Copilot SEO: how to get your brand named and cited by Microsoft Copilot

What is Copilot SEO, and why does it differ from Google work?

Copilot SEO is search optimization aimed at Microsoft's assistant rather than a results page: the goal is to be one of the sources Copilot quotes and links when it writes an answer. The main difference from Google work is the index underneath. Copilot retrieves from Bing, so a site that ranks well in Google but is thin, blocked or stale in Bing starts at a disadvantage.

Copilot appears in more places than most marketing teams track: the Copilot app and copilot.microsoft.com, AI answers inside Bing results, the Edge sidebar and Windows. Microsoft's AI Performance report groups these as citations across Copilot, AI-generated summaries in Bing and select partner integrations.

The case for doing it separately is measurement, not size. Bing is a minority engine almost everywhere: StatCounter's September 2026 figures put Bing at 9.89% of US searches, against 85.19% for Google. Yet Microsoft reports citation share and competitor gaps per grounding query, so Copilot is often the easiest engine to prove progress in. If most of your buyers are in the United States, Canada or Australia, where Bing holds about a tenth of searches, the audience is large enough to plan for.

How does Microsoft Copilot choose which sources to cite?

Copilot rewrites the user's prompt into one or more search queries, retrieves candidate pages from Bing's index, and writes an answer grounded in the passages it selects. Microsoft calls those rewritten searches grounding queries, and they are what its reporting is organized around.

Two consequences follow. First, a page that Bing has not indexed, or ranks far down for the grounding query, is unlikely to be in the candidate pool at all. Second, the grounding query is often narrower than the buyer's prompt, so a page that answers one specific subtopic clearly can be cited even if it does not rank for the head term.

Copilot does not lean on the same sources as other engines. Temso's Q3 2026 study of 111,279 citations in accounting found clutch.co was the top source in Copilot, and that ChatGPT's and Copilot's sources overlapped by only 5%. Work done for ChatGPT SEO does not carry over automatically, which is why you should log Copilot's cited domains for your own prompts before deciding where to spend.

Which Bing settings can remove you from Copilot answers?

Check this before anything else, because a single meta tag can keep a page in Bing results but out of Copilot. Search Engine Journal reported in February 2026 that Microsoft rewrote the Bing Webmaster Guidelines to treat grounding results and citations as eligibility outcomes, and spelled out what each directive does to Copilot.

Many sites carry NOARCHIVE or NOCACHE from an old CMS template or a legal request years ago, with nobody remembering why. Crawl your site for these values in robots meta tags and X-Robots-Tag headers, list every page affected, and decide page by page. Bingbot does the crawling for Copilot, so also confirm in your server logs and CDN that Bingbot receives HTTP 200 responses. Cloudflare's AI crawler settings target bots such as GPTBot and ClaudeBot; Bingbot should not be caught by them, but verify it rather than assume. This is the first group of checks in our technical AEO reviews, alongside AI crawler access for the other engines.

How do you rank in Microsoft Copilot answers?

Ranking in Copilot is mostly Bing SEO plus pages written to be quoted. Fix Bing first, then rewrite the pages your buyers' prompts hit, then work on the sources around you.

Start in Bing Webmaster Tools: verify the site, submit the XML sitemap with accurate lastmod dates, and set up IndexNow so new and changed pages reach Bing quickly. Check the URL inspection results for your key pages and fix anything Bing reports as not indexed or blocked.

Next, rewrite the pages that match commercial prompts: pricing, comparison pages, alternatives pages, category pages and how-to guides. Write answer-first content, with each section opening on a one or two sentence answer under a heading phrased as the question. Use tables for comparisons, name the source of every figure and show a last-updated date. Copilot lifts passages, so each section should stand alone.

Keep brand facts identical across your site, Organization schema with sameAs links, LinkedIn and the directories your category uses. Treat schema markup as a consistency tool rather than a citation lever, and treat llms.txt as optional hygiene that no major engine has confirmed using.

Do not try to write for the model directly. Search Engine Journal reported that the rewritten Bing guidelines added a section on prompt injection and AI manipulation and expanded keyword stuffing to cover content designed to trigger citations or AI responses. Hidden instructions, fake FAQs and repeated brand claims risk the opposite of what you want.

