Bing's quiet role in AI search

Reference reading: For platform-specific Microsoft Merchant Center setup, the forthcoming Microsoft Merchant Center guide will cover the implementation. For the broader AI search landscape, see Traditional SEO vs. AI search optimization.

Most ecommerce SEO and AI-search optimization budgets allocate to search engines roughly in proportion to organic search share. Google gets the majority of attention and budget; Bing gets a small slice; everything else gets a token mention. That allocation usually underweights Microsoft’s footprint in AI search specifically.

Three reasons, all structural.

Microsoft Copilot is real

Copilot is the AI assistant integrated across Windows, Microsoft 365, Edge, and Bing. When a Copilot user asks a shopping question, the answer pulls from Bing Shopping, Microsoft Merchant Center, and the broader Bing index — the same backend across each surface.

For an ecommerce catalog, surfacing in Copilot means surfacing across multiple Microsoft surfaces simultaneously. The exact user counts shift over time and Microsoft does not publish them consistently — for current counts, refer to Microsoft’s quarterly investor reports — but the deployment surface is broad enough (Windows default, Microsoft 365 ubiquity, Edge install base) that treating Copilot as a marginal AI consumer underestimates the backend it pulls from.

Bing’s index is consumed by more than Bing

The part most operators miss. Bing’s index is licensed and consumed by AI services beyond Microsoft’s own products.

OpenAI’s ChatGPT has historically relied on Bing as a primary search backend for web answers. The arrangement has been reported across tech press since the launch of Bing Chat (later renamed Copilot) in 2023; OpenAI continues to evolve its own retrieval layer, but the underlying point — that a meaningful share of ChatGPT’s web answers have routed through Bing’s index — is well-established.

DuckDuckGo and Yahoo also use Bing’s index as a primary organic results layer. This is documented in DuckDuckGo’s “Sources” help article, which lists Bing among the sources DuckDuckGo’s results draw from. AI features built on top of those engines inherit Bing’s data.

The practical consequence: a catalog optimized for Bing surfaces across a wider set of AI experiences than the Bing-only surfacing share would suggest.

Microsoft’s structured-data discipline is more permissive than Google’s

Bing has historically accepted a broader set of structured-data types than Google narrows to over time. Google removed FAQ rich results from non-authoritative sites in August 2023 and removed HowTo rich results from desktop on September 13, 2023; those policy changes were Google-specific.

Microsoft’s October 2025 ads guidance for AI search explicitly recommends structured data as part of preparing content for inclusion in AI-generated answers. The framing is concrete — Microsoft is publishing operator guidance that says it uses structured data for AI surfacing.

For ecommerce catalogs already producing Schema.org markup for Google, the markup is reusable on the Microsoft side without additional work.

What’s commonly underweighted

Three Bing-specific optimizations operators commonly skip:

Microsoft Merchant Center. The Bing-side equivalent of Google Merchant Center. Free to set up, structurally similar to GMC, and the input layer for Bing Shopping and Copilot product recommendations. Catalogs running GMC already have most of the data in the right shape; the work is duplicating the feed setup. For the specific feed schema and onboarding steps, refer to Microsoft Merchant Center’s product feed documentation.

Bing Webmaster Tools. The Bing analog of Google Search Console. Index status, click-through data, and structured-data parsing diagnostics — these don’t appear in GSC. Operators tracking organic performance only in GSC don’t have visibility into Microsoft-side indexing or surfacing.

FAQ and HowTo markup retention. A common 2024 reaction to Google’s narrowing was to strip these schemas. The schemas remain valid Schema.org and Bing continues to read them. Stripping them removes the input that was carrying Microsoft-side surfacing.

What’s not actually different

A few things sometimes overstated about Bing-side optimization:

• The product schema is the same. Schema.org is a standard. A Product block that validates for Google validates for Bing.
• GTINs and identifiers are the same. Universal identifiers do the same disambiguation work across both indexes.
• Crawler behavior is broadly similar. Bingbot and Googlebot both read JSON-LD at index time. The direct-fetch pattern Searchviu’s October 2025 test surfaced applies to both — index-time reading appears to be what feeds the AI surfaces; AI agents themselves may not consistently extract JSON-LD on direct page fetch.

The Bing optimization story isn’t “do something different.” It’s “don’t skip what’s already done for Google when the surface is Microsoft.”

What to do this week

Three concrete moves, in priority order:

1. Set up Microsoft Merchant Center if it isn’t already. The catalog feed is largely the same as the GMC feed. The surfacing payoff hits Copilot product responses and Bing Shopping simultaneously.
2. Verify the site in Bing Webmaster Tools. Index coverage, structured-data diagnostics, query data — all distinct from GSC, all useful for Microsoft-surface optimization.
3. Audit any 2024-era schema removals. If FAQ, HowTo, or other schemas were stripped in response to Google’s deprecation, restore them. They remain valid Schema.org and Bing continues to use them.

The honest position on Bing in AI commerce in 2026: it is one backend powering several AI surfaces, the catalog work to optimize for it largely overlaps with the Google-side work most catalogs are already doing, and skipping it forfeits the surfacing without the implied savings — because the work is already largely done.