Microsoft has started routing review-star clicks in its Copilot AI assistant directly to Google Maps business profiles, a move that surprised the search industry when it was first flagged by SEO consultant Nathan Gotch on X and later confirmed by Search Engine Roundtable on August 15, 2025. The behavior means a user searching for “best coffee near me” in Copilot and clicking the star rating or review snippet gets taken straight to the corresponding Google Maps listing—not a Bing or Copilot-hosted review page, and not Bing Places, Microsoft’s own local business directory.

This seemingly small UI choice carries significant strategic weight. It redirects local discovery traffic out of Microsoft’s ecosystem and into Google’s, at a time when AI assistants are reshaping how consumers find businesses. The observation, documented by Barry Schwartz with screenshots and a GIF, shows that Copilot’s local answer cards now link to an external, competing platform for user-generated reviews, raising immediate questions about data sourcing, monetization, and the cross-platform reality of AI-powered search.

What exactly was observed

When a user asked Copilot for a local recommendation—such as “best coffee near me”—the AI returned a card with the business name, a star rating, and a short snippet of review information. The critical detail: clicking that star rating or the reviews element opened the business’s Google Maps profile in a web browser or the Google Maps app. This was not a one-off glitch. The behavior was replicated independently by Search Engine Roundtable and by other SEOs who tested similar queries after the initial report.

The distinction is stark. In a traditional Microsoft-run search experience, clicking a review link would typically lead to a Bing Places page or a proprietary overlay within Copilot. Here, the single click removes the user from the Copilot/Bing environment entirely and drops them into Google Maps, where Google controls the review dataset, the user interface, and every subsequent conversion action—directions, calls, website visits, or bookings.

How this is technically possible

Copilot is not a walled garden. Microsoft has explicitly built its AI assistant to pull from the open web. In March 2025, the company announced open web search for Copilot Studio agents, giving them the ability to “surface high-quality, real-time information from the web when configured knowledge sources fall short.” That capability is now generally available to all users, and it is exactly the architecture that makes linking to Google Maps feasible. Copilot can reference or link to any external source it deems authoritative, whether that’s Bing Places, Yelp, TripAdvisor, or—as has now been observed—Google Maps.

Further, Copilot’s rollout of agentic web actions and deeper browsing features demonstrates Microsoft’s intent to let the assistant interact with external services. The same plumbing that allows Copilot to book a restaurant reservation or check a flight status can just as easily construct a link to a Google Maps listing. The routing observed in local cards does not require any special partnership or API integration with Google; it is simply an emergent behavior of Copilot’s web-sourcing orchestration, which synthesizes what it finds on the public internet.

Why would Microsoft point its own users to a competitor?

The move seems counterintuitive, but several plausible explanations exist—none mutually exclusive.

Data completeness and freshness. Google Maps houses the world’s largest and most frequently updated corpus of business reviews. For many listings, especially outside the U.S. or for less digitally engaged businesses, Google’s review set is far richer than Bing Places’. Copilot may be defaulting to the most comprehensive source available, improving answer quality and perceived accuracy, even if that source belongs to a rival.

User intent and experience. When a user clicks a review, they often want to see the full timeline of customer comments, user photos, or the ability to get directions. Google Maps excels at these tasks. Copilot’s summary card might not satisfy the user’s real need, so linking out to the canonical host of that rich content can be the most helpful—and pragmatic—design choice.

Engineering trade-offs. Displaying full reviews inline would require crawling, caching, and re-serving content verbatim, which raises copyright and freshness challenges. By linking to the original source, Copilot attributes the content correctly and offloads the heavy lifting of maintaining a review database, while still providing the user with a path to deeper information.

Emergent sourcing, not deliberate favoritism. It’s possible that Copilot’s algorithm simply considers Google Maps the best available source for detailed reviews on a given business, based on web signals like traffic, recency, and authority. No explicit decision to “favor Google over Bing” may have been made; rather, the open web search capability may naturally surface Google as the top result for review content.

