Traffic from ChatGPT, Gemini and Claude now lands in its own default channel rather than getting bucketed into Referral. The technical details: medium becomes ai-assistant, campaign becomes (ai-assistant), and the channel group is AI Assistant.
If you’ve been hand-rolling regex filters to pull AI traffic out of Referral over the past year, those can be retired. The new channel is a default on every property, no opt-in required. Comparison reports against organic, paid and direct become a one-click exercise rather than a series of explore-tab workarounds.
A few caveats to flag. Perplexity isn’t in the supported source list as of the May announcement, which is an odd omission given Perplexity’s share of AI search referral volume. Google hasn’t confirmed whether other AI engines (Anthropic Claude direct, Mistral, Copilot) will get added later or whether you’ll need to keep custom filters running for those.
What this changes practically for agency reporting. You can now speak to clients in plain GA4 channel language about AI traffic alongside organic and paid. Conversion rates by AI source become a question you can answer in a screenshot rather than a query. The longer-term shift is more interesting: once AI traffic is a first-class GA channel, the bar for treating it as a meaningful business signal in board-level reporting drops sharply. Expect AI traffic to start appearing in attribution and media-mix modelling within the next two quarters as a direct consequence.
Ahrefs Tests 1,885 Pages: Schema Markup Doesn’t Move AI Citations
Here’s a finding that will land badly for half the SEO industry. Ahrefs tested 1,885 pages that added JSON-LD schema markup and measured AI citation rates 30 days before and after, against matched control pages from different domains. The result: schema markup did not measurably change AI citation rates on any major platform tested.
The numbers. Google AI Overviews showed a 4.6% decline (within noise). Google AI Mode showed a 2.2% gain (also within noise). ChatGPT showed a 2.4% gain (likewise). Across four separate test rounds and 1,885 pages, schema added to already-indexed content had no detectable lift on whether AI engines chose to cite those pages.
Caveats Ahrefs flagged explicitly:
- Every tested page already had 100+ AI citations before the schema change, so this is a test of “more schema on already-cited content,” not “schema on uncited content.”
- Schema additions often coincide with other site changes that could mask the effect.
- A separate searchVIU experiment found AI systems ignore JSON-LD entirely when fetching pages directly, suggesting the structured data layer simply isn’t feeding retrieval the way Google’s classic crawler used it.
The 30-day window may miss slower effects, and the study doesn’t isolate schema types (Article vs Product vs FAQ etc). But the headline finding is hard to wave away: at scale, adding schema didn’t move AI citations.
The implication for client conversations. If a client believes schema is the lever to pull for AI visibility, they’re optimising the wrong layer. Investment is better placed in entity clarity, author attribution and content depth, the signals AI engines are actually weighting. Schema isn’t worthless (it still does work for rich results and structured retrieval in classic search), but the AI-citation premium most teams have priced in isn’t backed by this data.
Google Search Ranking Volatility Spikes Across Mid-May
Heavy ranking volatility hit Google search this week. The third-party tracking suites (Mozcast, SEMrush sensor, Serpstat) all spiked into the red on 13 and 14 May. Google hasn’t confirmed an update; the working theory is either an unconfirmed core component refresh or product-layer changes around AI Mode and AI Overviews.
For SEO teams whose clients felt the move, the routine applies. Document the affected keywords before the dust settles. Pull comparison snapshots from two weeks earlier. Don’t make panic changes for at least seven days; ranking volatility this acute usually has some reversion built in by the second week.
Marie Haynes: “Crawled, Currently Not Indexed” Numbers Are Climbing
Marie Haynes flagged a meaningful uptick in pages classified “Crawled – currently not indexed” inside Google Search Console over the past two weeks. Multiple practitioners are reporting the same pattern, especially on AI-assisted or template-heavy site sections.
Haynes ties the trend to two intertwined Google behaviours. The first is tighter sampling of crawled content against the quality threshold (consistent with the Dan Taylor analysis covered in last week’s roundup). The second is a shift in how Google decides which crawled URLs make it into the index. Pages that pass crawling but fail the quality gate land in this limbo state.
