Category: Uncategorized

What just changed in GA4

Google Analytics 4 has opened a native product link to Google Business Profile. Phone calls, taps for directions, bookings, and messages generated from a GBP listing now flow into the same GA4 property that already records website sessions and form submissions. The link lives under GA4 Admin, in the Product Links section, and it is rolling out to accounts gradually at no additional cost.

For site owners who run technical SEO audits, the headline is that high-intent local actions that used to vanish into a separate GBP dashboard are now first-party data inside Analytics. Auditors who previously had to flag “attribution gap” as a known unknown on local SEO reports now have a measurement surface they can test, verify, and recommend.

Which metrics land in GA4 and where they appear

After linking, a dedicated GBP section becomes visible in the standard reports. Seven interaction types populate the view: total interactions, website clicks, calls, directions, messages, bookings, and menus. The Google Analytics Help documentation confirms that linking a Google Business Profile to an Analytics property surfaces Business Profile performance data inside Analytics, with no UTM parameters required because the connection is server-side.

Two practical audit notes come out of that:

• The metrics show up in acquisition and engagement reports, so auditors reviewing those report templates should expect a new data source and re-check any filter or segment logic that assumed only website traffic populated those views.
• Data is retained for six months. Even if a property’s data retention setting is longer, GBP events older than six months will not appear, which affects any year-over-year comparison an audit might run.

What the numbers say about why this audit gap mattered

According to BrightLocal’s 2025 Local Marketing Survey, the average Google Business Profile receives 1,009 interactions per month, and 42 percent of those actions are calls or direction requests. The same survey reports that 73 percent of consumers have used a business’s profile to call or get directions. Until the GA4 link shipped, none of those interactions were visible inside GA4 reports, so any audit that relied on Analytics to model local conversion value was working from an incomplete dataset.

That is the core reason auditors should treat this as a measurement fix, not just a dashboard convenience. A plumber, salon, or bakery that drives half of its profile actions through calls was, by definition, underreported in GA4. The audit implication is direct: every prior conversion baseline an auditor signed off on needs a recalibration note.

How to audit the link itself

The setup path is short enough to verify during an audit run:

• GA4 Admin, then Product Links, then Google Business Profile links.
• Click Link and pick the profile or profiles to connect.
• Confirm, with editor or administrator rights on the GA4 property and owner or manager rights on the GBP.

Auditors should check three things on a real account:

1. The link actually exists and is active, by opening the GBP section under Reports.
2. The seven metric types are populating with non-zero values once enough time has passed.
3. Access controls match the team’s permission model. GBP owner access is broad, and a link created with the wrong account can quietly expose a profile to an unrelated GA4 property.

The aggregation problem for multi-location sites

If a brand links more than one Business Profile to a single GA4 property, every metric collapses into one aggregated dataset. There is no built-in per-location filter or segment in GA4 at this stage. For a multi-location SEO audit, that is a real limitation, because location-level decisions, such as which storefront to invest in, still require the GBP dashboard or the Performance API.

Auditors working with multi-location clients should flag this clearly in the report. The integration improves top-level visibility but does not replace the granular location exports that franchise and chain operators depend on.

What to add to the audit checklist this week

For any client whose organic strategy leans on local search, the audit should now include:

• A confirmation that the GBP-to-GA4 link has been created, with a screenshot for the deliverable.
• A check on GBP completeness. A profile that is missing hours, categories, or service items will underperform on calls and directions before any tracking question even matters, because profile quality drives whether a user taps at all.
• A name, address, and phone number consistency check across the major directories, since GBP remains a primary surface for local discovery and AI-driven answers.
• A note in the conversion baseline that pre-link GA4 data did not include calls or directions, so any historical comparison needs a footnote.
• A separate segment or exploration that isolates GBP-driven calls, because they will now distort totals in acquisition reports if treated as standard traffic.

What is likely to change next

Google has framed this as a phased rollout, which means the feature set will expand. Multi-location operators are an obvious constituency for per-profile segmentation, and a deeper tie-in with Google Ads is the natural next step. Once calls and direction taps can feed bidding algorithms as conversion signals, local search campaigns gain a measurement loop that did not exist before. Subproperties and audience building are also plausible extensions, since both already exist elsewhere in GA4.

AI-driven discovery adds a second reason to watch this area. When a voice assistant or AI agent surfaces a business and the user calls, that call now lands in GA4 as a trackable event, giving auditors a way to measure leads that originate outside the traditional web journey.

FAQ

What metrics does the GA4 and GBP integration track?

The integration tracks seven interaction types from a linked Google Business Profile: total interactions, website clicks, calls, direction requests, messages, bookings, and menus. They appear in GA4 under acquisition and engagement reports once the profile link is confirmed.

Can an audit see per-location data when several Business Profiles are linked?

No. Linking more than one GBP to the same GA4 property aggregates every metric into a single dataset with no per-location segmentation. Multi-location brands should keep using the Business Profile dashboard or the Performance API for location-level detail.

How long does GA4 keep Google Business Profile data?

GBP metrics in GA4 are retained for six months. Reports will not show events older than six months even if the property’s general data retention setting is longer, so periodic exports are needed for long-term trend analysis.

In early 2026, SpaceX, Google, and Anthropic jointly confirmed plans for AI1, a solar-powered AI data center designed to operate in low Earth orbit at roughly 550 kilometers. The cluster, which pairs Google TPU v6e accelerators with Anthropic inference engines, is scheduled to ride a SpaceX Starship to orbit in Q3 2027. The project reframes the geography of compute: rather than drawing power from terrestrial grids or pulling in air for cooling, it runs on sunlight and sheds heat into the vacuum of space.

Why an orbital data center changes the audit checklist

Most technical SEO audits treat latency as a function of distance to a fixed data center. AI1 breaks that assumption. The cluster orbits overhead, so a request from Mumbai or Berlin may terminate on hardware passing directly overhead rather than routing to a server in Virginia or Frankfurt. The ground-to-orbit signal leg from 550 km is around 3 milliseconds, compared with the 40 to 100 milliseconds a continental round trip typically takes. When inference moves to a node in the sky, the path between user and server compresses in ways that traditional traceroutes will not show.

For a site owner, this shifts what counts as edge computing. AI-powered features like chat assistants, voice agents, real-time translation, and local recommendation widgets will pull from orbital nodes that pass within line of sight of a ground station. The relevant question becomes whether your structured data is precise enough for an orbital inference layer to retrieve and serve during a 10-minute ground pass.

How AI1 produces and dissipates energy

The cluster is built as a modular set of compute nodes mounted on a SpaceX satellite bus. Each node carries Google TPU v6e silicon and Anthropic fine-tuned inference engines, fed by a pair of unfolding solar arrays that span more than 40 meters tip to tip. Peak output is roughly 100 kilowatts. Above the atmosphere, sunlight delivers about 1.36 kW/m², roughly 40 percent more peak irradiance than the strongest ground-based solar farms achieve, and that energy is available nearly continuously.

Cooling is the harder engineering problem. Vacuum blocks convection, so heat can only leave through radiation. AI1 uses a two-phase pumped loop to pull heat away from the chips and dump it into large deployable radiators coated in a high-emissivity white paint. The radiators are sized for a continuous 30-kilowatt thermal load, which keeps chip junction temperatures below 85°C during full utilization. SpaceX ran early radiator prototypes on Transporter rideshare missions and confirmed stable temperatures across the 90-minute eclipse cycle.

A sun-synchronous orbit means AI1 crosses the same ground points at roughly the same local solar time each day, which simplifies scheduling for the optical laser downlinks that feed it. Twelve ground stations, each capable of 100 Gbps, handle result delivery and ingest new model shards during each pass. The design also reflects Google Cloud’s carbon-intelligent computing direction. Thomas Kurian, CEO of Google Cloud, framed the project as a step toward proving that orbital infrastructure can cut the carbon cost of serving billions of daily AI queries.

What the numbers mean in practice

• 100 kW of solar generation can sustain about 600 TPU v6e chips while the satellite is in sunlight (Google Cloud, 2026).
• 3 ms signal leg from 550 km altitude to a ground station, versus 40-100 ms for transcontinental terrestrial round trips.
• 30 kW continuous heat rejection through deployable radiators, with chip junctions held under 85°C.
• 12 ground stations handling 100 Gbps laser up- and downlinks each.
• Zero water consumption for cooling, compared with the millions of gallons per day used by large terrestrial AI campuses.

What to verify on your own site today

Audit work changes in three concrete ways once orbital inference enters the picture.

First, structured data. AI serving layers, whether terrestrial or in orbit, lean on schema markup to resolve entities quickly. Confirm that your local business, product, and FAQ schemas are complete and consistent across pages. Missing fields force the inference layer to fall back to slower retrieval paths, which negates the latency advantage an orbital node provides.

Second, location signals. AI1 will favor results that carry clean geographic tags during a short ground pass. NAP consistency, geo coordinates, and hreflang coverage all matter more when an orbital node has only minutes to resolve a query before moving on. Run a local SEO audit the same way you would for a new search feature rollout.

Third, real-time AI integrations. Chatbots, voice assistants, and recommendation widgets that rely on generative inference should be tested with fresh prompts from multiple continents. If response times stay flat regardless of origin, the provider may already be routing through edge or orbital tiers. If they spike in regions you serve, document the gap and flag it for the vendor.

Is orbital compute economically realistic?

Launch cost remains the swing factor. Starship pricing sits between $1,500 and $2,000 per kilogram to low Earth orbit. A 100 kW payload complete with radiators, eclipse batteries, and radiation shielding would weigh somewhere between 8 and 12 metric tons, putting total launch cost in the same range as building a small ground data center. Early internal estimates from Google suggest a payback window of three to five years for inference-only workloads, assuming the satellite hits 99.9 percent uptime.

Radiation is the second risk. Cosmic rays and solar particle events can flip bits in memory, so AI accelerators need either hardened silicon or triple-redundant error correction. Starlink has shown that SpaceX hardware can survive thousands of orbits with minimal failures, but AI accelerators are denser and more complex than routing chips. The first twelve months of AI1 operations will double as a silicon stress test.

Timeline and next steps

AI1 is scheduled to launch in Q3 2027 aboard Starship. The first six months in orbit will be a research phase focused on thermal stability, radiation hardening, and how liquid-cooled TPUs behave in microgravity. Google has already committed to at least three follow-on launches if AI1 hits its KPIs, with a longer-range plan for a constellation of 40 to 60 orbital nodes functioning as a distributed supercomputer. Anthropic is exploring whether its Constitutional AI training framework can run entirely on station, which would mark the first major model update performed without drawing on a terrestrial grid. SpaceX is also looking at whether the Starlink laser mesh can tie multiple AI1 nodes into a space-based data center mesh and cut downlink hops.

FAQ

What is SpaceX AI1?

AI1 is a solar-powered AI data center that SpaceX, Google, and Anthropic are building together. It runs Google TPU v6e accelerators and Anthropic inference engines on a SpaceX-built satellite bus in low Earth orbit at roughly 550 km.