Immediate impact on businesses and local SEO

The most immediate effect is a shift in traffic routing. A potential customer who clicks a review in Copilot now lands on Google Maps, not on the business’s own website or a Bing property. This changes the conversion and attribution path: a sale that originated with an AI assistant may end up being credited to Google Maps, masking the true source of the lead.

Visibility dependence deepens. Businesses that have neglected their Google Business Profile (GBP) while focusing exclusively on Bing Places or other directories may suddenly find that even Copilot users are funneled to Google. Maintaining a strong GBP—complete with accurate hours, photos, and responses to reviews—becomes essential for any business hoping to be discovered through AI.

Cross-platform optimization is no longer optional. The old mantra of “claim your listing everywhere” takes on new urgency. Copilot could reference Yelp tomorrow, or TripAdvisor, or Apple Maps. Local SEO practitioners must manage and monitor their presence across multiple platforms, not just the one associated with a particular assistant. Inconsistencies in name, address, phone number, or reviews can lead to confusion when an AI cites conflicting sources.

Analytics must evolve. Traditional rank-tracking and organic referral reports won’t capture the full funnel from AI assistants. Teams need to instrument conversions differently: watch for referrer strings from copilot.microsoft.com, analyze downstream Google Maps interactions (calls, direction requests), and correlate them with sales. The old model of measuring clicks on a blue link is insufficient when an AI answer card can both serve a summary and then hand the user off to a third-party platform.

Competitive and regulatory angles

Routing users to a direct competitor’s property is highly unusual for a major search product. It raises questions about search neutrality and platform steering. For years, Microsoft has argued that Google’s practices in search and advertising are anticompetitive. Sending Copilot traffic to Google Maps could complicate that narrative, though it might also be viewed as a pragmatic, user-first decision that demonstrates Microsoft’s willingness to put the best answer ahead of corporate allegiance.

Monetization consequences. Google Maps is a lucrative advertising and conversion surface for Google. If AI assistants like Copilot start regularly funneling users there, Microsoft loses the chance to monetize that local intent on its own platforms. Bing’s local ads, Bing Places calls-to-action, and any planned Copilot-native booking features are bypassed entirely. This could undercut Microsoft’s ability to build a sustainable local revenue model for Copilot, making the assistant more dependent on other monetization paths like the Microsoft Store or enterprise subscriptions.

Privacy and data handling. When a user clicks from Copilot to Google Maps, they leave Microsoft’s authenticated session and enter Google’s domain, which may have different cookie policies, tracking mechanisms, and data-collection practices. This abrupt transition could create transparency and consent issues, especially if the user is not clearly warned that they are being handed off to a third-party service.

Strengths of Microsoft’s approach

Despite the apparent contradictions, the strategy is not without merit.

  • Better answers now. If Google Maps has 1,000 reviews for a restaurant while Bing Places has 12, linking to Google provides a more complete, helpful experience. Prioritizing usefulness over platform ego can build trust in Copilot as an honest assistant.
  • Provenance and transparency. Sending users to the original source of review content is intellectually honest. It avoids the black-box problem where an AI summary is the only thing visible, and it lets users verify claims themselves.
  • Flexible architecture. Copilot’s ability to incorporate best-of-breed sources aligns with Microsoft’s broader vision of an assistant that integrates information from anywhere, not just Microsoft properties. Open web search and agentic capabilities are core to that design.

Risks and downsides

  • User confusion. People expect Copilot to be a self-contained experience. Being abruptly dropped into Google Maps can feel disjointed, especially if the user was not prompted to leave the assistant.
  • Lost value capture. Every user who follows a review link to Google Maps is a potential customer whose subsequent actions—ad clicks, calls, bookings—occur under Google’s monetization umbrella. Microsoft cedes control of that downstream journey.
  • Trust and perception. Some may interpret this as Microsoft admitting that its own local index is inferior. That could hurt Copilot’s brand as a serious search alternative, even if the decision is purely algorithmic.
  • Uncertain staying power. This behavior might be an A/B test, a regional experiment, or a temporary glitch. The Search Engine Roundtable report notes it was reproducible at the time, but Microsoft has not commented publicly. Any strategic reliance on this routing could be upended by a future code change.