The audit move on any client site showing this: pull the affected URL list out of GSC, sample 20-30 pages, and look for shared signatures (thin content, duplicate templates, low entity coverage). If a pattern shows up, the fix is editorial, not technical. If pages are unique and substantive, the diagnosis is harder and likely involves topic authority signals across the broader cluster.
Google Discover Publisher Pages Get Links and Featured Posts
Google added two new modules to Discover publisher pages on 12 May. The first is a links section pointing to the publisher’s home page and key category pages. The second is a “featured posts” pinned-content block the publisher can curate.
For publishers chasing Discover traffic, the practical setup is straightforward. Audit existing publisher page content, push the featured-posts module towards evergreen, high-converting content, and verify the category links match the navigation structure on the public site. The links module is essentially free traffic redirect; the featured posts block is a managed promotion surface for owned content.
Inside ChatGPT Search: Web.run, Fan-Out Queries, and a 20% Drop in Cited Domains
How does ChatGPT actually decide which sites to cite? Olivier de Segonzac at RESONEO published a detailed reverse-engineering analysis of ChatGPT’s internal search architecture this week. Three findings stand out for anyone working AI visibility.
First, the tool that fetches web content is web.run, and its behaviour changed materially with the GPT-5.3 Instant model rollout on 4 March. The same tool that previously took simple text commands now uses structured JSON with 12 operations including search_query, open, find, click and screenshot. ChatGPT’s search workflow is closer to an automated browsing session than a single query lookup.
Second, ChatGPT performs what RESONEO calls fan-out queries: typically 10 or more separate searches per response, often using site: operators to restrict results to specific high-authority domains. A new undocumented operation, browse_rewritten_queries, handles product-comparison flows by fetching specs for each item individually. This is structurally different from how Google composes an AI Overview.
Third, and this is the data that will get cited heavily for the next few months, the switch to GPT-5.3 Instant on 4 March cut the average number of unique domains cited per response from 19 to 15. That’s a 20% drop. Unique URLs fell from 24 to 19. The pattern held across 400 daily prompts measured over 14 weeks, so it’s not a sampling artefact.
ChatGPT is consolidating around fewer, higher-authority sources.
The implication for practitioners. If your client’s brand was in the bottom tier of ChatGPT citations before March, the citation gap to top-tier brands has widened, not narrowed. If you’ve been treating ChatGPT visibility as a long-tail SEO problem, the new behaviour rewards a much smaller set of trusted sources. RESONEO’s piece also identifies “parametric visibility” (training data presence) and “dynamic visibility” (real-time retrieval) as distinct layers operating in parallel, which is a useful frame for diagnosing why a brand might appear in some ChatGPT responses but not others.
Conde Nast CEO: “Plan as if Search Is Zero”
“Plan as if search is zero.” That’s the directive Conde Nast CEO Roger Lynch gave his teams during a TBPN interview this week. Lynch said internal forecasts at Conde Nast have underestimated actual search traffic decline in each of the past three years, prompting a more aggressive planning baseline.
A few specifics Lynch flagged. Digital subscriptions are up 29% in revenue year on year, which is the channel Conde Nast is investing in to replace the search shortfall. Third-party data he cited shows search referrals fell 60% for small publishers over two years. Media leaders Lynch talks to expect over 40% further search traffic decline within three years.
He also described what he calls the “barbell effect” in publisher economics. Large authoritative brands (Vogue, The New Yorker) and small niche publications with loyal audiences (Pitchfork) perform well in an AI-mediated discovery world, while mid-market publishers without deep category authority or direct audiences face the steepest pressure. The strategic implication is that publisher M&A in 2026-2027 is likely to consolidate the mid-market into either the strong-brand top tier or the niche-loyal bottom tier.
For agency-side reading, the implication is that any client whose acquisition model depends meaningfully on organic search referrals should already be modelling a 30-50% reduction over three years and planning the migration to direct, paid, social, and AI-channel discovery. Lynch isn’t saying search will literally go to zero. He’s saying plan as if it could, because the projections you’ve been using are systematically wrong in one direction.
