Can AI agents built on Grounding with Bing Search find my website?

Microsoft's agent documentation for Azure AI Foundry lists two web tools on its new agents API: Web Search and Grounding with Bing Search. As of that documentation (last updated July 2026), both are generally available, and Web Search needs no separate Bing resource. Both draw on the Bing index.

That gives you a clear chain of reasoning:

1. An agent built with either tool retrieves web results from Bing.
2. A page Bing hasn't indexed can't come back in those results.
3. So Bing index presence is a precondition for being found by these agents.

The chain stops there. Being in Bing's index makes your page a candidate. It doesn't make the agent retrieve it, use it or cite it. Microsoft describes a multi-step grounding pipeline, and nothing in that documentation says how pages are picked from the index for a given answer.

If you want the consumer version of this story, see how Microsoft Copilot grounds answers in Bing. The same index sits underneath both.

Four conditions stack up. You control the first three.

Bingbot can reach you

robots.txt has to allow bingbot, and your network layer has to let it through. A firewall, CDN or bot-management rule can turn away a crawler that robots.txt allows. The operators document that mechanism, though how often it happens isn't measured. Our guide to bingbot, Copilot and your firewall covers the usual causes.

Bing has indexed the page

Reachable isn't the same as indexed. Only Bing Webmaster Tools tells you which URLs Bing actually holds.

You haven't suppressed the content

Bing announced in October 2025 that content marked with data-nosnippet is still indexed but excluded from snippets and AI summaries. noindex and nosnippet meta directives do similar work at page level. These tags often get copied from a template by someone hiding a cookie banner.

The agent actually searches

This one belongs to the agent's builder. A developer adds the search tool or leaves it out, and the system prompt steers when it fires. Retrieval is per query and discretionary across AI platforms. For the wider picture, see whether AI model APIs search the web by default.

The mechanism is documented. The effect size is not. No study in our evidence base measures how often a Bing-indexed page gets cited by a custom agent built on Grounding with Bing Search. Treat any vendor number on that as unverified.

The broader research points the same way. A July 2026 critical survey of 45 GEO studies found that no reviewed technique shows a stable, longitudinal, cross-platform causal effect on organic discoverability. Access is the part you can check. Selection is not.

One public test illustrates the index link. In OtterlyAI's 2026 experiment with two AI-generated sites, Google deindexed both while Bing kept them. Microsoft Copilot, which grounds on Bing, kept citing them: Otterly tracked 2,718 of 3,934 citations for one site coming from Copilot. ChatGPT cited them almost never, even while they ranked in Bing. That's one pair of low-stakes sites run by a tracking vendor, so read it as an illustration, not a measured rule. It fits the reasoning above: index presence is necessary, not sufficient.

It also cuts against a common shortcut. For the ChatGPT side, see whether ChatGPT uses Bing to find websites.

• You can't see inside enterprise agents. A company's internal agent runs in its own Azure tenant. Nothing public tells you whether it searched, what it queried or what it cited.
• The queries aren't the prompts. Agents generate their own search queries from the user's request. A list of prompts you test is a proxy for an input you can't observe.
• Grounding query data has a limited scope. Bing Webmaster Tools' AI Performance report covers Microsoft Copilot, AI summaries in Bing and select partner integrations. Microsoft doesn't state that every custom agent built on its APIs is included, so don't assume your report reflects them.
• IndexNow is hygiene, not a lever. Bing consumes IndexNow pings. The specification never mentions AI, and a 200 response confirms only that the ping was received.
• Third-party pages can misrepresent you. Agents grounded on web results can pick up manipulated content about your brand. Hidden text or injected instructions are illegitimate tactics, and they're a risk to watch for, not one to use.

These are all checks, not findings. Run them before you assume anything about your site.

1. Verify your site in Bing Webmaster Tools. It's the only source of Bing index status, and the only place that exposes grounding queries. If you're unsure, read whether Bing Webmaster Tools is worth setting up for AI search.
2. Inspect your key URLs. Pricing, product and about pages matter most. Confirm each one is indexed, not just crawled.
3. Read robots.txt for bingbot. Look for a blanket AI block or a stray Disallow that catches it.
4. Test the network layer. Request a page with bingbot's user-agent and look for a 403, a 429 or a challenge page. The AI Crawler Checker shows which crawlers your firewall turns away.
5. Grep for suppression tags. Search templates for data-nosnippet, nosnippet, max-snippet:0 and noindex, and confirm each one is intentional.
6. Export your grounding queries. The AI Performance report has no API for them yet, so export and analyze them by hand. Microsoft notes grounding query data is a sample.

None of this guarantees an agent will cite you. It removes the blocks you control.