What businesses and SEOs should do now

  1. Audit cross-platform listings immediately. Make sure your Google Business Profile is claimed, verified, and fully populated with accurate information, high-quality photos, and an active review response strategy. Do the same for Bing Places, Apple Maps, and Yelp. Inconsistencies across platforms will only become more problematic as AI assistants cross-reference data.
  2. Rethink conversion tracking. Set up custom UTM parameters or link-shortening services for links placed in directories and social media that might be scraped by AI. Monitor for “copilot” in referrer strings and map those sessions through to on-site conversions or Google Maps interactions. If Google Maps now receives AI-driven traffic, ensure your GBP has clear call-to-action buttons and that you monitor Insights for directional, call, and website clicks.
  3. Optimize for mixed-navigation flows. Design user journeys that assume a customer may discover you in Copilot, read reviews on Google Maps, and then visit your website separately. Ensure that your Google Maps messaging aligns with your website, and that your site is mobile-friendly, since many Google Maps clicks come from smartphones.
  4. Stay informed. Microsoft may adjust this behavior at any time. Watch official Copilot and Bing blogs, follow SEO news outlets like Search Engine Roundtable, and test Copilot queries regularly to see where local review clicks land.

What journalists and researchers should test next

  • Scope and geography. Is this global, or limited to certain markets? Reproduce the behavior from different IP addresses and language settings.
  • Source alternation. Does Copilot ever link to Yelp, TripAdvisor, or Bing Places for reviews? Under what conditions does it choose one over another?
  • Referral path. Examine the actual URL that Copilot uses to link to Google Maps. Is it a direct maps.google.com link, or does it pass through a Microsoft intermediate like go.microsoft.com? Check for referral headers and URL parameters that might indicate attribution.
  • Account type and tier. Does the behavior differ for free Copilot versus Copilot Pro or enterprise accounts? Licensing may influence source selection.

Verification and caveats

The core observation is well-documented. Search Engine Roundtable’s report includes screenshots and a video clip showing the routing in action, and Nathan Gotch’s original X post sparked the discussion. Microsoft’s own documentation validates that Copilot can use open web search, making the linking technically legitimate. However, without an official statement from Microsoft, we cannot determine whether this is a permanent product decision, an emergent outcome of web-sourcing algorithms, or an experimental feature that may be rolled back. Readers should test the behavior in their own Copilot environments and treat it as potentially transient.

The Copilot–Google Maps routing crystallizes a defining tension of the AI era: who owns the user’s intent and the conversion pathway when assistants synthesize answers from multiple sources? In a traditional search results page, clicks are distributed by ranking positions, and each destination is a separate website. With AI-generated answer cards, the assistant may summarize content without a click, but when a click is necessary, the assistant must choose a destination. That single choice can silently redistribute enormous value across the ecosystem.

For users, the priority is the most useful answer; sometimes that means linking to a platform with richer data, even if it belongs to a competitor. For platform owners, capturing the full funnel is critical for monetization and ecosystem lock-in. For regulators, these hybrid flows introduce new questions about disclosure, fairness, and competitive dynamics. Should an assistant be required to tell users why it picked one source over another? Should Microsoft be obligated to treat its own properties on equal footing? These are not theoretical questions—they are now live in production.

Final analysis

Microsoft’s quiet routing of review clicks to Google Maps is a small feature change with outsized consequences. It underscores that AI assistants do not operate in silos; they are multiplexers that can favor any web source based on quality, completeness, and relevance—even if that source is a direct competitor. The immediate lesson for businesses is clear: visibility on Google is now essential for AI-driven discovery, regardless of which assistant the customer uses. For SEO professionals, the era of single-ecosystem optimization is over. Cross-platform readiness, rigorous analytics that span from answer card to in-store visit, and a willingness to adapt quickly are the new table stakes.

Whether this routing becomes a permanent design choice or a fleeting experiment, one truth holds: the wall between search engines is crumbling. AI assistants will continue to stitch together the web’s best parts, and the winners will be the platforms that host the most trusted, complete, and actionable local signals. For businesses, that means maintenance across Google, Bing, and every other relevant listing is no longer optional—it is the price of being found in an AI-intermediated world.