AI Visibility Isn’t One Problem. It’s Three Layers
Duane Forrester (who also wrote the citation-ROI piece in last week’s roundup) followed up with a useful diagnostic framework. AI visibility, he argues, isn’t one problem. It’s three distinct layers that fail for different reasons and need different fixes.
Layer one is retrieval. Can the AI even pull your content? This is the technical layer: crawlability, parseability, clean rendering for non-JS-executing bots. If retrieval is broken, schema and editorial work won’t matter.
Layer two is the knowledge graph. Are you a clean, recognisable entity? Consistent naming across the web, schema where it helps with entity disambiguation, brand mentions, presence on high-trust platforms. The knowledge graph layer is what Wikipedia entries and Google entity panels reflect.
Layer three is the context graph, which is newer. How does an AI agent reason about your brand internally to its current task? This is the layer where enterprise customer systems start to matter, because agents are increasingly grounded in business-specific context rather than open-web retrieval alone. Gartner expects 40% of enterprise applications to feature task-specific AI agents by year-end, up from under 5% last year.
The diagnostic test Forrester suggests: when a brand isn’t showing up in AI responses, identify which layer is breaking before pouring effort into the other two. Most teams default to “write more content,” which only addresses one layer.
Adobe Q2 2026: AI-Referred Retail Conversions Up 393% Year on Year
393% year-on-year growth in AI-referred traffic to US retailers in Q1 2026, peaking at 1,151% in December 2025. That’s the headline number from Adobe’s Q2 2026 AI Traffic Report, released earlier this week. The growth rate is impressive but not the most interesting finding.
The most interesting finding is the conversion reversal. Twelve months ago, AI-referred visitors to retailers converted at roughly half the rate of other channels. By March 2026, those same visitors converted 42% better than non-AI traffic. That’s a flip from underperforming to outperforming, the kind of channel-maturation curve we last saw with paid social around 2018-2019.
Other engagement metrics Adobe shared:
- Engagement (Adobe’s composite metric) up 12% versus non-AI traffic.
- Time on site up 48%.
- Pages per visit up 13%.
- Revenue per visit up 37%.
Adobe also flagged a “Citation Readability” finding. Top-performing retailers’ homepages scored 62% higher on machine readability than bottom performers. The mechanical drivers of citation readiness (clean HTML, semantic markup, lean JS) are correlating directly with AI-referred revenue.
The methodology caveat Adobe acknowledged: data is self-reported from retailers on Adobe’s analytics platform, and Adobe sells AI optimisation tooling alongside this research. Both points are worth holding in mind, though the underlying conversion-rate flip is the kind of finding that’s hard to fake at this volume of accounts. For retail clients in particular, this is the strongest data so far that AI-referred traffic is becoming a high-quality acquisition channel rather than a tyre-kicker channel. Budget allocation conversations should reflect that.
Google’s ALDRIFT: AI Answers That Sound Right and Actually Are
Google Research published a paper this week introducing ALDRIFT (Algorithm Driven Iterated Fitting of Targets), a framework for evaluating whether AI-generated answers are actually correct, not just plausible. The paper itself was published on arXiv in May 2026.
The problem ALDRIFT addresses is the “plausibility trap.” AI systems generate answers that sound coherent but fall apart in execution. ALDRIFT pairs a generative model with an external scoring process that checks whether candidate answers actually achieve the underlying goal, iteratively refining toward better solutions while preserving enough viable options to avoid dead ends.
This is theoretical research, currently tested only on small models (GPT-2 in the paper), and probably won’t ship in production AI Overviews for many months. But the direction matters. Google is moving from “AI Overviews that look right” toward “AI Overviews that are right.” That bar will rise across every AI surface, and the citation logic that follows will reward sources that consistently support correct answers, not sources that consistently support plausible-sounding ones.
