In This Article
Knowing that “Copilot” is a family of products, and that only part of it runs through Bing, is half the job. The other half is knowing what to fix on one specific page. This article is the page-level workflow: a seven-block template you can run against any URL, and how NovaX shows you which block is failing.
In How to Optimize Your Website for Copilot I covered what Microsoft documents: which Copilot products have a web grounding path, what Bing Webmaster Tools controls, and what is guesswork. I will not repeat that here. This article asks a narrower question. Given a single page, what do you look at, in what order, and how do you know afterwards whether anything changed?
I use NovaX for that. It scores each page against engine-specific signals, one engine at a time. For Copilot, it models one particular idea: that the engine behaves like a commercial and action-oriented filter on top of the Bing index, one that prefers complete schema, primary sources and pages that lead somewhere. That model is useful, and it is also a model, not a view inside Microsoft. I keep the two apart throughout, and the section on evidence levels says exactly where the line sits.
What is in This Series
This article is part of the cluster How to Measure Your Presence in AI Search Engines. The hub also contains AI Search Visibility Share of Voice, and a set of articles for each engine:
- ChatGPT: official guide, how to measure and how to optimize with NovaX
- Perplexity: official guide, how to measure and how to optimize with NovaX
- Claude: official guide, how to measure and how to optimize with NovaX
- Gemini: official guide, how to measure and how to optimize with NovaX
- Copilot: official guide, how to measure and how to optimize with NovaX
- Grok: official guide and how to measure
The official guide articles say what the engine’s owner documents. The measure articles say how to track presence over time. The NovaX articles sit between them: they are the hands-on step where you inspect a page and decide what to change. NovaX has engine profiles for five engines, and Grok is not one of them, which is why it has no NovaX article.
The method is the same in each. The engine profile, and so what the blocks check, is different. The Gemini profile leans on schema and identity, ChatGPT on depth, Perplexity on density and Claude on structure and authorship. The Copilot profile leans on a complete schema bundle, canonical correctness and action paths.
What the Copilot Engine Is?
Copilot is the commercial and action-oriented engine: it asks whether a page is a primary source, whether its trust markup is complete, and whether it leads a reader somewhere. Those are the three questions the engine profile is built around, and every signal in the report maps to one of them.
- Is this a primary source? Is the page an official product page, a documentation hub or a whitepaper from the party that produced the information, or a third-party description of it? This is a research-backed signal in the report, and it is not one of the four weighted components.
- Is the trust bundle complete? Do Article, Organization and Person schema appear together, linked through the author and publisher fields? Schema and metadata carry the largest weight, 38%.
- Is the page clean and does it lead on? Is there one canonical address and one H1 (structural integrity, 28%), enough substance to synthesize an answer from (extractability, 22%), a credible update date (freshness, 12%), and a clear next step for the reader?
None of the three is something Microsoft documents as a selection criterion. The profile describes Copilot as powered by the Bing index and the Microsoft Knowledge Graph, and says canonical correctness and schema completeness matter more than prose depth. The Bing index part is consistent with what Microsoft documents for the Copilot products that use web grounding. The Knowledge Graph part, and the idea that Copilot validates authorship against it, are NovaX’s modelling choices.
Two points of scope matter before you read any score. First, “Copilot” covers Copilot Chat, Microsoft 365 Copilot, Copilot in Edge, the consumer app and several developer and security products. As I set out in the optimize article, Microsoft documents a web grounding path through Bing only for some of them. The NovaX profile scores page-level signals that are relevant to the Bing path. It does not tell you which Copilot product cited you, and it cannot see your Bing Webmaster Tools account or your robots.txt. Second, unlike most of the other engines in this series, Copilot has a first-party report. The AI Performance report in Bing Webmaster Tools shows citation activity for your site. A page score can suggest where to look, and that report is where you check whether anything happened.
How to Read the NovaX Copilot Report
Read two numbers first, then the signals underneath them, and never act on the headline alone. A Copilot report opens with an Engine Score out of 100 and a Citation Probability as a percentage. They answer different questions, and they can disagree.