How does AI1 cool itself without air?

AI1 uses a two-phase pumped liquid loop to carry heat from the chips to deployable radiators coated with high-emissivity white paint. The radiators shed heat as infrared radiation, handling a continuous 30 kW thermal load and keeping chip junctions below 85°C.

When does AI1 launch and what comes after?

AI1 is slated for Q3 2027 aboard Starship, followed by a six-month research phase. Google has committed to at least three follow-on launches if KPIs are met, with a longer-term vision of a 40 to 60 node orbital constellation.

Mythos-1, the first publicly released Mythos-class AI model, launched from research collective Mythos AI with 1.2 trillion parameters, native support for five input modalities, and tool calls generated directly in the response stream. Access is offered through a public API plus downloadable weights under a non-commercial license, with commercial use requiring a separate agreement. For teams that audit how their content surfaces inside AI systems, the release resets the baseline for what “capable” means.

Why a Mythos-class model matters for technical SEO audits

Until now, the largest language models in public circulation have been text-first systems. Vision, audio, and structured inputs were handled by bolt-on encoders that translated everything into tokens before reasoning began. Mythos-class architecture inverts that pattern: vision, audio, text, code, and structured tool instructions share a single representation space from the first training step. For audit work, the practical consequence is that a model can consume a screenshot, a page’s HTML, a CSV of log data, and a voice note in one inference pass without losing fidelity between modalities.

Site owners who have been optimizing for text-only retrieval and text-only summarization are about to face systems that read pages the way a human reviewer would: looking at the rendered layout, listening to a clip, reading structured data, and acting on it. The crawl budget and schema audits that have defined technical SEO for years still apply, but the bar for “AI-readable” now includes fidelity across modalities.

Three design choices that separate Mythos-1 from prior releases

Three architectural decisions in Mythos-1 change the audit checklist, and each one maps to a specific thing to test on your own properties.

Interleaved latent fusion across modalities

Mythos-1 replaces the classic transformer stack with a novel design that encodes all input modalities into a shared latent space before reasoning. There is no intermediate text translation step. For auditors, this means you should verify that pages with rich visual content (infographics, product photos, video thumbnails) carry alt text, captions, and surrounding context that still make sense when the visual signal is degraded or absent. Models that fuse modalities natively still fall back to text when an image is unreadable.

Action tokens for native tool use

The model includes a dedicated vocabulary of action tokens that produce tool calls, API requests, database queries, and browser actions directly in the response stream. No external orchestrator or chain-of-thought post-processing is required. The audit implication: any workflow your site exposes to an agent (forms, booking endpoints, product filters, schema.org actions) needs to be reachable through a clean, documented interface. Mythos-1’s AgentBench success rate above 85 percent for multi-step tasks means agents will attempt these workflows; surfaces that are undocumented or partially blocked will produce failures that look like content gaps.

Trillion-parameter scale with sub-300ms serving

At 1.2 trillion parameters, Mythos-1 is the largest model distributed both as an API and as downloadable weights. The custom serving stack keeps typical prompt latency under 300 ms. For SEO purposes, latency matters because response time influences whether an agentic system will retry, abandon, or fall back to a cached answer. Pages and APIs that respond slowly during agent traffic are likely to be skipped on subsequent passes.

Benchmark numbers worth weighing in your next audit

Standard evaluation suites show Mythos-1 performing at or above leading peers on reasoning (MMLU and MMLU-Pro, following Hendrycks et al., 2021), code generation (HumanEval and MBPP), and multimodal perception (SEED-Bench 2 and MMBench). Two figures stand out for audit planning:

• AgentBench multi-step task completion above 85 percent, the highest publicly reported for any model of this scale.
• A 1 million token context window, enough to ingest an entire codebase, a full-length film, or days of audio in a single prompt.

These numbers don’t change what a single page should contain. They change how thoroughly a model can evaluate a property. Auditors should expect AI-driven review tools to ingest full site archives, cross-reference internal links, and run multi-page diagnostic flows without losing context mid-task.

What Mythos AI has signaled for the roadmap

Mythos AI has framed Mythos-1 as the first entry in a planned family. A smaller “Mythos-1 Mini” aimed at on-device deployment is expected within months, and the team is working with cloud providers to make the full 1.2T model runnable on commodity GPU clusters later in the year. The non-commercial weight release has already produced a wave of community fine-tuning experiments, which suggests vertically specialized Mythos-class models (legal, medical, financial, ecommerce) will follow quickly. A version trained on robotics sensor streams, adding touch and spatial understanding to the native modality set, is on the research roadmap.

For site owners, the roadmap signals that native multimodal ingestion is the new floor, not the ceiling. Audits planned for the next two quarters should assume agents will see, hear, read, and touch (through structured interfaces) anything your site makes available.

Practical audit items to add before Mythos-class agents hit your stack

Four checks move up the priority list based on what Mythos-1 demonstrates is now possible.

1. Verify that every visual asset on priority pages has descriptive alt text, caption, and adjacent body copy that remains coherent when the image is removed. Native multimodal models degrade gracefully, but they still rely on text fallback.
2. Document every agent-reachable endpoint: forms, booking flows, product filters, JSON-LD actions, sitemaps. Confirm each one returns a clean response under load, because action tokens will exercise these paths directly.
3. Stress-test page and API latency at the edge, not just from a single region. Sub-second response times influence whether an agent retries or moves on.
4. Audit structured data for completeness across the full page object, not just the headline entity. A 1 million token context window means an agent can compare every property on a page against your schema and flag silent mismatches.

The bigger picture for sites preparing for agentic traffic

Mythos-1 marks the point where the largest publicly available model treats the full range of human perception and action as its native language. For technical SEO audits, the takeaway is direct: the systems evaluating your pages can now see, hear, read, and act in a single pass, and they will reach for your endpoints the same way a human reviewer would. Sites that document their interfaces, describe their visuals, and respond quickly will be the ones that surface correctly in agent-driven discovery.

FAQ

What makes a model “Mythos-class”?

A Mythos-class model is built from the first training step for native multimodal processing across vision, audio, text, code, and structured tool instructions. It includes built-in agentic reasoning that emits tool calls directly in the response stream, and it operates at parameter scales exceeding one trillion.

Who built Mythos-1 and how can developers access it?

Mythos-1 was built by Mythos AI, a research collective that kept a low profile before launch. It is available through a public API and as downloadable weights for non-commercial research use. Commercial deployment requires a paid agreement with Mythos AI.

What audit priorities shift with Mythos-1’s release?

Auditors should prioritize descriptive alt text and captions for visual assets, documentation and load testing for agent-reachable endpoints, edge latency measurement, and full-page structured data completeness. These checks align with Mythos-1’s native multimodal fusion, action-token tool use, sub-300 ms serving, and 1 million token context window.

When Google surfaces an answer directly, how often does a user actually leave the page to visit a third-party site? A new analysis from seoClarity puts a number on it. Across 1.5 million anonymized search sessions logged in April 2026, AI Mode held users in conversation for almost three times longer than AI Overviews while cutting external clicks to less than a third. For anyone tracking referral traffic, the gap between these two AI surfaces is now the gap to plan around.

What the study measured

Researchers at the SEO platform seoClarity compared two session types side by side: searches that triggered an AI Overview sitting atop a traditional results page, and searches conducted entirely inside AI Mode, the fullscreen chat interface Google introduced as an opt-in experiment in early 2026. The dataset captured session duration, follow-up query count, external click-through behavior, and query category for each session.

Both interfaces draw on similar underlying language models, so the contrast in user behavior comes down to interaction design rather than answer quality. That distinction is the one worth understanding before changing a content strategy.

The click-out gap

External click-through behavior is where the two surfaces diverge most sharply. Sessions that included an AI Overview recorded a 12% click-out rate to third-party sites. Sessions inside AI Mode produced a 4% click-out rate, a threefold reduction. Median source clicks per session followed the same pattern: 1.4 for AI Overviews, 0.3 for AI Mode.

Even with citation links present, the conversational interface keeps the user asking the next question rather than opening a new tab. That single behavioral pattern reframes what a top placement inside an AI summary is actually worth.

How long users stay and how deep they go

AI Mode sessions averaged 3 minutes 12 seconds. AI Overview sessions averaged 1 minute 8 seconds. The extra time corresponds to follow-up activity: AI Mode users typed an average of 2.8 additional queries per session, building a multi-turn thread instead of a single search-and-leave action.

Query intent also split along predictable lines. Roughly 62% of AI Mode queries were informational or exploratory. Transactional and navigational searches still gravitated toward the classic results page beneath AI Overviews, where product listings, local packs, and direct site links remain visible.

What an audit of your own pages should now check

Reading the study as a site operator turns the findings into a checklist. A few items deserve a fresh look the next time you run a technical audit.

Track AI Overview impressions and clicks separately in Search Console

Google has begun reporting AI Overview impression and click data as its own line in Search Console. Pull that report, filter to the last 90 days, and compare click-through rate against your non-AI queries. A page that historically pulled a 4% organic CTR may show something different when an AI summary sits above it, and that delta is the real signal worth watching.

Audit your citation readiness

AI Overviews still link out, so source credibility still matters. Crawl your top 20 pages and check whether each one includes clear author bylines, primary source links, structured headings, and a publication or update date. Pages missing those elements are less likely to be chosen as a citation source, which directly affects the 12% click-out pool.

Confirm your brand entity is machine-readable

AI Mode rarely clicks out, but it still names brands when answering. Search for your brand name in a knowledge graph viewer, then check that your company name, address, leadership, and product lines match across your site, Wikipedia, Wikidata, Crunchbase, and major directories. Inconsistent entity data is a quiet reason an LLM mentions a competitor instead of you inside an AI Mode reply.

Map content to query intent, not just keywords

The study found informational and exploratory queries dominant in AI Mode and transactional queries dominant in the standard SERP. Run your existing pages against an intent classification and split them: informational guides should be optimized for inclusion in AI answers, while product, pricing, and location pages should be optimized for the classic results that still drive clicks.

Watch the ad roadmap

Google has started testing ad placements inside AI Mode. If you currently buy branded search ads, revisit whether the same spend will reach users inside a chat surface. A paid placement inside AI Mode may soon be the only paid way to interrupt a conversational thread that never clicks out.

Why the divergence is structural, not temporary

It is tempting to file the click-out gap under “early product behavior” and assume users will start clicking more once they learn the interface. The study’s session length and follow-up data push against that read. Users are not bouncing off AI Mode; they are staying and refining. Each follow-up replaces what would have been a new search on Google, let alone a visit to another site. That substitution effect, not unfamiliarity, is what suppresses outbound clicks.

Internal roadmap reporting suggests AI Mode could become the default mobile entry point for broad informational queries later in 2026, with AI Overviews continuing to anchor the desktop layout. If that rollout happens, the 4% click-out figure is the one to plan around, and the audit checklist above is the practical bridge.

FAQ

What is the difference between AI Mode and AI Overviews?