The Consensus Gap: Only 2.37% of AI Citations Overlap Across Engines
Only 2.37% of cited URLs appear in all three of ChatGPT, Perplexity and Google AI Overviews for the same prompt. That’s the finding from Kevin Indig’s “Consensus Gap” analysis, drawing on 3.7 million citations across a 20,000-prompt sample. 91% of citations appeared in only one engine. The pattern held across four samples from Q3 2025 through Q1 2026, so it’s not a moment-in-time artefact.
Commercial queries showed 2.4% overlap; informational queries showed 2.0%. Format mattered too. Guides achieved 2.3% cross-engine overlap; homepages just 1.1%.
The dashboard problem Indig flags is the one we’ve seen on multiple client accounts. Aggregate AI visibility scores compress three separate ranking systems into a single number. A brand can look dominant on a dashboard while being completely invisible to two of the three engines. The score is technically accurate but operationally misleading.
“The strategic question is no longer ‘how do we rank in AI?’ but ‘how do we build assets that survive different engine preferences?'”
Indig recommends measuring three distinct metrics in parallel. Presence (any-engine visibility), portability (all-engine visibility), and concentration (distribution across platforms). The concentration metric is the one most agency dashboards don’t yet report on.
For our own client work, this lines up with what we’ve been seeing on AI visibility dashboards. A brand can dominate Perplexity and be near-invisible in AI Overviews, or vice versa, with no obvious topical reason. The Consensus Gap data gives that pattern a structural explanation: the engines are pulling from genuinely different source pools, not just ranking the same pool differently. Strategies need to reflect that the AI visibility map is three maps, not one.
AI Overviews Are Surfacing Negative Reviews Users Didn’t Ask For
A new analysis showed AI Overviews are surfacing negative review content for branded queries even when users didn’t ask for reviews. The pattern: a user searches for [brand name + general info], the AI Overview includes a “what users say” or “common complaints” panel pulled from review sites and forums, and negative sentiment appears prominently without any user intent signal asking for it.
For brands with thin review velocity or unmanaged negative content on Trustpilot, Reddit or industry forums, this creates a reputational surface that didn’t exist a year ago. The audit step is straightforward: search 10-15 client branded queries in AI Mode and AI Overviews, screenshot the responses, and flag any that surface negative review content. The remediation work is harder, requiring coordinated review velocity, response management, and proactive content for brand defensiveness in AI-visible content sources.
iPullRank’s “Cloaking for LLMs”: A Sharp Provocation
What if you cloaked AI crawlers instead of users? That’s the provocation Mike King at iPullRank floated this week. The technique inverts traditional cloaking: rather than showing search engines content hidden from users, you show users content hidden from LLM crawlers, while keeping Googlebot’s view fully intact.
The mechanism uses Cloudflare Workers to block specific JavaScript files for known LLM user agents, serving Markdown stubs instead. Since most AI crawlers don’t execute JavaScript, they don’t see the protected content. Real users get the full experience.
Whether this is a workable strategy depends on what you’re protecting. King acknowledges the obvious counters. User-agent spoofing breaks the wall, you have to monitor that Google still sees your content via URL Inspection, and you can’t cloak content you actually want surfaced in AI Overviews. For high-value proprietary content (research methodology pages, premium thought leadership, paid-tier guides), the technique is defensible. For public marketing pages, it’s not.
We’re not recommending it for client sites yet. But the broader argument King makes, that publishers and brands need new tools to control AI access independent of search access, is a directionally correct read of where 2026-2027 publisher rights conversations are heading.
Google Quietly Changes Search Terms Reporting for AI Queries
Google quietly updated its documentation this week to confirm that search terms in Google Ads reporting for AI-powered surfaces “may not always reflect a user’s exact query.” For AI Mode, AI Overviews, Lens and autocomplete experiences, the reported terms may now show Google’s interpretation of intent rather than the literal query the user typed. Anthony Higman spotted the documentation change on LinkedIn; Google hasn’t formally announced it.