- Engine Score: how completely the page’s signals match what the Copilot profile looks for, or in other words, how well Copilot understand your page. The report adds a score trend, and a comparison with your workspace average once you have inspected more pages.
- Citation Probability: a more cautious eligibility estimate. In the report I am using as the example, a page scored 68 out of 100 and 46% citation probability, and NovaX labelled that “moderate eligibility”, with a note to address the critical issues to push above 60%.
Those two are not the same. A page can match the profile well and still be a middling bet to be cited, because a retrieval system picks among competing pages and the choice depends on more than this one page. When the two diverge, open the detail before you celebrate either number.
The engine weights its score in four parts:
| Component | Weight | What it covers for Copilot |
|---|---|---|
| Schema and metadata | 38% | JSON-LD as the primary trust signal, with Person, Organization and Article together as the ideal |
| Structural integrity | 28% | A single H1, logical headings and canonical alignment with the Bing index |
| Extractability | 22% | Word count and structured content for answer synthesis |
| Freshness | 12% | dateModified in JSON-LD to confirm recency, at a moderate weight beside schema |
The report also scores entity clarity as a fifth core signal, which sits beside these four without a listed weight for Copilot. I read it as supporting context.
One more thing about these weights is worth knowing, because it lets you check the report yourself. In this report, a plain weighted average of the signals reproduces the headline score. The five core signals read schema 50, structural integrity 75, extractability 85 and freshness 81 for the four weighted ones. The arithmetic is 0.38 × 50, plus 0.28 × 75, plus 0.22 × 85, plus 0.12 × 81, which is 68.4, shown as 68. I ran the same check on the ChatGPT, Perplexity and Claude reports used in the sibling articles, and it reproduces each of those scores too. So the weights are not decoration. They tell you what a given signal is worth to the headline number.
The tabs then split the page into parts you can act on:
| Tab | What it tells you | What you do with it |
|---|---|---|
| Overview | Score, trend, citation probability, top actions, issues, engine-specific checks | Decide where to start |
| Signals | The five core signals scored individually | Find the weakest block of the template |
| Simulator | Projected score impact of each fix | Order fixes by payoff before you touch the page |
| Queries | The questions found on the page and how eligible each is | Spot weak or noisy question headings |
| Entities | Entities detected, their type, schema status and clarity | Decide which entities need markup or a clearer definition |
| Recommendations | Specific issues raised by the Inspector | Fix the cheap, concrete ones, and read them with care |
| Freshness | Freshness score and sensitivity for the engine | Decide whether the page is due an honest update |
One caution on the Recommendations tab. It comes from the Inspector, which looks at the page across engines, and not from the Copilot profile alone. In the example report, one recommendation says authorship signals matter for Perplexity and Gemini, and another says Gemini scores visual context separately. The four recommendations are word for word the same as in the Claude report for the same page. That is the evidence: they describe the page, and not the engine.
The Simulator is the part I use most. You tick the fixes you plan to make and it projects the score change. In the example, adding Article schema was projected at +16 and adding Person schema at +12. I come back to whether those numbers hold up in the worked example.
The Universal Copilot Page Template
Seven blocks cover everything the Copilot profile checks, and they apply to any page on any site. Work through them in order. The first is about not locking yourself out. The next two are the structural blocks. The fourth is the schema bundle, which carries the largest weight. The last three cover the research-backed signals and the content around them.
1. Access: do not exclude yourself
- Goal: Bing can crawl the page, the sitemap says when it last changed, and any grounding preference is a decision, not an accident.
- NovaX checks: whether bingbot is blocked. The report says that if it is, the page is invisible to Copilot and citation probability falls to zero.