AI Overviews are an AI-generated summary that appears at the top of a standard Google results page, with linked citations and the traditional blue links still visible below. AI Mode is a separate chat-style interface that opens a fullscreen conversation and answers follow-up questions without showing a classic results page.

Do users click on websites more from AI Overviews or AI Mode?

Users click out far more often from AI Overviews. The seoClarity study found a 12% external click-out rate for AI Overview sessions, compared with 4% for AI Mode sessions. Median source clicks per session were 1.4 for AI Overviews and 0.3 for AI Mode.

How long do users stay in AI Mode compared with AI Overviews?

AI Mode sessions averaged 3 minutes 12 seconds with 2.8 follow-up queries per session. Sessions involving an AI Overview averaged 1 minute 8 seconds. The longer time in AI Mode reflects multi-turn refinement rather than a single search action.

A user behavior study on Google AI Mode found that 88% of searchers take the synthesized answer without ever scrolling back to compare the underlying sources. In the older AI Overviews interface, that figure sat around 50%. The shift cuts the funnel from ten links a buyer might browse to one to three citations the model decided to trust.

Why this changes a technical SEO audit

For two decades, ranking meant competing for a click on a results page. You could sit in position five, write a strong title, and still earn traffic. That math no longer holds inside AI Mode. If your page is not in the citation set, it is not in the consideration set, and an audit needs to start looking for citation readiness, not just blue-link readiness.

Google has confirmed that AI features in Search now reach more than a billion people. Once 88% of those users accept the answer the model writes, being a citable source becomes the job. Audit checklists built around rank tracking and click-through curves are looking at the wrong dial.

What the query fan-out does to your content

AI Mode works by fanning a single question into several sub-queries, pulling passages from a handful of trusted pages, and composing one response. The retrieval step weighs sources it can verify and attribute. A named author, a stable employer, a consistent business listing, and a clear publication trail all help the model treat a page as quotable. A thin anonymous article offers none of those signals.

This is the mechanism behind a second finding from the same study: LinkedIn has climbed to the number two most-cited source in AI answers, behind only YouTube. Posts on those platforms carry a verifiable identity by default. Random blogs do not, unless the owner goes out of the way to attach one.

What to check in your own audit right now

Treat the three areas below as a starting point for any AI-readiness review. Each maps directly to a signal the model’s retrieval step is known to weight.

Entity and contactability

• Claimed and verified business profile on Google, with name, address, phone, and category fields completed.
• Same NAP data echoed consistently across directory listings and social profiles.
• Organization markup on the home page with a sameAs array pointing to every official profile.
• A working contact path the model can resolve, not a contact form with no address or phone behind it.

Authorship and expertise signals

• Author bylines on every substantive page, linked to a populated author entity.
• Bios that name a real person with a role, employer, and a credential the model can cross-reference.
• First-hand detail in the prose: original data, named case examples, dated observations. Commodity restatements of common knowledge carry no signal.
• Existing strong pages refreshed and expanded before new thin pages are published.

Off-site footprint and entity consistency

• LinkedIn company page and key executive profiles active and aligned with the brand description on your site.
• YouTube presence where the topic supports video, with channel authorship and consistent branding.
• Review profiles (Google, industry-specific directories) maintained, since review signals travel with citation trust.
• Social publishing cadence steady enough that the entity reads as alive, not abandoned.

The operational data you need to preserve before it disappears

Two deadlines on the Google side also affect what gets measured, not just what gets cited.

Starting in June 2025, Google Ads begins deleting hourly, daily, and weekly granular data older than 37 months. Any year-over-three seasonality model that relies on that grain of reporting breaks once the rows are gone. Pull the historical exports now if the audit touches paid performance benchmarking.

By January 2027, standard Display campaigns migrate into Demand Gen, Google’s goal-based format spanning YouTube, Discover, and Gmail. Manual placement control narrows. Goal signals and creative quality carry more weight, which raises the cost of sloppy creative audits.

Paid placement is moving into the same surface

On the open side of the market, ChatGPT is introducing pay-per-conversion ads inside its conversational surface. Pricing on outcomes rather than clicks is the first real ad inventory inside an AI assistant. For operators running audits that span both organic and paid, the implication is that the distinction between a citation and a placement is going to soften. Both will live inside the same answer box, and being a trusted, verifiable entity remains the price of admission to either.

What separates a citation-ready page from a non-citable one

Pages that the model cites tend to share four traits, and an audit should score each page against them.

1. The page answers a specific question the model is likely to fan out on, not a vague topical hub.
2. The answer sits above the fold in clear paragraphs the model can lift as a passage.
3. The page carries a verifiable author and publisher with linked entities the model can resolve.
4. The same facts stated on the page line up with the same facts stated on the brand’s profiles elsewhere. Contradictions across surfaces are a quiet demerit.

The shape of the audit report to write next

If you are documenting AI readiness for a client or for your own site, the report needs a different structure than a traditional SEO audit. Lead with entity completeness. Show the gap between the brand’s claimed profiles and the actual on-page markup. Then list the top pages by traffic or revenue and score each on the four citation traits above. End with a paid-side section that flags any Display campaigns still on the legacy path before the January 2027 migration.

The wider pattern is straightforward. The web is shifting from a place people browse to a place machines summarize, and the 88% acceptance figure is the clearest measure yet of how far that move has gone. An audit built to track that shift checks whether the machine can identify, verify, and quote the business. Everything else is secondary.

FAQ

What did the AI search behavior study find about AI Mode?

The study found that 88% of users accept Google AI Mode results as-is, without scrolling back to compare the cited sources. In the older AI Overviews format, roughly 50% of users scrolled backward to compare. The funnel has tightened from ten possible links to one to three sources the model chose to trust.

Why is LinkedIn the number two most-cited source in AI answers?

LinkedIn sits at number two behind YouTube because posts on both platforms carry verifiable authorship by default: a named professional with a job title, or a named channel. AI models favor content where the author can be identified and the expertise cross-referenced, which is exactly what a random blog post usually fails to provide.

How can a business get cited in AI Mode results?

Claim and standardize the business listing so name, address, phone, and category data match across every surface. Put a named author with a real bio behind every substantive page, and write content that shows first-hand detail. Keep LinkedIn, YouTube, and review profiles active and consistent with what the site says. Citation goes to the entities that are easiest to identify, verify, and quote.

AI-driven search interfaces such as Google AI Overviews, Gemini, Perplexity, and ChatGPT Search have shifted the goalposts for SEO teams. Rather than rewarding raw publishing volume, these systems tend to cite a small set of pages that are accurate, current, and trustworthy at the moment a query is asked. For site owners running technical audits, that changes what is worth measuring and what is worth fixing on existing pages.

Why volume stopped being the strategy

Content velocity once dominated SEO playbooks. More indexed URLs meant more keyword targets and more chances to rank, which produced sprawling calendars of near-duplicate city pages and minor keyword variations. Answer engines break that loop by returning one synthesized answer and only a handful of citations per query.

Google Search Central documentation on AI features reinforces the same underlying criteria that drive classic rankings: helpfulness, reliability, and people-first content. When citation slots are scarce, extra low-value pages can dilute topical authority rather than add to it. A page that wins in AI search is usually the one that best represents the answer when someone asks, not the one with the newest publication date.

How do AI systems treat freshness?

Freshness still matters, but in a different sense. AI retrieval pulls and synthesizes information on the fly, and stale data adds uncertainty. An article that still recommends outdated tactics such as keyword stuffing or exact-match anchor spam is less likely to be cited, regardless of writing quality.

This puts a hard line between cosmetic updates and meaningful ones. Changing the publication date without improving the page rarely produces movement. A genuine refresh swaps in current statistics, adds recent examples, updates screenshots, includes new expert input, expands the FAQ, tightens internal links, and broadens semantic coverage around the topic. Visible “Last Updated” timestamps, refreshed metadata, and dateModified schema markup help crawlers notice the change, but they amplify substance rather than replace it.

This also maps to the Query Deserves Freshness model. Queries most likely to demand recent answers fall into a few buckets: time-sensitive topics such as pricing, software releases, AI developments, and statistics; recurring seasonal pieces like holiday campaigns and tax guides; and high-authority existing pages that already earn backlinks and trust. A fourth bucket is easy to miss: brand consistency across every platform where a business appears.

What should an SEO audit measure now?

Traditional signals remain the baseline after any update: rankings, impressions, click-through rate, and organic traffic. AI-era signals are additive rather than a replacement: citation frequency inside AI answers, brand mentions across the web, referral traffic from AI assistants, and a comparative share-of-voice against competitors in answer engines.

A concrete refresh checklist for an audit looks like this:

• Update any data points, screenshots, or examples that are more than roughly a year old.
• Clarify sections where wording has drifted away from current best practice.
• Expand coverage to include emerging subtopics and adjacent questions.
• Improve internal links so authority flows toward the page and toward supporting entities.
• Strengthen entity relevance by referencing recognized companies, products, and people directly.
• Confirm the page gets re-crawled and re-indexed quickly so the improved version can surface sooner.

The myth to retire

The assumption that publishing more pages automatically produces more AI citations does not hold up under the current retrieval patterns of answer engines. More URLs without strong entity backing tend to create thin, overlapping content that fragments the signals a model needs to trust a brand. A smaller set of authoritative pages that are kept current tends to outperform a large catalogue of lightly maintained ones.

What is the practical workflow for SEO teams?

A repeatable refresh cycle keeps maintenance predictable rather than reactive. Start by identifying pages whose rankings or traffic have slipped. Prioritize ones that already carry backlinks and authority. Check whether the underlying search intent has shifted since publication. Then update facts, add new entities and sources, expand missing coverage, refresh internal links, update freshness markers, and resubmit the URLs for indexing.

AI-visibility tooling is also maturing. Tracking how often a brand is cited, in which contexts, and against which competitors gives a clearer read on share-of-voice than rankings alone. Pairing citation data with traditional analytics produces a fuller picture of where refreshes are paying off.

In competitive verticals such as real estate, banking, eCommerce, and travel, refreshing top-performing assets usually returns more than launching dozens of new low-value pages. The strategic balance is creation plus maintenance, weighted toward maintenance on the URLs that already work.

What should site owners check first?

For a small business, the operator move is a focused audit, not a publishing spree. Pull the handful of pages that already rank or already convert customers and run one question against each: would an AI answer engine trust this version of the page to represent the business right now? If services, pricing, or examples are even a year out of date, the gap is recoverable without writing anything new.

Brand consistency deserves its own line on the checklist. AI systems evaluate a business across its site, directories, review platforms, forums, and social profiles, not the homepage alone. Conflicting addresses, hours, or service descriptions create uncertainty and weaken the entity trust a model needs before it will cite a brand. Cleaning up listings and keeping profiles on-message is now part of the freshness routine rather than a one-time setup task.

The bigger picture

Fresh content for AI visibility is less about producing a steady stream of new posts and more about keeping the right pages accurate, consistent, and citation-ready. As answer engines narrow each query to a small set of trusted sources, the brands that win are the ones whose information holds up the moment a question is asked. Refreshing what a site already owns is no longer maintenance work. It is the core strategy for staying inside the answer.