The substantive question is how much “interpretation” we’re talking about. A literal “wedding photographer London” search reported as “wedding photographer London” is one thing. A user query that AI Mode rewrites into “best London wedding photographers with packages under ?2,000” before retrieving sources, and that gets reported back as the rewritten version, is something else entirely. Google’s documentation doesn’t yet draw the line.
For regulated verticals, this is genuinely uncomfortable.
If you’re advertising in financial services, healthcare or legal, your compliance team has been auditing reported search terms for brand-safety and consent purposes. If those reports are now showing Google’s inferred intent rather than literal user input, your audit basis has changed. Negative keyword strategies also break. You can’t reliably negative-match on terms that may have been generated by Google rather than typed by the user.
The right response is to flag this change with internal compliance, audit whether your current negative keyword lists are doing what you think they’re doing, and push Google for clarification on the distinguishability of literal versus interpreted terms in reporting.
Microsoft Advertising Extends LinkedIn Profile Targeting to CTV
Microsoft Advertising extended its LinkedIn professional profile targeting into connected TV inventory this week. Advertisers can now build CTV audiences using LinkedIn’s professional data layer (job title, seniority, company size, industry) and run video ads against them on connected TV.
For B2B brands, this is the first real CTV targeting option that doesn’t collapse to consumer demographics. Combined with brand-awareness measurement on the same platform, it allows B2B campaigns to chase upper-funnel CTV reach with audience signal that actually maps to B2B buying committees. The use case to test first on relevant client accounts is enterprise SaaS brand campaigns aimed at IT decision-makers, where consumer-CTV targeting has previously been too imprecise to justify the spend.
Google Ads Teases Gemini-Powered Dashboards
A teaser, not a launch. Google previewed a set of Gemini-powered dashboards inside Google Ads this week, showing natural-language query input, AI-generated narrative summaries of performance trends, and automatic anomaly detection across campaign metrics. No launch date yet; Google framed it as “coming soon.”
This is the direction Google’s been heading since the first AI summaries appeared in Search Console. If shipped well, it removes a meaningful chunk of the manual reporting work agencies do for clients. If shipped badly, it generates plausible-sounding narratives that paper over the actual metric movement. Watch for the launch window, and benchmark Gemini-generated narratives against your standing report templates before relying on them in client decks.
Yoast Launches an AI Content Planner Inside Premium
Yoast shipped an AI Content Planner inside Yoast SEO Premium on 12 May. The tool analyses existing site content, suggests new post ideas tuned to the site’s topical authority, and generates structured drafts with outlines, focus keyphrases, and meta descriptions ready to refine.
It’s a sensible move from Yoast. Most WordPress users with SEO plugins were already pasting site URLs into ChatGPT to get content ideas; this brings that workflow inside the CMS with platform-aware context. The drafts will need editorial review (no surprise there), and the focus-keyphrase suggestions are only as good as the underlying topical model. But for SMB clients running Yoast Premium, this is a useful new starting point for content briefs, not a replacement for content strategy.
Strategic Direction: Where Search Marketing Is Heading
Three themes from this week’s announcements.
One. AI traffic is going mainstream as a measurement channel. The GA “AI Assistant” channel and Adobe’s Q2 conversion-flip data both say the same thing from different angles. AI-referred traffic is no longer the experimental long-tail. It’s a quantifiable, conversion-positive acquisition channel that belongs in standing reporting and budget conversations.
Two. The SEO orthodoxy on AI visibility is being tested in public. The Ahrefs schema study, the Consensus Gap research, the RESONEO ChatGPT architecture analysis. All three challenge the comfortable narrative that “good SEO is good AI SEO.” Schema doesn’t move citations, engines don’t agree on which sources to cite, and ChatGPT specifically has consolidated around fewer high-authority sources since March. These aren’t reasons to panic, but they are reasons to rebuild measurement and strategy from first principles rather than porting 2024 SEO instincts forward.
