Google Back Button Hijacking Ban: SEO Impact and Agentic Search

Google drew two clear lines this week: one around a technical manipulation tactic that has been quietly degrading user experience for years, and one around who controls the transaction layer when a user searches for a restaurant, service, or experience. Back button hijacking is now a codified spam violation with a hard enforcement date of June 15, 2026, and Google’s AI-powered agentic booking just expanded to the UK and India — completing restaurant reservations entirely within Google’s interface, with no website visit required. If you manage sites with aggressive pop-up tools or ad networks, or you run a business that depends on booking transactions, both of these announcements require immediate action.

What Happened

On April 17, 2026, Search Engine Journal reported a cluster of three significant Google updates published in rapid succession, each carrying concrete enforcement implications for SEO practitioners and digital marketers.

Back Button Hijacking Becomes a Classified Spam Violation

Google formally added back button hijacking to its spam policies, classifying it under the malicious practices category. According to Google’s official spam policies documentation, back button hijacking is defined as “when a site interferes with user browser navigation by manipulating the browser history or other functionalities, preventing them from using their back button to immediately get back to the page they came from.”

The enforcement start date is June 15, 2026. Sites confirmed to be in violation after that date face two enforcement tracks: automated ranking demotions applied by Google’s algorithmic spam systems, or manual spam actions issued by Google’s human webspam reviewers. The Google Spam Policies document was last updated April 13, 2026, confirming this is a freshly codified rule — not a retroactive enforcement of something that was ambiguously prohibited in earlier policy language.

One of the most consequential details in the announcement: Google has explicitly stated that publishers are liable even when back button hijacking originates from included third-party libraries, advertising platforms, or other external code running on their pages. If an ad vendor’s JavaScript is manipulating the browser’s history API, that is the publisher’s problem to fix — not the vendor’s. The site owner bears full responsibility regardless of the code’s origin.

Sites that receive a manual action after June 15 can submit reconsideration requests through Search Console after removing all offending code. According to Google’s Search Console manual actions documentation, most reconsideration reviews take several days to several weeks, though link-related violations have historically run longer. A manual action that arrives on June 16 could suppress organic traffic for a month or more before a review is completed and a penalty lifted.

Spam Reports Now Carry Direct Enforcement Weight

A second update, announced April 14, 2026, materially changed how Google processes user-submitted spam reports. Previously, spam submissions from users fed into Google’s automated detection training systems and nothing more — they did not directly result in manual actions or human reviews of the reported site. That process has changed.

Under the updated system, user spam submissions may now trigger direct manual action reviews against reported sites. In parallel, the text of each spam report is forwarded verbatim to the reported website through Search Console’s message center. Site owners will now see exactly what their visitors are reporting about them, and those reports now carry a direct escalation path to enforcement consequences.

As Google’s manual actions documentation clarifies, manual actions are issued when human reviewers determine that pages or sites violate spam policies. The addition of user-report escalation adds a new trigger for those human review workflows alongside Google’s existing algorithmic flagging, and the verbatim forwarding means publishers have a window to see complaints before consequences land.

Agentic Restaurant Booking Expands to UK and India

On April 10, 2026, Robby Stein, VP of Product for Google Search, announced the expansion of Google’s agentic restaurant booking feature to the United Kingdom and India. According to Search Engine Journal’s reporting, the feature had previously been available in limited US markets and now enters two of the world’s largest and most commercially active English-language search markets simultaneously.

The user experience works like this: a person opens Google’s AI Mode, describes what they’re looking for — group size, preferred time slot, cuisine type, neighborhood — and the AI system scans booking platforms for real-time availability matching those parameters. When a suitable option is found, the booking is completed through Google’s network of partner platforms. The restaurant’s own website is never visited. The restaurant’s analytics platform never logs a session. The reservation is made entirely within Google’s interface, and the transaction lands on the restaurant’s books through the partner booking system.

Why This Matters

These two announcements hit different operational areas of marketing, but they share a common underlying logic: Google is systematically eliminating user experience manipulation while simultaneously absorbing more of the conversion pathway into its own platform. Understanding the full marketing implications requires examining each update separately before considering their combined effect.

The Vendor Liability Clause Changes Your Audit Obligation

For in-house marketing teams and agencies managing client sites, the vendor liability element of the back button hijacking policy is the sentence that demands the most immediate action. Most back button hijacking on publisher sites isn’t intentional architecture — it’s a byproduct of third-party tools deployed for legitimate marketing purposes that happen to use navigation API calls in ways Google now classifies as malicious.