Which sources does Copilot lean on in your category?

It depends heavily on the industry, so log the domains Copilot cites for your own prompts before spending. Published studies show how different the mix can be.

In professional services, Temso found Clutch at the top of Copilot's accounting citations, so accounting firms and agencies should treat a complete Clutch profile with verified reviews as Copilot work, not just lead generation. Amsive's category research found forbes.com took 24.91% of Copilot's citations for health insurance prompts, which points insurers and financial services brands toward editorial coverage and digital PR rather than more pages on their own site.

For B2B SaaS that usually means G2, Capterra and editorial best-of lists; for ecommerce brands, marketplaces and review-rich product pages; for local businesses, Bing Places, Google Business Profile and local directories. Earned mentions only: buying placements or seeding forum posts creates exactly the manipulation signals the Bing guidelines now describe.

How do you measure Copilot citations with Bing Webmaster Tools?

Use Bing Webmaster Tools AI Performance as your primary source, and back it with your own prompt runs and referral data. Microsoft introduced the report in public preview in February 2026, showing total citations, cited pages and the grounding queries behind them.

In June 2026 Microsoft added four views. Intents classifies grounding queries into groups such as Informational, Commercial and Local. Topics clusters related queries. Citation Share is, in Microsoft's words, the percentage of citations attributed to your site out of all citations shown across all sites for the same grounding query. Compare overlays the current period against a previous one.

On 13 October 2026, Microsoft's Bing team added Topics Portfolio as a beta tab. It shows each topic's approximate total citations across all sites, how many sites are cited and how many distinct queries it contains, plus opportunity queries: queries where other sites are cited and yours is not. Microsoft notes that topic labels are AI-generated and can be imperfect, figures are approximate, and smaller sites may see a short list or no tab at all.

Opportunity queries are the most useful new input for a content plan, because they name the exact questions where a competitor is cited instead of you. Export them monthly, map each to an existing page or a gap, and check the result in the next period with Compare. Alongside this, run a fixed prompt set in the Copilot app, since the report covers Microsoft surfaces only, and add a GA4 segment for referrals from copilot.microsoft.com and bing.com. Our AI visibility tracking service reports citation share from these sources each month.

What does a 90-day Copilot SEO plan look like?

Fix access and indexing in the first fortnight, rewrite in the middle, and judge on trends at the end. Ahrefs' self-promotional content experiment, in which a new brand went from about 0% to 65% of Copilot answers in three months, also found pages were cited on only about a third of eligible days, so read monthly trends rather than daily results.

• Days 1 to 15: verify Bing Webmaster Tools, export the AI Performance baseline, run your prompt set in Copilot, and audit NOARCHIVE, NOCACHE and snippet directives plus Bingbot responses in server logs.
• Days 15 to 45: fix indexing gaps, set up IndexNow, and rewrite the five to ten pages that match your highest-intent grounding queries, answer-first with tables and sources.
• Days 30 to 75: align schema and profiles, then close gaps in the third-party sources Copilot cites for your category, starting from the opportunity queries list.
• Days 60 to 90: compare citation share, cited pages and Copilot referrals against the baseline, and reprioritize from what moved.

Should you hire help for Copilot SEO?

Hire help when nobody owns Bing today, when the directive audit turns up problems across templates, or when your team lacks time to rewrite pages and earn mentions. Keep it in-house if your SEO team already manages Bing Webmaster Tools and only needs the Copilot-specific checks above.

Do not buy a separate Copilot program if your buyers are mainly in markets where Bing is small, such as India, where StatCounter's September 2026 data shows Google with almost all searches, or if your pages are not yet indexed and answer-first in any engine. In those cases, fix the foundations under a single AI search optimization program that covers Copilot alongside Google AI Overviews, Perplexity SEO and ChatGPT.

When you talk to an AEO agency, ask to see their prompt set, how they read citation share in Bing Webmaster Tools and what they would change in your robots meta tags. Be wary of promised positions, case-study lifts without baselines, or schema sold as the main fix. No one can promise a citation. For market rates, see our guide to what AEO costs. We start with a free AEO visibility check, quote a fixed price, and measure the same prompt set before and after changes so you can verify what moved.