FAQ

Does publishing more content still help AI visibility?

Not in the way it once did. Classic SEO rewarded publishing volume because more indexed pages meant more keyword targets. AI search engines synthesize one answer and cite only a few sources, so raw volume offers little advantage. Extra pages can also produce thin, overlapping content that fragments topical authority. A smaller set of authoritative pages with regular refreshes tends to perform better than a large catalogue of lightly maintained ones.

What counts as a real content refresh versus fake freshness?

A real refresh changes the substance of the page: updated data and statistics, new examples and screenshots, expanded topic coverage, added FAQs, stronger internal links, and clearer entity references. Fake freshness is changing the publication date or making trivial edits without improving the content, which rarely moves the needle. Visible “Last Updated” timestamps, refreshed metadata, and dateModified schema markup help crawlers notice the change, but they only amplify genuine improvement.

Why does brand consistency across platforms affect AI visibility?

AI systems evaluate a business across its website, directories, review platforms, forums, and social profiles, not the homepage alone. Conflicting addresses, hours, services, or messaging across those surfaces introduce uncertainty and weaken the entity trust a model needs before citing the brand. Keeping information aligned everywhere is now part of content freshness and directly influences whether answer engines treat the business as a reliable source.

Some companies have started celebrating how many AI tokens they burn through, treating consumption as a stand-in for success. Practitioners have begun calling the pattern tokenmaxxing, and it shows up in reported annual AI bills reaching roughly half a billion dollars at one major cloud provider, with no matching improvement in disclosed profit. Gartner has tied this kind of behavior to its forecast that at least 30% of generative AI projects will be scrapped after proof of concept.

For anyone running technical SEO audits, the tokenmaxxing lens matters because the same vanity-metric mindset that bloats LLM bills also bloats pages, sitemaps, and crawl budgets. If a team can’t connect an AI feature to a business outcome, it usually can’t connect a content page to revenue either.

Why Raw Token Counts Are the Wrong Audit Signal

Generative AI is projected to add between $2.6 trillion and $4.4 trillion to the global economy each year, but only where organizations capture real productivity gains. The trouble starts when leaders start quoting token milestones in earnings calls, all-hands slides, or board updates. “Our developers generated 10 billion tokens last quarter” reads as momentum on a dashboard, but if those tokens produced drafts that needed heavy rewrites, hallucinated snippets, or chatbot chatter no customer asked for, the number is a costume, not a result.

Gartner’s 2023 forecast warned that through 2025, at least 30% of generative AI projects would be abandoned after proof of concept because of poor data quality, rising costs, or unclear value. Measuring consumption instead of outcomes accelerates exactly that failure mode: sponsors chase output quantity and never build the instrumentation that would show whether any of it worked.

What Tokenmaxxing Looks Like in Practice

Tokenmaxxing isn’t one bad decision; it is a stack of small incentives that compound. The pattern has a few recognizable shapes that show up across departments:

• Prompt bloat by default. Engineers send massive system prompts for short answers, or chain multiple summarization passes when one would do. Each pass adds to the meter.
• Thin workflow integration. A model is bolted onto an existing process without redesign, so the output is a rough draft that a human has to fix. The fix work is invisible; the tokens are not.
• Internal usage quotas. Some teams set AI usage targets that push employees to route simple tasks through LLMs because the dashboard rewards activity.
• Committed-spend pressure. A multi-year contract with a model vendor creates a budget hole that someone has to fill, so the metric becomes “did we use what we paid for” rather than “did we get value from it.”

That last shape is what made the rumored Amazon Claude bill, reported at around $500 million a year, so visible. A line item that size, without a parallel story about margin or revenue, signals that consumption became the goal.

The Numbers Behind the Failure Rate

Three data points frame how widespread the gap between AI activity and AI value has become:

• Gartner projects more than 30% of gen AI projects will be abandoned by 2025, citing cost, data quality, and unclear value as the main causes.
• A survey by a major analyst firm found that roughly 48% of AI initiatives never progress past the pilot stage, often because organizations cannot show business impact beyond usage stats.
• A 2025 Foundry and CIO.com study reported that only about 14% of CIOs actively track tangible business outcomes from their AI investments, while the rest rely on adoption counts and satisfaction scores that look a lot like tokenmaxxing.

Rita Sallam, Distinguished VP Analyst at Gartner, summed up the disconnect: “The bar for generative AI success is high, and many organizations are struggling to prove and realize value.” When the bar is high and the measurement is loose, projects die quietly in pilot purgatory.

What to Audit on Your Own Stack

If you run technical SEO audits, the tokenmaxxing framework maps cleanly to the way you already check a site. The audit questions are nearly identical: is the input earning its keep, or is it just generating output?

• Tie each LLM call to a measurable event. A chatbot reply should map to a resolution event, a deflection, or a conversion. If a feature can’t name the event it influences, it is decorative.
• Check for chained calls that duplicate work. Multiple summarization or rewriting passes on the same content are the prompt equivalent of redirect chains; they cost tokens without changing the answer.
• Look at committed spend vs. realized value. A large annual contract with a model provider should show up in your cost-per-resolved-ticket or cost-per-conversion data, not just in finance dashboards.
• Track error rate and rework, not just volume. High token output with high human correction rates is a sign the model is doing work twice; so is a content pipeline where every AI draft needs a full editorial pass before it ships.

The same logic applies to content pages. A URL that gets crawled, indexed, and never converts is doing for SEO what a token call without a downstream event does for AI: burning budget without producing outcome. Crawl-budget waste and token-budget waste are the same problem wearing different clothes.

Where the Industry Is Heading

Value-based AI observability is the term gaining ground for tools that correlate LLM traces with business events: tasks completed per dollar, time-to-insight reductions, error-rate improvements, and revenue-influenced pipelines. Advisory firms are pitching outcome scorecards that replace token counts with metrics a finance team can audit. Anthropic and OpenAI have both added granular cost controls, prompt caching, and batch processing, partly because providers recognize that bloated bills without matching outcomes damage trust.

The shift in language is small but telling. Two years ago the question on slide decks was “how many tokens did we consume?” Now it is “what did those tokens actually accomplish?” That is the same pivot SEO has been working through for a decade, from ranking reports to revenue reports, and the same audit discipline applies.

The Audit-Friendly Takeaway

Token volume is a usage signal, not a success signal. The teams that come out ahead will be the ones that refuse to report on tokens alone, and that build lightweight internal scorecards tying every AI call to a named business event. If a feature, page, or prompt cannot point to that event, it is a candidate for the same treatment you would give an orphan URL: measure the cost of keeping it, measure the cost of cutting it, and make a call.

FAQ

What is tokenmaxxing?

Tokenmaxxing is the practice of optimizing for the total number of tokens a company consumes from large language models, such as Claude or GPT, as a vanity metric, without tying that consumption to any measurable business result. It treats raw output as proof of AI maturity.

Why does a reported $500 million Claude bill raise concerns?

A reported annual Claude spend near half a billion dollars, with no comparable disclosed gain in revenue or cost savings, illustrates the risk of decoupling AI investment from business value. The figure suggests that burning tokens had become the goal rather than a side effect of doing useful work.

Which metrics should replace raw token usage?

Outcome-based metrics such as tasks automated per dollar, time saved per process, revenue influenced, error-rate reduction, and cost per resolved customer ticket tie AI activity to financial and operational KPIs. These make it clear whether a given AI spend is paying back its cost or just filling a quota.

At Google I/O 2026, Google showed an updated AI Studio that turns a plain-English prompt into a native Android app, exportable to a phone within minutes. The feature opens with a limited scope (personal utility apps only) and is the clearest mobile signal so far that prompt-driven “vibe coding” is leaving the desktop and reaching the devices most customers carry. Play Store distribution still runs through Google’s normal review process.

For anyone who runs a website, the story matters because the same AI models that now build apps are the ones that crawl, parse, and reason about your pages. A prompt-built weather widget or job-site tracker is only as good as the data the underlying model can find, cite, and trust.

What changed at I/O 2026

AI Studio’s new flow accepts a written description and produces a native Android build that can be side-loaded to a device in a matter of minutes, based on Google’s demo. The model under the hood is Gemini, which means it can reason about APIs, data shapes, and interface patterns the way an engineer would, then hand back a working binary instead of a mockup.

Two scope limits deserve attention:

• Only personal utility apps ship in the initial release. Trackers, calculators, checklists, dashboards. Broader categories remain gated.
• Distribution rules did not change. Anything headed to the Play Store still has to clear Google’s existing review process.

Google also previewed AI-generated widgets at the Android Show, including examples that surface specific weather metrics or suggest recipes. These single-purpose widgets are the first concrete piece of what Google is calling a “generative UI,” where the phone composes interface elements on demand rather than the user navigating a fixed grid of icons. Android president Sameer Samat tempered the vision on stage: “While I don’t think we want to wake up every morning and have our devices have different UI, I do think there’s a level of personalization and customization to the user that could be delightful.”

How this changes a site audit

A site audit used to mean checking titles, schema, speed, and links. With vibe-coded apps pulling live data from your pages, three extra layers now belong on the checklist.

1. Can the model read your structured data

A prompt-built widget that surfaces today’s booked appointments, this week’s ad spend, or invoices past 30 days needs clean, machine-readable source material. Run your key pages through a structured-data validator and confirm that prices, dates, product names, and review counts parse without errors. If your JSON-LD throws warnings, the widget will quietly render blanks or hallucinate.

2. Is your business entity consistent across the web

Gemini, ChatGPT, Claude, and similar models reason about a business using the public mentions they can find. If your name, address, phone, and hours disagree between your site, your Google Business Profile, and the top directories, a vibe-coded app that tries to look you up will return a garbled answer. Audit the top ten citations for your brand and align the fields exactly.

3. Are your contact paths reachable, not just visible

An AI agent that wants to act on your business (booking an appointment, pulling a price, confirming stock) needs an endpoint it can call. Static contact pages with no schema, no API, and a form protected by an aggressive CAPTCHA will register as a dead end. Check that at least one contact method is reachable without JavaScript-rendered gates, and that the markup includes the right Schema.org type.

Why Apple matters to this story

Apple is reportedly building a prompt-driven path for iOS Shortcuts that lets users describe an automation in words (for example, “open the transit app when I get to the bus stop”) instead of assembling it block by block. iOS 27 is the expected vehicle. The mechanics differ from Google’s full-app generator, but the user-facing pattern is the same: describe the outcome, skip the assembly.

Once both platforms treat “describe what you want” as a first-class input, internal tooling for small operators stops being a budget question and becomes a prompt-writing question. The bottleneck moves from engineering to knowing which workflow to describe first.