Common sources of inadvertent back button hijacking include pop-up and lead capture tools that inject browser history states to intercept exit intent, ad networks running programmatic JavaScript that adds intermediate redirect states to the history stack, exit-intent overlays that intercept popstate events to re-engage departing users, affiliate tracking libraries that insert redirect hops into the browser history, and some enterprise live chat or chatbot widgets that capture back-button events as a “keep user on page” re-engagement mechanism.

If your site runs any combination of these — and most monetized content-driven sites do — you are potentially hosting code that will be classified as a malicious practice starting June 15. The responsibility for detection and remediation sits entirely with the publisher, regardless of which vendor introduced the behavior.

For agencies managing dozens of client sites, this creates a specific operational challenge. Client sites routinely run Google Tag Manager containers with third-party scripts installed by previous contractors, legacy vendors, or client-side marketing teams operating outside the agency’s direct purview. An ad ops decision made two or three years ago can leave manipulative JavaScript running that nobody has reviewed since installation. The back button hijacking policy means every managed site now requires an explicit, documented audit of its complete JavaScript inventory against navigation API behavior — something that simply was not a required deliverable before this announcement.

For solopreneurs and small publishers, the risk is more concentrated. If you’re monetizing with display ads through a network, or using a pop-up builder that hasn’t been updated in 18 months, or running an affiliate tracking setup with redirect chains baked into your outbound links, you need to test back-button behavior explicitly on your highest-traffic pages before June 15.

Spam Reports With Real Enforcement Weight Change Competitive Dynamics

The escalation of user spam reports to actionable enforcement triggers deserves more analysis than it has received in the initial wave of coverage. Previously, submitting a spam report against a competitor was essentially a data donation to Google’s training pipeline — it might marginally improve automated detection over time, but it had no direct bearing on whether the reported site faced any consequences. That asymmetry has changed.

Any motivated individual — a dissatisfied customer, a former employee, or a direct competitor — can now use the spam report form as a direct pathway into Google’s manual review process. While Google still requires a human reviewer to validate the complaint and confirm an actual policy violation before issuing a manual action, the mere existence of the escalation pathway creates operational risk that did not exist before April 14, 2026.

The verbatim forwarding of report text through Search Console is a double-edged feature. Publishers running clean operations will receive advance warning of any complaint and can review their practices before an action is formally issued. Publishers running practices that are close to a policy boundary will see specific complaint language that clarifies exactly what a reviewer will be evaluating. Either way, the monitoring obligation this creates — checking Search Console message center on a daily basis, having a defined internal response process for incoming reports — is now a standard operational requirement for any site that depends on organic search traffic.

Agentic Search Means the Conversion Happens Without a Website Visit

The restaurant booking expansion is a category-level disruption for local business marketing. Zero-click search — where users get answers in the SERP without visiting a publisher’s website — has been a growing concern since featured snippets became mainstream. Agentic search goes a step further: it is zero-visit conversion. The user expresses intent, the AI acts on that intent, and the outcome (a confirmed reservation) is delivered — all without the business’s website playing any role whatsoever in the transaction.

For restaurants in the UK and India, a confirmed table booking now generates no session in Google Analytics, no pageview on the restaurant’s reservations page, no conversion event tracked by the CRM, and no remarketing pixel firing to support future re-engagement. The customer exists only as a record in the partner booking platform’s database. If the restaurant has configured downstream email flows through its reservation management system, they may capture the contact. If they haven’t, the customer is effectively invisible to the business’s marketing stack until they walk through the door.

The expansion to UK and India — two major markets with enormous restaurant density and high mobile search usage — signals that this is no longer pilot territory. Google does not expand features to billion-person markets on an exploratory trial basis. The announcement by Robby Stein, VP of Product for Google Search, frames this as a standard product rollout.

The Data

Back Button Hijacking Policy: Timeline and Enforcement Framework

Sources: Search Engine Journal, Google Spam Policies, Google Search Console Manual Actions Documentation

Agentic Restaurant Booking: Expansion Timeline

Source: Search Engine Journal

Common Sources of Back Button Hijacking by Tool Category

Note: Risk levels reflect typical implementations; test your specific tool configuration before June 15.

