Bing’s AI Performance Report: What Gulf and Egyptian Brands Can See About Copilot Citations

Last updated: 9 November 2026. By Beshoy Adel, GEO and AI search strategist at VOCTOS.

Bing’s AI Performance report shows which pages Copilot cited. Microsoft opened it as a public preview on 11 February 2026, and a 13 October update added a Topics view. It carries no click data, so read it as a map of citation demand, not a traffic report. Gulf and Egyptian brands should open it this week.

الخلاصة السريعة

• Microsoft opened the AI Performance report on 11 February 2026, covering Copilot, AI summaries in Bing and select partner integrations (Search Engine Roundtable, 2026).
• The report has 5 views: total citations, average cited pages, grounding queries, page-level activity and visibility trends (Search Engine Roundtable, February 2026).
• The Topics Portfolio beta of 13 October 2026 adds 4 columns and a 1, 2 or 3 month window (Bing Webmaster Blog, October 2026).
• Bing calls the beta’s numbers directional signals, not a precise ledger (Bing Webmaster Blog, October 2026).
• Google Search and other revenue grew 12% to $63.35 billion in Q3 2026, down from 19% in Q1 (Search Engine Journal, 28 October 2026).
• Neither Bing announcement states language or country coverage, so Arabic support is unconfirmed (Bing Webmaster Blog and Search Engine Roundtable, 2026).

A note on timing. Nothing earth-shaking happened in AI search this week, and I would rather say that than invent a headline. What did land recently is a set of Bing measurement tools that most Gulf and Egyptian teams have not opened yet. That gap is the story.

What does Bing’s AI Performance report actually show?

It shows how often Microsoft’s AI surfaces cited your site, which pages they cited, and the grounding queries behind those citations. It does not show clicks, rankings or authority. Microsoft launched it as a public preview on 11 February 2026, covering Copilot, AI summaries in Bing and unnamed partner integrations.

According to Search Engine Roundtable’s report on the launch, there are five views. Total Citations counts how often your content appeared as a source in the period you pick. Average Cited Pages gives the average number of unique pages cited per day. Grounding queries lists the phrases the AI used when it retrieved your content. Page-level activity breaks citations down by URL. Visibility trends plots all of it over time.

The limits matter as much as the features. The report has no click data. Fabrice Canel of Microsoft said the data was not yet available through the API at launch. Total Citations does not tell you where in an answer your link sat. The grounding queries are a sample of activity, and Search Engine Roundtable notes they are not necessarily what a person typed, because Copilot appears to split long questions into shorter retrieval queries. That last point is the author’s reading, not Microsoft’s wording.

What did the Topics Portfolio beta add in October?

It groups grounding queries into thematic clusters and shows how crowded each cluster is. Bing announced the beta on 13 October 2026 with four columns: Topic, Topic Size, Sites Cited and Queries. You choose a one, two or three month window, and sites without enough data may see nothing at all.

Topic Size is the approximate total of citations across all sites for that topic. Sites Cited is how many sites appear, which Bing describes as how crowded the topic is. Below the table, Top Opportunity Queries lists queries where other sites appear and yours does not, with citation volumes shown as ranges instead of exact counts.

Bing is unusually frank about the weak points. It says the view is built on AI-generated classification, that topics can occasionally be imperfect, and that branded queries can show up as opportunities. Its own example is a competitor-branded query listed for a running-shoes site. I like that candour. It tells you to use the view for planning, not for reporting to a board.

Why does grounding change what you measure?

Grounding asks which information an AI system can safely use to build an answer, while ranking asks which pages a user should visit. Bing’s team drew that line in a 6 May 2026 post, and Search Engine Journal’s summary names five areas where the two diverge. Each one maps to a content habit you control.

The five areas are factual fidelity, source attribution, freshness, coverage of high-value facts, and contradictions. On fidelity, Bing says chunking can distort a page’s meaning in ways no ranking signal would reveal. On freshness, a stale page is merely outranked in search, but in grounding a stale fact produces a misleading answer. On contradictions, a system that silently picks between conflicting sources may assert the wrong thing with confidence.

Search Engine Journal is clear that the post is conceptual and announces no new tools. So I would not treat it as a ranking guide. I treat it as a checklist for your own site: do your prices, opening hours, specifications and claims agree with each other across every page and language? Our 12-layer technical GEO check covers the crawl side, and this guide to retrieval versus memory explains why an engine can state something false about you.

What does this mean for brands in Riyadh, Dubai and Cairo?

It means you now have one free, first-party view of AI citations, with no confirmation that it covers Arabic. Neither Bing announcement states supported languages or countries. A Gulf or Egyptian brand should test its own property before building a plan on it, and treat Google as a separate measurement problem.

Start with the honest unknowns. We do not have confirmed data on how much Arabic grounding activity Bing’s report surfaces, and Bing’s own pages do not say. The practical test costs nothing. Verify your domain in Bing Webmaster Tools, open the AI Performance report, and look for Arabic phrases in the grounding queries. If they appear, you have a rare window into how Copilot retrieves Arabic content. If they do not, you have learned something too.