What to run on your own site this week

• Validate the structured data on every page an AI-built widget might pull from: product, service, FAQ, local business, and review schema.
• Pull your top ten directory listings and diff the core fields against your own site. Fix every drift.
• Open a private browsing window and ask Gemini, ChatGPT, and Claude to describe your business. Note what they get wrong; those gaps are your citation list.
• Check that at least one contact endpoint on your site responds to a curl request without a CAPTCHA or a render-blocking script.
• Crawl your own pages with a headless browser disabled. Anything that fails to load without JavaScript is invisible to most AI crawlers today.

The realistic ceiling

Reviewer Allison Johnson offered a useful caution after the I/O demos: “I’ve heard a lot of promises over the past few years from tech company execs about how AI will fundamentally change how we interact with mobile devices.” Until widgets ship at scale and survive contact with real users, the feature is a demo, not a platform shift. The build time claim (“a matter of minutes”) and the personal-utility gate are both signals that Google is keeping the scope tight on purpose.

For site owners, that caution cuts both ways. The technology is real, the demo is real, but the gap between a working widget on a reviewer’s phone and a stable platform that other people’s apps depend on is still wide. Audit your pages as if the apps are coming, because they are, and use the lead time to make sure your data is the kind an AI can use without guessing.

FAQ

What did Google actually announce at I/O 2026 about building Android apps?

Google showed an update to AI Studio that accepts a plain-English prompt and produces a native Android build, exportable to a phone within minutes. The release is limited to personal utility apps at launch, and Play Store distribution still goes through Google’s normal review process. Gemini powers the generation step, so the model can reason about APIs, data shapes, and interface patterns.

How should a site audit change because of vibe-coded apps?

Add three checks to a standard audit. Validate that structured data on key pages parses cleanly so AI-built widgets do not render blanks or hallucinate. Confirm that the business entity (name, address, phone, hours) is identical across your site and the major directories, since AI models reason from public mentions. Make sure at least one contact endpoint is reachable without JavaScript or CAPTCHA so an AI agent can act on the business.

Is Apple doing the same thing for iPhone?

Apple is reportedly building a prompt-driven path for iOS Shortcuts that lets users describe an automation in words instead of assembling it block by block. iOS 27 is the expected vehicle. The mechanic is different from Google’s app generator, but the user-facing pattern matches: describe the outcome and skip the manual assembly.

When DeepSeek published API rates in early 2025 that matched GPT-4o on standard benchmarks at $0.14 per million input tokens and $0.28 per million output tokens, the economics of every AI product on the market shifted overnight. OpenAI, Google, and Anthropic responded within 90 days, with Google cutting Gemini pricing by as much as 85%. The result is the sharpest pricing compression in enterprise software history, and it changes the calculus for every site owner running AI-driven features or counting on AI agents to find their pages.

What actually changed in early 2025

DeepSeek-V3 launched in December 2024, followed by the reasoning model DeepSeek-R1 in January 2025. Both matched frontier proprietary models on MMLU, HumanEval, and MATH benchmarks while charging a fraction of the going rate. The pricing was not a loss leader built on venture cash. It reflected architectural choices that cut the real cost of serving tokens.

Two design decisions drove the gap. First, DeepSeek-V3 uses a Mixture-of-Experts (MoE) layout with 671 billion total parameters but only about 37 billion active per forward pass, so most of the model sits idle on any given query. Second, Multi-Head Latent Attention (MLA) compresses the key-value cache during inference, trimming memory and compute. DeepSeek also reported training V3 on a cluster of Nvidia H800 GPUs for roughly $5.6 million, a figure widely debated, but the inference efficiency is reproducible in independent benchmarks.

The price gap that broke the market

Here is how the frontier API rates compared per million tokens at launch:

• DeepSeek-V3: $0.14 input, $0.28 output (cache-hit input as low as $0.014).
• OpenAI GPT-4o: $2.50 input, $10.00 output, roughly 18x and 36x more expensive.
• Google Gemini 1.5 Pro: $1.25 input, $5.00 output. Google countered with Gemini 2.0 Flash at $0.10 input and $0.40 output, an 85% cut.
• Anthropic Claude 3.5 Sonnet: $3.00 input, $15.00 output. Anthropic followed with Claude 3.5 Haiku at $0.25 input and $1.25 output.
• DeepSeek-R1: $0.55 input, $2.19 output, undercutting OpenAI o1 at $15.00 input and $60.00 output by roughly 27x on both sides.

OpenAI added tiered caching discounts and launched GPT-4o mini in the same window. The direction across every lab was the same: down, and fast.

Three forces pushing inference cost toward zero

The cuts are not a one-time event. They reflect structural pressure that will keep compressing margins.

Open-weight releases from DeepSeek, Meta (Llama), Mistral, and others prevent any proprietary lab from holding a large premium for long. Hardware efficiency is compounding, with Nvidia’s Blackwell generation, custom inference silicon from Groq and Cerebras, and serving tricks like speculative decoding each cutting the cost per token. And usage is exploding, because cheaper tokens unlock applications that were previously uneconomical, which drives aggregate consumption up even as unit prices fall, a textbook Jevons paradox. The global AI market was projected by Statista to reach $243 billion in 2025, and the price war is reshaping how that spend gets allocated.

What to audit on your own site

For teams building or buying AI features, the price war is a green light to revisit every line item tied to inference. AI agents that crawl business listings, verify contact details, evaluate reputation signals, and route leads to your site or a competitor are now running on dramatically cheaper tokens. That has direct consequences for technical SEO work.

Start with these checks:

• Structured data and NAP consistency. Agents verifying business information will cross-check name, address, and phone across many sources. Run an audit to confirm your structured data matches what is rendered on the page and what appears in major listings. Inconsistencies now get caught faster and routed around faster.
• Directory presence. The directories and platforms that AI agents query are no longer optional listings. They are infrastructure for AI-mediated discovery. Confirm your business is present and accurate on the sources your buyers’ agents actually pull from.
• Server response for agent traffic. If you block or rate-limit user agents that look automated, you may be hiding from the very crawlers whose recommendations drive leads. Review your robots.txt, firewall rules, and CDN rate limits with an eye to legitimate AI crawlers, not just the big search engine bots.
• Content freshness signals. Cheaper inference means more agents re-checking pages on shorter cycles. Make sure publish and update dates are accurate, sitemaps are current, and canonical tags are correct, so re-checks see the freshest version of your page.
• Page speed on the routes that get cited. When an agent decides which source to surface, response time and Core Web Vitals still matter. A page that is slow to render or returns intermittent 5xx errors gets deprioritized by agents that have to choose among many candidates.

What this means for product builders

Products that were economically marginal six months ago, including customer support bots handling millions of daily tokens, real-time content moderation, and AI lead qualification, are now within reach of small teams. A workload that produced a five-figure monthly inference bill at GPT-4 rates can run for a fraction of that on the new pricing floor. If you shelved an AI feature in 2024 because the unit economics did not close, it is worth re-modeling with current rates.

Margins across the model layer will compress. Labs with high fixed costs may struggle, and the surviving players will likely push toward platform plays, enterprise tooling, and application-layer revenue rather than relying on raw token sales. The companies that win the next cycle will be the ones building on top of the cheap-inference layer, not the ones still trying to charge for access to the model itself.

FAQ

How much cheaper is DeepSeek than OpenAI right now?

DeepSeek-V3 charges $0.14 per million input tokens and $0.28 per million output tokens, compared to GPT-4o at $2.50 input and $10.00 output. That is roughly 18x cheaper on input and 36x cheaper on output. For reasoning workloads, DeepSeek-R1 at $0.55/$2.19 undercuts OpenAI o1 at $15.00/$60.00 by about 27x on both sides.

Did Google and Anthropic actually cut their prices in response?

Yes. Google introduced Gemini 2.0 Flash at $0.10 input and $0.40 output per million tokens, an 85% reduction from Gemini 1.5 Pro. Anthropic launched Claude 3.5 Haiku at $0.25 input and $1.25 output, down from Claude 3.5 Sonnet’s $3.00/$15.00. OpenAI added tiered caching discounts and introduced GPT-4o mini. Both labs also expanded free tier access.

What should I audit on my site now that AI agents are cheaper to run?

Verify that your structured data, NAP information, and directory listings are consistent and current, since cheaper inference means more agents cross-checking them. Review robots.txt, firewall, and CDN rules so legitimate AI crawlers are not blocked alongside scrapers you want to keep out. Confirm canonical tags, sitemaps, and publish dates are accurate, because agents are re-checking pages on shorter cycles. Finally, check Core Web Vitals and 5xx rates on the pages most likely to be cited, since agents deprioritize slow or unreliable sources.

U.S. app installs for DuckDuckGo climbed as much as 30.5% in a single day during late May 2026, and iOS installs spiked nearly 70% on the peak day, according to figures the company shared covering May 20 through May 25. The driver was not a marketing campaign. Users were leaving Google after its annual developer conference rolled out AI-first changes to Search, and DuckDuckGo’s “no AI” search page became the clearest alternative on offer. For site owners running technical audits, the takeaway sits underneath the privacy framing: search behavior is splitting across multiple surfaces, and a one-engine playbook now leaves gaps.

Why a 30% Spike on a Small Player Matters

DuckDuckGo remains a minor player next to Google in U.S. search, and the company does not command a share that should trigger panic on its own. The number to watch is the direction, not the size. A measurable, sustained migration of users toward engines that allow them to switch AI off is the kind of structural shift that rewires where your next customer starts their query. Anyone who has watched a single traffic source decay year over year knows how quietly that loss compounds.

The May 20 to May 25 window captured six straight days of growth, with a U.S. app install lift averaging 18.1% week over week against the prior week. On iOS specifically, average weekly install growth hit 33%, with a near-70% peak on the strongest day. Visits to DuckDuckGo’s AI-disabled search page grew 22.7% on average week over week, peaking on May 24. Founder and CEO Gabriel Weinberg tied the surge directly to users being “force-fed” AI features without a real opt-out, and to results he described as getting “worse, not better.” That language captures the friction, and friction is what drives switching.

What Actually Changed in Google Search

Google used its developer conference to push a sweeping overhaul that swaps the familiar list of blue links for AI-generated answers capable of summarizing information, completing tasks, and monitoring queries in the background. Critics have argued the AI-first design undercuts the open web by siphoning traffic away from the publishers and small businesses whose content the answers are built on. The technical audit implication is direct: when an answer engine responds without a click, the page that earned its information may never receive a visitor.

AI Overviews now ship with follow-up chat prompts attached, which nudges even users who skip AI Mode toward a chatbot-shaped experience. The complaint is less that AI exists in search, and more that the AI surface has become hard to avoid. For site owners, that means the question of whether your content is cited inside an AI summary now competes with the older question of whether your page ranks in the classic results.

How the Alternatives Position Themselves

DuckDuckGo’s response was a dual play. The main app continues to grow, and the company actively promotes a dedicated “no AI” search page that disables AI-written summaries and synthetic image results by default, returning a deliberately traditional result set. That same product line still includes Duck.ai, a chatbot that routes to models from OpenAI, Anthropic, and Meta while stripping IP addresses, declining to permanently store chat histories, and refusing to use conversations for model training. The stance is control and privacy, not a blanket rejection of AI.