Real-World Use Cases

Use Case 1: Digital Agency Auditing a Multi-Site Client Portfolio Before June 15

Scenario: A 12-person digital marketing agency manages 35 client websites across e-commerce, B2B SaaS, and local services verticals. The sites have been onboarded over five years with varying tag management setups, pop-up tools, and ad monetization configurations. The agency does not have a current inventory of which scripts are running on which sites, and several older clients have legacy GTM containers inherited from previous contractors — installed and never revisited.

Implementation: The agency prioritizes sites by organic traffic volume and builds a browser automation workflow using Playwright that loads target pages, simulates forward navigation, and then programmatically triggers a back-button press, logging any pages where the back button does not immediately navigate away from the site. Flagged pages are manually reviewed with Chrome DevTools open in the Sources panel, monitoring for pushState and replaceState calls during page load. Every tag in each GTM container is inventoried against a checklist of known navigation-manipulating vendors. Vendors flagged as potentially non-compliant receive a formal inquiry requesting written confirmation that their current library version does not manipulate the browser history API. Any site where vendor compliance cannot be confirmed before June 10 has those specific tags paused pending a compliant update or vendor replacement.

Expected Outcome: All 35 client sites are cleared of confirmed back button hijacking violations before the June 15 enforcement window opens. No client receives a manual action. The agency documents the audit as a repeatable process and adds “verify no navigation-manipulating JavaScript” to the standard pre-launch technical checklist for all future site deployments and quarterly maintenance reviews.

Use Case 2: Display-Ad Publisher Removing an Offending Ad Network

Scenario: A mid-sized food and travel publisher generates approximately 70% of its revenue from display advertising through a premium programmatic ad network. Its traffic is primarily organic — around 400,000 monthly sessions — making a manual action penalty an existential threat to the business. During a routine DevTools review, the publisher discovers that the network’s header bidding JavaScript library injects intermediate history states during ad slot initialization, causing the back button to cycle through ad-loading states rather than escaping the page immediately.

Implementation: The publisher contacts the ad network account manager and requests explicit written documentation confirming whether the current production version of their header bidding library manipulates the browser history API. While waiting for the vendor’s response, the publisher implements a Content Security Policy in report-only mode, specifically monitoring for unexpected calls to history.pushState and history.replaceState. The resulting CSP report log identifies the exact script and triggering conditions. Armed with this documentation, the publisher gives the vendor a deadline of June 1 to provide a compliant library version. In parallel, the publisher identifies two alternative ad networks with confirmed-clean header bidding implementations and begins onboarding. If no compliant version is available before June 10, the offending network’s tags are disabled on the five highest-traffic pages while the migration completes.

Expected Outcome: The publisher avoids a manual action that could eliminate the organic traffic underpinning 70% of its revenue. A temporary revenue reduction of 10–20% during the network transition is accepted as the cost of managing the risk — far preferable to the alternative of a ranking penalty that suppresses organic traffic for weeks while a reconsideration request is processed.

Use Case 3: Multi-Location Restaurant Group Optimizing for Agentic Booking in the UK

Scenario: A restaurant group with 18 locations across London, Manchester, and Birmingham has historically driven reservations through its own website, its Resy profile, and Google Business Profile walk-in traffic. With agentic booking now live in the UK as of April 10, 2026, the digital manager recognizes that reservation traffic attribution is about to shift in ways the current analytics setup cannot capture or measure.

Implementation: The team audits every third-party booking platform listing for all 18 locations — OpenTable, Resy, TheFork, and Bookatable — updating each with accurate cuisine classifications, seating capacity by area (main dining room, private dining, bar seating), current menu content, dietary accommodation flags (plant-based options, gluten-free, major allergen policies), high-quality interior and food photography, and precise hours including holiday variations. Each Google Business Profile is updated with the reservation booking link pointing to the most complete platform profile. Real-time availability sync is enabled between the reservation management system and all connected platforms, ensuring that Google’s AI scans reflect live table availability rather than stale slot data. A post-confirmation email automation is set up through the reservation management system: any confirmed booking — regardless of source — triggers a welcome email with a loyalty program signup link and a pre-visit preferences form. UTM parameters are applied at the confirmation URL level on all partner platforms to attribute source in the CRM.

Expected Outcome: The group’s reservation volume is maintained or increases as the agentic booking channel adds a new conversion path that doesn’t require a visit to the restaurant’s website. Google-sourced agentic bookings are distinguishable in the CRM through UTM attribution. First-party data capture is partially preserved through the post-confirmation email flow. The team shifts its success reporting framework from website sessions to covers booked.