Three. The publisher and PPC reporting layers are both losing transparency. Conde Nast’s “plan for zero” framing and Google’s quiet change to AI search terms reporting are different signals from the same direction. Publishers can’t count on the open-web economics of the last decade. Advertisers can’t count on literal query data in reporting. Both groups need new measurement and risk-management frameworks, and the burden of that work will land on agencies and in-house teams over the next 12-24 months.
Plan accordingly.
Key Takeaways
- Validate that the new GA “AI Assistant” channel is populating cleanly on every client account, and retire any custom Referral filters that were pulling AI traffic manually. If you rely on Perplexity attribution, keep a custom filter in place until Google confirms inclusion.
- Treat the Ahrefs schema study as a reason to reallocate effort from schema markup to entity clarity, author attribution, and content depth for AI-citation gains. Schema still earns its keep for rich results and classic search.
- Plan AI visibility measurement around presence, portability and concentration rather than a single composite score. The Consensus Gap data means three separate metrics or you’re flying blind on two of three engines.
- For publishers and content-heavy clients, model search traffic at zero by end of 2027 as a stress test. Even if reality lands at -30% to -50%, the planning discipline of running the zero case forces direct-channel investment.
- Audit AI Overview output for client branded queries quarterly, particularly checking for unsolicited negative-review surfacing. Build a remediation plan now for any client with unmanaged negative content on review sites or forums.
- For regulated PPC verticals, flag the AI search terms reporting change with internal compliance and verify negative keyword effectiveness before the next reporting cycle.
Frequently Asked Questions
What is the new Google Analytics “AI Assistant” channel and what does it track?
GA4 now categorises traffic from ChatGPT, Gemini and Claude into a default “AI Assistant” channel group. The medium is ai-assistant and the campaign is (ai-assistant). It’s a default on every property, no setup required. Perplexity isn’t included in the May 2026 announcement, so if you depend on Perplexity attribution, keep a custom filter running.
Did Ahrefs prove schema markup is useless for SEO?
No. Ahrefs proved that adding schema to already-cited pages doesn’t measurably improve AI citation rates on the engines tested (Google AI Overviews, AI Mode, ChatGPT). Schema still works for rich results, structured retrieval in classic Google search, and entity disambiguation. The finding is narrow: adding schema isn’t a lever for AI visibility specifically.
How seriously should I take Conde Nast’s “plan for zero search” memo?
Seriously, but interpret it correctly. Lynch isn’t predicting literal zero. He’s saying internal forecasts have underestimated decline in each of the past three years, so the planning baseline should be more aggressive. For clients with material organic dependency, modelling a 30-50% reduction over three years and planning direct-channel investment is a reasonable interpretation.
What changed in Google’s search terms reporting for AI queries?
Google quietly updated documentation to confirm that reported search terms for AI Mode, AI Overviews, Lens and autocomplete may show Google’s interpretation of user intent rather than the literal query typed. The practical impact is that negative keyword strategies and compliance audits relying on literal query data are now less reliable for those surfaces. Flag this with internal compliance for regulated verticals.
Should I start using “cloaking for LLMs” on client sites?
Not yet. The technique is defensible for genuinely proprietary content where you want to block LLM access while preserving Google indexability. But for general marketing pages, it risks blocking yourself from AI visibility you actually want, and it’s brittle against user-agent spoofing. Treat it as an option for premium content protection, not a default policy.
Conclusion
This week’s news is about measurement and strategy catching up to where AI search actually is. Google Analytics finally treats AI traffic as a first-class channel. Ahrefs publishes the kind of negative result the industry needed to read. Kevin Indig’s research forces a rethink on how agencies score AI visibility. The Conde Nast memo names what publishers have been quietly modelling for two years.
For agencies, the practical move this quarter is to update reporting templates with the new GA channel, rebuild AI visibility scorecards around presence, portability and concentration, audit AI Overview attribution for any client with brand-defensibility concerns, and flag the AI search terms reporting change with PPC compliance teams. None of this is heavy lifting individually. Together it adds up to the next iteration of how the work gets reported and managed.
Need help adapting your search strategy for the AI era? Contact the Anicca team for expert SEO and PPC guidance.