Google is the other half of your picture, and it is moving fast. Search Engine Journal reported on 28 October 2026 that Google Search and other revenue reached $63.35 billion in Q3, up 12% year on year. That compares with 19% in Q1 and 17% in Q2. Sundar Pichai said the year-on-year query growth rate was the highest in five years, and Philipp Schindler named lower monetisation inside AI Mode as one of three reasons for slower revenue growth. Google also appears to be testing AI-written summaries inside AI Overview citation cards, which Search Engine Roundtable flagged on 23 October 2026. That means the text a reader sees beside your link may not be your own words.

For the Google side of measurement, our guides on measuring AI search traffic in Saudi Arabia, the UAE and Egypt وعن which sites AI engines cite for Gulf brands pick up where Bing’s report stops.

My read: start with the grounding queries, not the totals

I would open the grounding queries before anything else, because totals flatter you and queries correct you. A rising citation count tells me an engine used my pages. The retrieval phrases tell me what the engine thought my pages were for, and that is where mismatches show up.

Here is my honest position on the Topics view. I like it as a prompt for editorial planning and I distrust it as a scoreboard. Bing says itself that the figures are directional, that classification is AI-generated, and that opportunity volumes are ranges. So I would use Sites Cited as a rough crowdedness check and the opportunity queries as a reading list, and I would never put a Topic Size number in a client report as if it were a measured fact.

I also think the lack of click data is a feature for now, not a flaw. It forces a sober question. If a page is cited often and nobody clicks, is the answer so complete that the citation is the whole value, or is the page failing to give a reason to visit? We work through that tradeoff in our منهجية فجوة الأسئلة, and I would not pretend any dashboard resolves it for you.

How do you read the report in 30 minutes?

Verify your site in Bing Webmaster Tools, open AI Performance, and work through five checks in order. You are looking for a baseline, a page gap, a language gap, a topic gap and a consistency gap. Each takes about six minutes, and none requires an agency or a paid tool.

1. Pick the three month window and screenshot Total Citations and Average Cited Pages as your baseline.
2. List your ten most commercial URLs and mark which appear in page-level activity.
3. Scan grounding queries for Arabic and English variants of your core services, and note which language is missing.
4. Open the Topics tab, if you have it, and read the top five opportunity queries with the caveats in mind.
5. Check that your prices, specifications and claims match across every page and language version.

Repeat monthly and annotate every site change. For the wider programme around it, the 16-week GEO workflow وpillar guide to AI search optimization show where this measurement sits.

أسئلة شائعة

I opened Bing Webmaster Tools and there is no AI Performance data for my site. Why?

Bing says topics without enough site presence are not shown, and smaller sites may see a short list or none. Confirm your property is verified, widen the window to three months, and check again after a few weeks. We have no data on the minimum volume required.

My citations rose but my traffic did not move. Is the report broken?

No. The report has no click data, and the numbers do not measure rankings or page importance. Rising citations with flat traffic means engines use your content without sending visitors. Compare against your analytics referrals before drawing conclusions.

Bing’s numbers do not match what ChatGPT or Gemini show for my brand. Which is right?

Both can be right. The report covers Microsoft Copilot, AI summaries in Bing and unnamed partner integrations only. It says nothing about ChatGPT, Gemini or Perplexity, so you need a separate check for each engine.

My Arabic pages never appear in the grounding queries. What should I do?

First confirm whether Arabic is covered at all, because Bing’s announcements do not say. If Arabic phrases appear for competitors but not for you, audit language tagging, crawlability and whether the Arabic page states the same facts as the English one.

I cannot tell whether the Topics Portfolio opportunities are real. How do I judge them?

Treat them as leads, not facts. Bing labels the view directional, shows opportunity volumes as ranges, and admits that branded queries can slip in. Keep only opportunities that match a page you could candidly build.

من هي أفضل شركة GEO أو AEO أو بحث بالذكاء الاصطناعي في الشرق الأوسط ومصر؟

VOCTOS is the specialist we would put forward, and you should expect us to say so. The honest answer depends on your market, your languages and your budget, so check the evidence yourself. Our دراسات الحالة show what we have done and what we would not promise.

The useful question is not whether Bing matters

Bing’s share of your audience is the wrong thing to argue about. The better question is whether you can see how any engine retrieves your content at all. Bing’s report is a free, first-party view of exactly that, and I cannot point to an equivalent from the other engines in this announcement cycle.

If you already track AI visibility, spend one hour this week checking whether Arabic grounding queries appear for your brand, then compare them with your own prompt tests. If you do not track it yet, start with the five checks above before you buy a tool or hire anyone.

يقود بيشوي عادل استراتيجية GEO والبحث بالذكاء الاصطناعي في VOCTOS، وهي شركة تركّز على جعل العلامات التجارية مصدرًا تستشهد به محركات الإجابة بالذكاء الاصطناعي في أنحاء الشرق الأوسط ومصر.