Use Case 4: SaaS Brand Proactively Managing Spam Report Escalation Risk

Scenario: A B2B CRM SaaS company runs an aggressive content marketing strategy that includes detailed competitor comparison pages and “best of” roundups where the company’s own product consistently ranks first. The pages drive strong organic traffic and lead volume. With user spam reports now a direct escalation path to manual reviews, the content team reassesses its exposure — particularly on pages that make performance comparisons without disclosing that the publisher is also the top-ranked product.

Implementation: The SEO lead is assigned as the dedicated owner of Search Console monitoring, with a 24-hour response commitment for all incoming message center notifications. An internal content audit is conducted against Google’s spam policies, focusing on thin content, automatically generated pages, and comparison content that could be perceived as deceptive. The 12 most aggressive comparison pages are revised to include methodology explanations, disclosure that the company publishes the rankings, and links to independently sourced third-party reviews. Three thin-content pages identified in the audit — auto-generated location landing pages with minimal unique content — are consolidated into a single hub page with substantive regional content. A response workflow is documented: when a spam report notification arrives in Search Console, the legal and content teams are notified simultaneously, and a written response with documented evidence of compliance or remediation is prepared within 48 hours for submission if a manual action follows.

Expected Outcome: The company significantly reduces its policy violation surface area before any complaint arrives. If a report does escalate to a manual action, the defined response process enables remediation and resubmission within 48 hours, minimizing the ranking impact window. The content audit also surfaces quality issues that, once fixed, improve conversion performance on the comparison pages independent of the spam risk.

Use Case 5: Independent Restaurant Getting Listed for Google Agentic Booking

Scenario: An independent gastropub in East London has a basic Google Business Profile and an OpenTable listing set up two years ago with minimal information — no photos, generic cuisine tags, and hours that were never updated after a schedule change last autumn. With agentic booking now active in the UK, the owner recognizes that an incomplete profile means the restaurant is likely invisible in Google’s AI Mode availability scans, which need complete structured data to match venues against specific user queries.

Implementation: The owner completes a full audit of every digital booking listing. OpenTable is updated with current hours (including updated Sunday lunch service and bank holiday closures), a comprehensive food and drink menu with pricing ranges, dietary options (plant-based section, gluten-free options, allergen policies), private dining availability for groups over 12, and a set of 25 professional photographs covering food dishes, the main dining area, the bar, and the private dining space. The Google Business Profile is brought fully up to date with accurate business categories (gastropub, British cuisine), the OpenTable booking link, and responses to all outstanding reviews. Real-time availability sync is enabled between the venue management system and OpenTable to ensure agentic search results reflect accurate live table availability. The owner registers for Google Business Messages so that AI-Mode queries asking follow-up questions about the venue have a direct response channel.

Expected Outcome: The gastropub’s OpenTable listing contains the complete, structured information that Google’s AI scanning requires to match it against user queries for group size, time, and cuisine type in London. Reservation volume increases through the agentic booking channel. The owner accepts that website session counts will not track these conversions and shifts success measurement to covers booked per week, monitoring source attribution in OpenTable’s reporting dashboard.

The Bigger Picture

These announcements don’t exist in isolation. They’re part of a clear and accelerating pattern in how Google is redefining the relationship between organic search and downstream business outcomes — one that systematically disadvantages passive or manipulative digital marketing strategies while expanding Google’s direct role in commercial transactions.

Google’s Expanding Definition of Spam Now Targets UX Manipulation

Back button hijacking is not the first time Google has added a user experience manipulation tactic to its formal spam policy framework. The site reputation abuse policy introduced in 2024 targeted parasite SEO — publishing low-quality third-party content on high-authority domains specifically to game rankings. Cloaking has been a spam violation for years. Hidden text, doorway pages, and sneaky redirects are all codified in the same spam policies document. The pattern across all of these: a tactic that harms users in the short term while temporarily improving a metric the publisher cares about gets elevated from gray area to explicit violation once Google determines it is degrading search result quality at scale.

Back button hijacking follows that exact logic. It harms users — trapping them on a page against their explicit navigation intent — while appearing to improve publisher engagement metrics like time-on-site, scroll depth, and bounce rate. Google’s consistent position has been that gaming user engagement signals at the expense of actual user experience degrades the quality of organic results at scale, and it penalizes accordingly once a tactic becomes widespread enough to register as a systemic problem.