Other engines are picking up traffic with different hooks:

• Brave offers customizable result filters it calls Goggles, plus an AI on/off toggle.
• Ecosia markets reforestation funding built into its search activity.
• Startpage returns familiar Google results while acting as a privacy layer that does not hand over personal data.

All three are Chromium-based, which lowers the switching cost for users because most Chrome extensions still carry over. Leaving Google no longer means abandoning a familiar toolchain, and that drop in friction is exactly what makes a fragmentation signal durable.

What an Auditor Should Check Right Now

If you run technical SEO audits, the install spike is a prompt to expand the surfaces you test, not a reason to abandon Google. A practical first pass covers four areas.

Listing data consistency. Every alternative engine relies on the same underlying signals: business name, address, phone, hours, and category. Pull those fields and compare them across Google Business Profile, Apple Maps, Bing Places, Yelp, and the major data aggregators. Any mismatch becomes a candidate for incorrect answers in AI summaries or alternative SERPs.

AI-contactability. As more queries resolve inside conversational agents, the test is whether an AI can actually surface and reach your business. Run queries that describe your service in natural language across ChatGPT, Perplexity, Google AI Overviews, and DuckDuckGo’s AI-assisted surfaces. Note which answers name your business, which cite your URL, and which return a competitor instead.

Structured data coverage. Review your schema markup for Organization, LocalBusiness, Product, and FAQPage types. Engines that feed AI summaries lean heavily on structured data to ground their answers, so missing or inconsistent markup translates directly into missed citations.

Brand presence across non-search surfaces. Fragmentation does not stop at search engines. Conversational agents pull from social profiles, review sites, and press coverage when forming answers. Confirm your profiles on the major platforms are active, consistent, and link back to a canonical domain.

How to Read the Rest of 2026

Expect the menu of alternatives to keep widening through 2026 rather than consolidating back into a single dominant surface. DuckDuckGo’s combined offering, an AI-free search page plus a privacy-first chatbot, points at the broader direction: the split is not “AI versus no AI.” It is user-controlled AI. Some visitors will ask a Google AI Overview. Some will type into DuckDuckGo’s classic results. A growing share will hand the question to a conversational agent and never visit a SERP at all.

For site owners, that means the KPI discussion needs to widen. Raw click counts from one engine tell less of the story than they did two years ago, because zero-click answers and multi-surface discovery both pull weight. The businesses that hold up are the ones whose underlying data is clean enough that any surface, AI or classic, can find and represent them accurately.

The Practical Takeaway

DuckDuckGo’s install spike is small in absolute terms, but it is a leading indicator of a market that is splitting along user-controlled lines. The audit response is not to chase every new engine. It is to make the data about your business so clean, so consistent, and so well structured that any search surface can answer questions about you correctly. Optimize for being discoverable across the whole landscape, not for placement on a single company’s homepage.

FAQ

What did DuckDuckGo report about its U.S. app installs in May 2026?

DuckDuckGo reported an average week-over-week increase of 18.1% in U.S. app installs for May 20 to May 25, 2026, with growth running for six straight days and peaking at a 30.5% increase on May 25. iOS installs grew an average of 33% week over week, peaking at nearly 70% on the strongest day.

Why are users switching away from Google Search right now?

Users are reacting to Google’s rollout of AI-first changes at its annual developer conference, including AI Overviews with follow-up chat prompts that make the experience feel like a chatbot. DuckDuckGo founder Gabriel Weinberg said users were being “force-fed” AI features without a meaningful opt-out and that results were getting “worse, not better.”

Which search engines besides DuckDuckGo are gaining attention in 2026?

Brave, Ecosia, and Startpage are all picking up attention. Brave offers customizable Goggles filters and an AI on/off toggle, Ecosia markets reforestation funding, and Startpage returns Google-style results without handing over personal data. All three are Chromium-based, which keeps most Chrome extensions working after a switch.

Anthropic released Claude Opus 4.8 on June 16, 2026, the company’s most capable general-purpose model to date, with a heavy emphasis on extended reasoning, autonomous tool use, and self-auditing safety layers. The release lands as a growing share of organizations experiment with AI agents, and it directly raises the bar for what those agents can do without human oversight. For site owners and technical SEOs, the practical question is not whether the model is impressive on a leaderboard, but whether the pages those agents crawl, read, and cite are clean enough to be trusted.

Why Opus 4.8 Matters for Anyone Auditing a Website

The newest Claude extends the model’s hidden reasoning transcript into the thousands of tokens before it answers. That depth lets it backtrack, verify sub-conclusions, and weigh counterfactuals. In parallel, Anthropic rebuilt the function-calling loop so the model plans sequences of API calls, web searches, and database queries on its own, then self-corrects when intermediate results look wrong. A new “reflect-and-resume” mechanism lets the agent pause, inspect its own state, and decide whether to keep going, pivot, or stop.

Vision also got an upgrade: high-resolution diagrams, blueprints, and dense charts now get structural parsing rather than caption-level summaries. A parallel safety head runs during generation and flags potential policy violations or hallucinations before tokens reach the user.

For an auditor, the consequence is concrete. A reasoning agent that can sustain thousands of internal steps and cross-reference multiple sources will surface contradictions and stale data far more reliably than earlier models did. A NAP mismatch that an older LLM would have glossed over is now a candidate for a flag.

What the Benchmarks Actually Imply

Anthropic reports directional gains across the standard suite. GPQA Diamond, the graduate-level reasoning test, shows clear improvement on multi-step physics, biology, and chemistry problems that demand hypothesis testing. SWE-bench Verified, which measures whether a model can resolve real GitHub issues end to end, posts a substantially higher pass rate. On τ-bench, which simulates retail and airline customer service interactions, the success rate roughly doubles compared with the previous Opus family, driven by better policy lookups, data entry, and multi-turn decision loops.

Tool-use accuracy has tightened as well: fewer spurious or redundant API calls, because the planning and validation checks run inline. And on Anthropic’s most adversarial internal probes, harmful completions stayed below 0.5%, which the company frames as a new safety record for an unrestricted deployment.

The relevant signal for site owners is not the absolute score. It is that Opus 4.8 holds coherent plans across long sequences of tool calls, which means an agent working through your site, your directory listings, and your third-party profiles can chain facts together in ways a shorter-context model could not.

How to Audit Your Site for the Agent Era

Start with the data an agent would lean on first: business identity, contact details, and service descriptions. Run an NAP consistency check across your own pages, the major directories, social profiles, and any industry listings. Earlier models would often pick one source and ignore the others. A model that sustains long reasoning is more likely to notice when your address on your site disagrees with the address on your Google Business Profile, and to discount the source it judges less trustworthy.

Then audit your structured data. Validate schema.org markup on organization pages, service pages, and product pages. Agents that plan multi-step research will look for explicit machine-readable hints before they fall back to prose. Review pages for clarity and recency: update dates, authorship, and changelog signals help an agent decide whether to trust a piece of content. Where possible, expose a public changelog or feed so the agent has a hook for freshness rather than having to infer it.

Audit your third-party footprint too. Reviews, citations, and directory entries are exactly the kind of cross-source signals a long-horizon agent now weighs. Pay particular attention to inconsistencies in category tags, hours, and service names, since those are the fields most often duplicated incorrectly across listings.

Finally, check the documents an agent might pull and parse. PDFs, contracts, pricing sheets, and spec documents should be text-searchable, well-structured, and free of image-only text. The upgraded vision layer helps, but agents still prefer clean text and tagged headings when they can get them.

What to Watch in the Next Release Wave

Anthropic confirmed that Opus 4.8 will form the backbone of the advanced tier of Claude Assistant and the API endpoints used by enterprise developers building autonomous workflows. A streaming “thinking trace” viewer and an agent run-log dashboard are being released alongside the model to help teams debug long-running integrations.

A distilled, lower-latency variant codenamed Opus-4.8-Nova is in early testing for on-device and edge deployments, which points toward Opus-class reasoning reaching mobile and industrial hardware within months. Anthropic’s research arm also plans to publish papers on the interpretability methods introduced with Opus 4.8, giving the community a window into how the deliberative layer decides what to surface.

None of that changes today’s audit checklist, but it raises the cost of deferring it. As reasoning depth increases, the gap between a site that is AI-readable and one that is not widens. The pages that survive contact with an Opus-class agent are the ones whose facts are consistent across every surface an agent can reach.

FAQ

What is Claude Opus 4.8?

Claude Opus 4.8 is Anthropic’s most advanced large language model, released June 16, 2026. It extends the Opus line with a much larger hidden reasoning budget, native multi-tool orchestration, upgraded vision, and a parallel safety head that audits outputs during generation.

How does Opus 4.8 differ from earlier Claude models?

The biggest shifts are the multi-thousand-token reasoning headroom, the rebuilt tool-use loop that plans and self-corrects, the reflect-and-resume mechanism for long-running agents, and the parallel safety monitor. Together they let the model sustain longer workflows with fewer cascading errors than Claude 3.5-era systems.

What should site owners audit first for agent visibility?

Start with NAP consistency across your site and every major directory, then validate schema.org on organization, service, and product pages. Confirm that PDFs and key documents are text-searchable, that authorship and update dates are visible, and that third-party listings have matching category tags and service descriptions. A long-horizon agent is more likely to flag, and to downweight, any contradiction it finds across those surfaces.

Across ecommerce, organic sessions are slipping year over year while revenue holds steady or climbs. The shopper who used to click through early research now gets answers on the results page, through AI summaries, product grids, review snippets, and LLM chats, so the click that survives arrives later, with more intent and more value. For site owners running technical audits, the practical problem is that the traffic chart you have trusted for years can now tell the opposite story about whether SEO is working.

What changed in the ecommerce search journey

Ten years ago a shopper searching “best leather belts” would bounce through review sites, category pages, and affiliate posts before refining the query, opening more tabs, and finally landing on a product listing page or buying guide on a retailer domain. Most of that messy research happened on the merchant’s site, which made reporting straightforward.

Modern mobile SERPs behave differently. Before a shopper ever reaches a retailer, results pages can surface sponsored products, organic product grids, review snippets, prices, discounts, star ratings, filters, and an AI Overview. In effect, the SERP is starting to behave like the product listing page. The shopper compares items, scans reviews, and checks price ranges without clicking through. When the shopper asks an LLM instead, the shortlist can arrive before any site is visited.

That shift pulls awareness, interest, and consideration off the merchant’s domain. SEO still matters, but the click now arrives later. By the time someone reaches a retailer, they often already know what they want, with less browsing and less comparison, and much closer to purchase. The product detail page, long treated as the final step, is increasingly the first page a shopper sees.

Why the old dashboard hides the real signal

For years, total organic clicks were how teams judged whether they were keeping pace. That signal is weak now. Rankings can hold, product visibility can stay strong, and motivated shoppers can still reach the site, yet total clicks fall because the SERP satisfied the research stage first. With AI Overviews and shopping modules answering more questions on the results page, a sinking traffic line can be misread as SEO failure. For a small operator with a lean budget, that misread can kill funding for the work that is keeping revenue alive.