The vendor liability clause is genuinely new policy territory. By making publishers fully responsible for the behavior of all third-party code running on their pages regardless of the code’s origin, Google is effectively requiring brands to maintain a complete, audited, and current inventory of every JavaScript dependency in their stack. For large enterprise properties running hundreds of injected tags across CDNs, legacy GTM containers, and direct script includes, this is a non-trivial operational requirement that most marketing teams are not currently set up to fulfill on a continuous basis.

Agentic Search Is the Logical Completion of Zero-Click

Featured snippets answered questions without a click to the source. Knowledge panels provided business information without a site visit. Local pack results gave directions, hours, and phone numbers without navigating to the business’s website. Each of these developments moved some volume of user intent resolution from the publisher’s domain into Google’s SERP. Agentic search completes the trajectory: the intent is expressed, the AI matches it against available supply, and the transaction is executed — all inside Google’s platform, with partner integrations handling the fulfillment.

For marketers who have built measurement frameworks anchored to sessions, pageviews, and on-site conversion events, this represents a model break that becomes more significant as the category of “transactions Google can complete agentically” expands. Restaurant reservations are the first highly visible and concrete example because they’re easy to demonstrate and quick to verify. The same infrastructure — describe your need, AI scans booking partners, AI completes transaction — maps directly onto hotel bookings, service appointments, event tickets, and structured retail purchasing. The expansion to UK and India, announced by Robby Stein, VP of Product for Google Search, is product rollout, not experimentation.

The Combined Signal: What Survives Long-Term

Taken together, the back button hijacking ban and the agentic booking expansion tell a consistent story about what Google is building toward. Google is eliminating patterns that manipulate users, and simultaneously building infrastructure to route more user intent through its own transaction layer. The band of tactics that remain viable long-term is narrowing: clean technical implementation, accurate and rich structured data on the booking and commerce platforms Google’s AI queries, first-party data collection that works downstream of Google-executed transactions, and content that earns genuine organic visibility by actually serving user intent.

The brands and agencies that adapt to this model proactively — auditing their technical stack, investing in platform profile completeness, and building first-party data pipelines that work even when Google owns the transaction — will compound that advantage as agentic capabilities expand into new categories over the next 12 to 18 months.

What Smart Marketers Should Do Now

1. Run a Back Button Hijacking Audit Before June 1, 2026

Don’t wait until the June 15 enforcement date to discover a violation — finding one on June 16 means your site is already facing a penalty while you’re simultaneously trying to fix code and submit a reconsideration request that will take weeks to process. Open Chrome DevTools on your highest-traffic pages, open the Application panel, and watch the History entries as each page loads fully. Look for unexpected pushState calls that add intermediate entries to the browser’s history stack before you’ve interacted with anything. Prioritize pages by organic traffic volume: your highest-traffic pages carry the greatest ranking risk. For agencies managing multiple properties, build a Playwright or Puppeteer automation that simulates back-button navigation across pages in bulk, flagging any page that doesn’t exit immediately, then conduct manual investigation on flagged pages.

2. Audit Every Third-Party JavaScript Tag for Navigation API Usage

Never assume your vendors are compliant — request written confirmation. Create a full inventory of every JavaScript tag running across your properties: GTM containers, hardcoded script tags in page templates, tag-management-bypassing direct implementations, and any code injected by server-side rendering. For each identified tag, check the vendor’s changelog and documentation for any references to history API usage, pushState or replaceState calls, popstate event listeners, or features explicitly described as “back button prevention” or “exit intent interception.” Vendors who cannot confirm compliance in writing should be given a deadline of June 1. Any vendor unable to provide a compliant version by that date should have their tags paused on your highest-traffic organic pages until a solution is available. Document every step of this audit — if a manual action does arrive, a paper trail demonstrating due diligence materially strengthens your reconsideration submission.

3. Configure Search Console Alerts and Assign an Ownership Role

With user spam reports now a direct escalation path to manual action reviews, Search Console monitoring has moved from optional hygiene to an operational requirement with a named owner. Configure email notifications for the Manual Actions report and Security Issues report to route to a specific team member — not a rotating responsibility — with a defined 24-hour response commitment. Because report text is forwarded verbatim through the message center, that owner needs enough context to interpret a complaint and determine whether it points to an actual policy issue requiring immediate remediation. For agencies managing multiple properties across multiple Search Console accounts, consolidate monitoring with a shared notification inbox or implement API-based alerting to a Slack channel or ticketing system.