SEO has always had a “Red Queen” quality. The metaphor from Lewis Carroll fits: teams have to keep running just to stay in the same place, refreshing pages, fixing technical debt, updating templates, and responding to competitors just to hold position. When the click that survives is worth more than several of the clicks that used to arrive, smaller traffic can sit beside stronger revenue. The shape of the funnel changed, not the demand behind it.

Which ecommerce SEO KPIs should you trust now?

• PDP clicks, where the money is. Clicks into product detail pages tell you far more than broad traffic totals when the surviving shopper arrives later in the funnel.
• PDP conversion rate. If a ready-to-buy visitor lands and does not convert, something is missing: shipping, sizing, returns, or context the page assumes the shopper already saw elsewhere.
• Organic revenue, the language leadership understands. When someone asks whether SEO works, revenue carries the argument.
• Click-through rate. In crowded ecommerce SERPs the question is not only whether your product appears but whether your listing earns the click.
• Merchant Center clicks and impressions. Feed quality, titles, images, pricing, promotions, and review signals shape how products look in shopping surfaces.
• Year-on-year traffic for PLPs, blogs, and evergreen content. Treat this as the metric most likely to mislead. Seasonality, a competitor’s price change, a core update, and a PR spike all move at once.

Before-and-after reporting fails for that reason: everything moves together. The fix is controlled SEO A/B testing, where the only honest question becomes what happened while your change was live.

What to audit on your own pages

If you are running a technical SEO audit today, the checklist has expanded. Start by checking AI contactability: whether AI assistants and agents can find, describe, and recommend your business when a shopper asks. Then tighten business listings so name, price, availability, and reviews are consistent everywhere a product grid pulls from. If a listing has not been claimed and verified, the product may not be eligible to appear at all.

Audit the PDP as a landing page. If product grids and AI-assisted journeys surface product detail pages directly, those pages have to do more jobs: reassure the shopper they are in the right place, surface delivery, returns, sizing, availability, reviews, and specs, and answer the last questions before purchase. Work that older journeys pushed onto guides and PLPs now has to live on the page where conversion happens.

Pair that work with controlled testing using Google Merchant Center feed data, and you get a reporting story that survives a confusing dashboard.

Should you build reporting on AI-native metrics yet?

Some teams will be tempted to jump straight to brand sentiment inside AI models, share of voice in AI chats, and query clusters in LLM tools. Watch these, but do not yet build board reporting on them. Outputs vary, personalization changes answers, and the same prompt may not return the same response twice. There is no reliable AI “search volume” data equivalent to classic search. Interesting, but not ready to carry your reporting.

The bigger picture for ecommerce operators

Ecommerce is not shrinking and organic is not over. The commercial demand is still there; the journey just changed shape. Some of the clicks once counted as SEO wins are being displaced into SERP features, AI summaries, and LLM chats, and the click that remains is often worth more. Meet shoppers at the moment of intent, make the page they land on do the convincing, and prove your value with revenue and controlled tests instead of a number that no longer means what it used to.

FAQ

Why is organic traffic falling while revenue stays flat or grows?

More of the shopping research now happens before the click. SERP features, product grids, AI Overviews, and LLM chats let shoppers compare products, prices, and reviews without visiting the site. The early, low-intent research clicks once counted are absorbed upstream, while high-intent shoppers still click through when ready to buy. The result is fewer clicks, each carrying more buying intent, which can produce a smaller traffic number next to stronger revenue.

Which ecommerce SEO KPIs should I watch in 2026?

Weight metrics tied to intent and money rather than raw volume. PDP clicks, PDP conversion rate, organic revenue, click-through rate, and Google Merchant Center clicks and impressions matter most. PDP clicks show whether motivated shoppers reach conversion pages; conversion rate reveals friction; Merchant Center data shows how appealing products look in shopping surfaces. Year-on-year traffic for PLPs, blogs, and evergreen content is the metric most likely to mislead.

Are PDPs really becoming landing pages?

Yes. Older models assumed shoppers entered through PLPs, category pages, or buying guides and saw context before reaching the product. Now product grids and AI-assisted journeys surface PDPs directly, so the product page is often the first thing a shopper sees. The PDP has to carry trust signals, delivery and returns information, sizing, availability, reviews, and specs that earlier pages used to provide. The landing page is now the conversion page.

Google’s updated Guide to Optimizing for Generative AI Features on Google Search landed inside Search Central’s SEO Fundamentals section, tucked next to the SEO Starter Guide. The placement signals a clear position: optimizing for AI Overviews still falls under standard SEO. More importantly, the guide warns that commodity content, the kind any writer or model could produce, faces the highest risk of being bypassed by AI-generated answers.

For site owners running technical SEO audits, this guide is a checklist disguised as documentation. It confirms how retrieval-augmented generation works, names the technical blockers that keep pages out of AI features, and draws a hard line between replaceable and irreplaceable content.

How do AI Overviews actually pull your pages into answers?

Two mechanisms in the guide explain why traditional SEO still drives AI visibility. The first is retrieval-augmented generation (RAG). AI Overviews are assembled from real pages already in Google’s index. If your page is indexed, ranks well, and meets snippet eligibility requirements, it can be cited directly in an AI Overview.

The second is query fan-out. When a user asks a complex question, Google runs several related searches simultaneously and combines the results into a single generated answer. Your page doesn’t need to match the user’s exact phrasing. A comprehensive page that answers one of the related sub-questions can surface even when the original query wording differs. This makes topical depth and semantic coverage more valuable than narrow keyword targeting.

What technical settings quietly block pages from AI Overviews?

The guide specifies that to appear in any generative AI feature, a page must be indexed and eligible to show a snippet. A page carrying a nosnippet tag fails that eligibility check and stays invisible to AI Overviews, even if it ranks strongly for traditional results.

Many SEO audits treat nosnippet as a minor setting. In an AI-driven search environment, a stray nosnippet tag can silently remove your most valuable pages from the answer layer. The audit step here is straightforward: check your high-traffic pages for nosnippet directives, confirm snippet eligibility, and verify that nothing in your meta robots configuration is blocking generative features.

Which tactics does Google tell you to stop using?

The guide includes a section titled “What you don’t need to do.” The list reads like a direct response to the consulting market that has built products around AI-search optimization:

• llms.txt files. Googlebot gives them no special treatment. These files may still matter for Anthropic, OpenAI, and Perplexity crawlers, which operate differently.
• Chunking content into short paragraphs for AI parsing. Google’s systems already understand context across multi-topic pages.
• Rewriting copy to match AI phrasing. The model handles synonyms and semantic meaning. A page about fixing a lawn full of weeds doesn’t need that exact string to be cited.
• Inauthentic brand mentions. Planting fake references across forums and roundups does nothing; standard spam policies apply.
• Over-investing in structured data as an AI lever. No schema markup unlocks AI Overview eligibility. Keep structured data for rich results, not as an AI access pass.

What is the non-commodity content test?

The most actionable idea in the guide is the distinction between commodity and non-commodity content. Google uses “7 Tips for First-Time Homebuyers” as its example of commodity content: common knowledge, available from anyone, adding no unique insight.

Non-commodity content is the first-hand account, original data, or named experience that only you could produce. Google’s framing is a question worth applying to every page on your site: “Are we creating something useful enough that people, and AI systems, would miss it if it disappeared?”

If a generative model could write your page in seconds, that page was never a differentiator. The audit implication: flag any page that could be reproduced by an AI with no loss of value. Those are the pages most likely to be replaced by AI Overviews rather than cited by them.

What should your audit checklist include?

The guide points toward a practical, self-serve workflow that doesn’t require a consultant:

1. Run a non-commodity audit on your top-performing pages and flag anything a model could reproduce verbatim.
2. Audit snippet eligibility across high-value pages, checking for stray nosnippet tags or meta robots directives that block generative features.
3. Consolidate thin or overlapping cluster pages before publishing new ones; depth beats volume in a fan-out retrieval model.
4. Stop allocating budget to llms.txt files and AI-specific schema for Googlebot.
5. Invest in content formats AI can’t generate: first-hand results, original research, named professional experience.
6. For e-commerce and local businesses, audit your feed and listing layer, the structured business data that AI agents will pull from.

How does this change your measurement approach?

The guide drops against a backdrop of falling click-through rates on queries where AI Overviews appear at the top of results. When the answer is generated on the results page, fewer users click through to the source site.

This shifts what you should measure. Traditional ranking position still matters because RAG pulls from indexed, ranked pages. But you now have a second visibility layer: citation frequency inside AI Overviews. Tracking which of your pages get cited, and for which query fan-out sub-questions, becomes a new audit dimension. Tools that only check blue-link rankings will miss this layer entirely.

What is Google signaling about the future of search?

The guide arrived a week after Google I/O, where the company confirmed AI Overviews reach billions of users and previewed AI agents that browse the web autonomously. Search is moving from a list of links to a place where tasks get completed.

That direction raises the stakes on machine-readability. If an automated agent needs to find your business hours, read your service descriptions, and complete a booking, your structured business data, your feed quality, and your snippet eligibility all become gating factors. The audit frame expands accordingly: you’re no longer just optimizing for human clicks, you’re optimizing for automated retrieval by systems that act on your behalf.

The underlying message is consistent with what Google’s guide states directly: AI Overviews are built through retrieval from real indexed pages, so the same fundamentals that earn rankings also earn AI citations. No new playbook is required. The work is auditing which pages are commodity, confirming technical eligibility, and investing in the content only your business could produce.

FAQ

Does a nosnippet tag actually block my pages from AI Overviews?

Yes. The guide specifies that to appear in any generative AI feature, a page must be indexed and eligible to show a snippet. A page carrying a nosnippet tag fails eligibility and stays out of AI Overviews, regardless of its traditional ranking strength. Many audits treat nosnippet as a minor setting, but in an AI-search environment, a misapplied tag can silently remove high-value pages from the answer layer.

What is the non-commodity content test from Google’s guide?

It’s an evaluation the guide implies: could a generative AI model produce an equally useful version of this page? If yes, the page is commodity content, common knowledge available from anyone, the type AI Overviews are built to absorb. Non-commodity content is first-hand experience, original data, or named insights only you could publish. Google’s phrasing: “Are we creating something useful enough that people, and AI systems, would miss it if it disappeared?”

Is AI search optimization a separate discipline from SEO?

No, according to Google. The guide was published inside Search Central’s SEO Fundamentals section and explicitly folds AEO and GEO back under ordinary SEO. Because AI Overviews are built through retrieval-augmented generation from indexed pages, the same fundamentals that earn rankings also earn AI citations. A new playbook or a specialist retainer is not required; solid SEO basics and original content are the stated path forward.

DuckDuckGo recorded a 22.7% average weekly rise in visits to its AI-free search page between May 20 and May 25, peaking at 27.7% on May 24. iOS app installs climbed 33% on average and spiked 69.9% on May 25. The timing lines up with public discussion about Google’s push into AI-generated answers and a user base actively seeking alternatives with stronger opt-out controls.

What the traffic numbers actually show

The headline figures are unusually concentrated. A near-70% single-day install spike on iOS is not the kind of lift that comes from a routine app store placement or a minor product update. Combined with a 27.7% peak traffic lift on May 24, the pattern points to a deliberate shift by users, not a passing curiosity. DuckDuckGo also reported an 18.1% week-over-week rise in US app installs across platforms during the same window, suggesting the iOS spike was part of a broader move rather than an isolated Apple effect.

For site owners, the first question to ask is whether any of that traffic represents a real audience worth optimizing for. DuckDuckGo remains a small share of overall search activity compared with Google, which still holds roughly 85% market share, so a single-engine audit will miss most of the picture. But a concentrated spike among users who actively chose an AI-free experience is a signal about intent, not just volume.

Why users are choosing an opt-out path

DuckDuckGo frames the product around user choice rather than rejection of AI. The company offers AI features through Duck.ai, including GPT-5 mini and Claude Haiku 4.5, but routes them through a privacy-preserving interface that lets people decide when AI is involved in their results. CEO Gabriel Weinberg put the contrast plainly: Google is force-feeding AI with no way to opt out, and as a result their results are getting worse, not better.

The interesting detail for anyone auditing a site is the framing. The users moving to DuckDuckGo are not anti-AI in a blanket sense. They want the option to turn AI features off. That distinction matters because it changes what they expect from the pages they land on. Users who have deliberately chosen a quieter search experience are less tolerant of pages loaded with auto-playing video, intrusive interstitials, or AI-generated filler text. If your site leans on those patterns, the audience arriving from DuckDuckGo is the first place that friction will show up in engagement metrics.

What this means when you audit your own pages

Search fragmentation is not new, but the AI layer adds a new variable. A standard technical SEO audit checks how a page renders, how it loads, and how it ranks in Google. An audit that accounts for AI-driven discovery also needs to check how a page is parsed by systems that pull snippets, citations, and structured answers from multiple sources.

Start with three things that are easy to verify on your own pages.

• Structured data markup. Different AI systems parse schema differently, so inconsistent or missing markup cuts you off from answer surfaces on some platforms while working fine on others. Run your key templates through a structured data validator and confirm Organization, WebSite, Article, and Product types are all present where they apply.
• Citation consistency. AI assistants pull business information from directories, knowledge panels, and review sites. If your name, address, phone, and service descriptions disagree across profiles, the model has to pick one, and it may not be the version you prefer. Pick five core directories, compare them by hand, and fix the discrepancies.
• Content quality signals. AI summaries pull from sources that are clear, factual, and easy to extract. Pages padded with vague introductions or repeated boilerplate are harder to cite cleanly. Read your top landing pages and ask whether the main claim is stated in the first two sentences in a way a non-human reader could lift it.

How to think about AI visibility across engines

Google’s search revenue still grew 19% in Q1 2026, so Google’s own AI surfaces remain the largest single channel for AI-mediated discovery. Treating Google as the only target is still the rational default for most businesses. The DuckDuckGo spike is a reminder that it is no longer the only audience that matters.

Brave Search and Startpage have also seen increased interest from users exploring engines with granular AI controls. Each of these surfaces parses content a little differently, weighs citations a little differently, and exposes AI features on different terms. A page that performs well on Google can still be invisible to a smaller AI-driven engine simply because the structured data is missing or the entity information is inconsistent. The audit work is the same, but the tolerance for sloppy implementation is lower when you are trying to show up across the ecosystem rather than just one engine.

The practical posture is to treat AI visibility the way the industry treated mobile a decade ago. You do not abandon desktop, but you make sure the foundation holds up on the smaller screen. Here, the smaller screen is any AI-driven surface that reads your pages differently from Google.

The signal behind the spike

A 27.7% single-day traffic lift does not prove users are abandoning Google. It proves that a meaningful slice of users will move when they feel pushed. That is the audit-relevant insight: the audience that cares about AI opt-outs is also the audience most likely to notice when a page does not respect their choice. Lighter pages, cleaner markup, and consistent entity data serve both groups. Heavy interstitial flows, autogenerated content blocks, and inconsistent directory listings frustrate the opt-out crowd first and then start bleeding into the rest of the traffic mix.

FAQ

How large was DuckDuckGo’s traffic increase between May 20 and May 25?

Visits to DuckDuckGo’s AI-free search page rose 22.7% on average week-over-week during that window, peaking at 27.7% on May 24. US app installs rose 18.1% week-over-week, with iOS installs averaging 33% higher and spiking 69.9% on May 25.

Why are users switching to DuckDuckGo right now?

CEO Gabriel Weinberg attributed the move to Google force-feeding AI with no way to opt out, which he said is making results worse. DuckDuckGo lets users disable AI-generated results while still offering AI tools through Duck.ai, including GPT-5 mini and Claude Haiku 4.5.

What should a site owner change in an AI-era audit?

Validate structured data on key templates, reconcile business information across at least five core directories, and tighten the opening sentences on landing pages so AI systems can extract clean citations. These steps hold up across Google and smaller AI-driven engines.

Figure AI’s Figure 02 humanoid robot completed 200 consecutive hours of autonomous physical work inside a simulated logistics environment, with no remote control and no human intervention. The test, documented in a company endurance report, recorded more than 28,000 pick-and-place cycles, a 99.4% task-completion rate, and over 120 miles of walking. It is the first time a humanoid platform has sustained a full workweek of repetitive physical labor without any operator handoff.

Why endurance changes the embodied AI conversation

For most of the last decade, humanoid demonstrations have been measured in minutes or hours. A robot that could walk across a stage, fold a towel, or sort a handful of objects drew headlines, then went back to the charger. The 200-hour run is significant because it answers the question warehouse operators actually ask: can the machine survive a shift, and the shift after that, and the shift after that?

The International Federation of Robotics has projected that humanoid robots could surpass 1.5 million units deployed worldwide by 2035. That forecast assumes endurance, not novelty. A robot that runs for a full week without a human touching it moves the technology from research project to candidate workforce.

What the test setup actually looked like

Figure’s engineering team placed Figure 02 inside a climate-controlled mock fulfillment center stocked with standardized storage bins, conveyor belts, and pallet racks. The robot ran a closed loop of warehouse tasks: pulling items from bins, placing them into shipping totes, walking between stations, scanning barcodes, and managing its own power supply through autonomous docking and battery swaps.

No operator intervened at any point. The onboard neural networks handled full task planning, error recovery, and energy forecasting. According to Figure’s endurance report, the platform proactively routed itself to a charging dock before its battery state crossed a safety threshold rather than waiting to fail.

The numbers from the 200-hour window

• 200 hours of continuous operation, the rough equivalent of 25 standard eight-hour workdays run back to back.
• More than 28,000 successful pick-and-place cycles, with a total payload moved exceeding 12,000 kilograms.
• 99.4% task-completion rate. The remaining 0.6% triggered automatic retries caused by grip slip or docking misalignment, and every retry resolved on the robot itself.
• Over 120 miles walked across varied floor surfaces inside the test cell, demonstrating locomotive stability under sustained load.
• Zero remote-operator handoffs. All error recovery ran through the onboard planning stack.

Those figures are the meaningful ones for anyone evaluating whether the technology is ready for a paying contract. A 99.4% success rate across tens of thousands of cycles is the kind of reliability number procurement teams request from incumbent automation vendors. Figure is now in that conversation.

What is technically new

Three engineering choices carried the run. First, the electric actuation stack was tuned for power efficiency, so each battery cycle produced more work than previous generations. Second, the packs themselves are hot-swappable: the robot walked into a dock, swapped a depleted pack for a fresh one, and resumed work without an external technician. Third, a self-monitoring layer predicted energy state and dispatched the robot to a charger ahead of depletion, rather than reacting after the fact.

The fourth ingredient was a custom end-effector, the gripper, that held its grip reliability across the full run. Slipping grippers are the most common failure mode in pick-and-place robotics, and the report credits the gripper design with keeping retry rates under one percent.

What Figure is doing next

The hardware that ran the endurance test now moves into live pilot work. BMW has been evaluating Figure 02 for material handling at its Spartanburg manufacturing facility; the next milestone is pushing shift-length endurance onto a real automotive assembly line, where manipulation demands are less uniform than a test cell.

Figure is also building out multi-robot coordination. The roadmap calls for several humanoids sharing a task queue, dynamically reassigning work based on each unit’s battery state and physical location. Safety certification for human co-working environments is a parallel track, since any commercial deployment will require regulatory sign-off before a robot shares a floor with people who are not test engineers.

Beyond the factory, the targets are last-mile delivery depots and large-format retail backrooms, settings where the work is bounded but the product mix shifts constantly.

What to watch if you run an operation

For site owners and operations leads, the 200-hour result is a procurement signal rather than a purchasing signal. The cost curve is still steep, and the current pilots are confined to controlled cells with standardized bins. Before a humanoid makes sense in your facility, three things need to mature:

• Generalization to unstructured inventory. The test used uniform totes. Real warehouses hold irregular shapes, soft packs, and transparent films.
• Manipulation dexterity beyond pick-and-place. Tote packing, label placement, and induction onto conveyors are still hard.
• Safety certification for mixed human-robot zones. Until regulators publish a clear framework, deployment will be limited to fenced cells.

Leasing models for humanoid labor are likely to follow the path warehouse IoT took, with vendors offering per-shift or per-cycle pricing once fleets scale. Operators who map their current manual-handling hot spots now, the SKUs that move most volume, the stations with the highest labor turnover, will be best positioned to evaluate a pilot when one becomes available.

How this fits the wider AI agent trend

The software side of the industry has spent the last two years shipping agentic systems that book appointments, write code, and chain tool calls without supervision. Figure’s endurance result is the physical counterpart: an embodied agent that runs a task loop without a human in the loop. The two tracks converge when a software agent watching inventory levels hands a restock request to a humanoid that walks to the right shelf and replenishes it. None of that is productized yet, but the building blocks on each side are arriving.

FAQ

How did Figure 02 keep running for 200 hours without a human?

Efficient electric actuation, hot-swappable battery packs, and onboard energy-prediction routines let the robot route itself to a charging dock, swap packs, and resume work. A purpose-built gripper design kept slip failures rare, and onboard planning handled the small share of retries that did occur.

What did the robot do during the 200-hour test?

Inside a climate-controlled mock logistics center, Figure 02 picked items from storage bins, placed them into shipping totes, walked between stations, scanned barcodes, and managed its own recharging. The run produced more than 28,000 successful pick-and-place cycles and moved over 12,000 kilograms of product.

Is Figure 02 ready for commercial deployment?

Pilots are already running at BMW’s Spartanburg plant for material handling, and the endurance result clears the shift-length reliability bar. Widespread commercial rollout still needs safety certification, better generalization to unstructured inventory, and lower unit costs.