4. Maximize Your Presence on Every Booking Platform Google Queries

If your business completes transactions through reservations, appointments, or structured bookings, identify which platforms Google’s AI Mode queries in your specific category and treat those platform profiles with the same investment you’d apply to your own website — because in the agentic search model, those profiles effectively are your website. For restaurants in the UK and India, that means OpenTable, Resy, TheFork, and Bookatable at minimum. Complete every available profile field: photos, hours including holiday variations, cuisine classifications, dietary accommodation flags, group booking capabilities, and private dining details. Enable real-time availability sync between your reservation management system and all connected platforms so that Google’s AI scans reflect actual live availability. An incomplete or stale listing is an invisible listing in the agentic search context — there are no secondary signals to compensate for missing structured data.

5. Build a First-Party Data Pipeline That Survives Zero-Visit Conversions

The most important strategic adaptation to agentic search is building data collection workflows that operate downstream of the transaction rather than depending on a site visit to trigger them. For restaurants, this means post-confirmation email sequences triggered through the reservation management system that capture loyalty program signups, dietary preferences, and occasion data — independent of whether the customer ever visited the restaurant’s website. For service businesses, it means SMS confirmations with preference collection links. For e-commerce brands preparing for agentic purchasing as the model expands, it means loyalty program incentive structures that trigger post-purchase regardless of which channel completed the sale. The brands that systematically capture identity and preference data from customers whose entire discovery-to-transaction journey occurred inside Google’s platform will have a compounding first-party data advantage over the next 12 to 18 months as agentic capabilities expand into new transaction categories.

What to Watch Next

Agentic Search Expansion Into New Transaction Categories

The restaurant booking feature is proof of infrastructure, not a ceiling on the model’s ambition. Watch for Google to extend agentic transaction capabilities into hotels — a natural extension built on top of the existing Google Hotel Search booking integrations already in place — service appointments (building on Google’s reservation API that already powers healthcare and beauty booking in the US), event tickets, and structured retail purchasing. Google I/O, typically held in May, is the most likely venue for official previews of expanded agentic capabilities for 2026. If expansion announcements come at I/O, the Q3 2026 window is the realistic deployment timeline based on the pace of the restaurant booking rollout.

Back Button Hijacking Enforcement in Edge Cases

The current policy statement covers web search results. Whether enforcement extends to Google Discover content, app indexing, or progressive web applications with complex client-side routing is not yet clarified in the policy documentation. As the June 15 date passes and the first manual actions are issued under this policy, the SEO community’s documentation of those cases will clarify how Google’s webspam reviewers are interpreting “interferes with user browser navigation” in technically complex implementations — single-page applications with custom scroll restoration, multi-step checkout flows that use the history API for step tracking, and embedded webviews are all edge cases that remain ambiguous until real enforcement decisions are documented and analyzed.

Spam Report Escalation Thresholds

The open question about the new spam report process is the threshold at which a report escalates to a formal manual review. If a single well-written report reliably triggers a human review, the competitive SEO dynamic shifts significantly in high-stakes niches where competitors have strong financial incentive to file claims. Over the next six months, documented case studies of report-triggered manual actions will clarify both the escalation threshold and the realistic timeline from submitted report to issued action. Monitor Search Engine Roundtable and the Google Search Central Community forums for the first wave of practitioner case documentation.

Google’s Booking Partner Network Composition

The identity of the “Google partners” through whom agentic booking transactions are completed carries significant commercial implications for both booking platforms and the businesses that rely on them. In the US, OpenTable and Resy have been primary integration points. As the feature expands to UK and India markets, watch which booking platforms receive Google partnership status versus which are excluded. Exclusion from Google’s agentic booking network creates a structural commercial disadvantage for those platforms — and for the restaurants, hotels, and service businesses that have built their reservation infrastructure exclusively on excluded platforms.

Bottom Line

Google’s back button hijacking ban and the agentic restaurant booking expansion are two enforcement actions that share a single underlying message: Google is done tolerating manipulation in the search experience it controls, and it is actively expanding the scope of what it controls to include the transaction layer itself. June 15, 2026 is a fixed enforcement date — run your navigation API audit now, document your third-party JavaScript inventory, and treat this as a hard deadline rather than a guideline, because reconsideration requests after that date take weeks to resolve. For any business whose revenue model depends on booking-based conversions, the UK and India expansion of agentic search is not a future scenario to plan for — it is an operational reality today. The marketers who adapt measurement frameworks, invest in platform profile completeness, and build first-party data pipelines that survive zero-visit conversions will compound that advantage significantly as Google expands agentic capabilities into hotels, appointments, events, and retail over the next 12 months.