- What Microsoft documents: as I covered in the optimize article, Microsoft’s new settings govern how content is used, and not whether it can be crawled. A robots.txt rule that blocks bingbot still blocks crawling. Sitemaps remain a core signal for coverage, with accurate ISO 8601
lastmoddates, and IndexNow complements them for faster updates, though it guarantees nothing. Grounding preferences (Allow, DisallowExtended and Disallow) can be set per page, and that is a decision that deserves legal input. - You check by hand: robots.txt for bingbot, noindex tags and leftover staging rules, the site’s status in Bing Webmaster Tools, the sitemap’s
lastmodvalues, and any grounding preference in page headers or tags. A page scanner cannot see most of these. For wider crawl and indexation problems, start with the Indexation and Crawl Diagnostic. - Done when: you can name the rule that governs bingbot on this site, someone owns the Bing Webmaster Tools login, and the sitemap dates match real edits.
2. Canonical: one address per page
- Goal: every page has one canonical URL, and it is the URL you want cited.
- NovaX checks: that a canonical URL is set. The report’s reasoning is that Bing de-duplicates heavily, so without a canonical Copilot may cite a different variant of the same page.
- Fix: set a self-referencing canonical on every indexable page, matching the address in the sitemap. Check variants: with and without a trailing slash, with tracking parameters, over http, on a staging host. Make sure internal links point at the canonical form.
- A rule I apply: do not point a canonical at a different page to consolidate signals unless the pages really are duplicates. A canonical that contradicts the content is worse than none.
- Done when: the canonical, the sitemap entry and the internal links all name the same URL.
3. Structure: one H1, clean headings
- Goal: the page has one clear topic, and its headings show where each idea begins and ends.
- NovaX checks: a single H1, a logical heading order, and canonical alignment. The report says Bing uses the primary heading as the main topic anchor for answer generation.
- Fix: one H1 that names the page’s subject and closely reflects the title tag. H2s for the main sections, H3s only inside them, with no skipped levels. Check the rendered HTML as well as the editor, because themes and page builders sometimes add a second H1.
- A rule I apply: phrase each heading for what the section answers. “Learn more” tells a retrieval system nothing about the section beneath it.
- Done when: the outline, read on its own, describes the page.
4. Schema bundle: Article, Organization and Person, linked
- Goal: the page carries the full trust bundle, and the three parts point at each other.
- NovaX checks: Article (or BlogPosting, FAQPage or HowTo) schema, and Person plus Organization schema. The report treats the three as a single required bundle, linked through the author and publisher fields, and carries a separate “schema completeness” check for it.
- Fix: an Article whose
authorpoints to a Person and whosepublisherpoints to an Organization, using shared@idvalues so the three nodes connect. Give the Person aname,jobTitleand author page URL. Give the Organization aname,urland a logo. Keep every value in line with the visible byline and footer. The skeleton is below. - A rule I apply: the markup describes the page, and it does not replace it. Validate it, and then compare it with what a reader sees. Anything in the JSON-LD that a reader cannot find on the page is a defect.
- Which pages this applies to: article pages. A category archive or a listing page is not an article, and marking it up as one would describe something the page is not. I make that argument in full in the Claude article, and it applies here too. The Organization part is different, because a publisher belongs on every page of a site.
- Done when: every article page has the three nodes, each with the same identity as on the page, and the author and publisher fields resolve to them.
5. Official source: be the primary source, visibly
- Goal: pages that are the first-hand source of something say so plainly, and carry the markers that tie them to the brand.
- NovaX checks: “official source signals”, meaning whether the page is an official product page, a documentation hub or a whitepaper from the primary source. The report says Copilot penalises third-party aggregators in favour of primary manufacturer or publisher data, and that an official spec sheet would be cited over a review describing the same specs.
- Fix: work out which pages on your site are first-hand: your product documentation, your specifications, your pricing, your published research, your own frameworks. On those pages, name the organization and its domain clearly, use Organization schema with the brand’s domain, and make sure the domain matches the brand entity as other sources describe it.
- A rule I apply: official status is a fact about who produced the information, not a tag you add. The report suggests an “Official Documentation” or “Product Specification” schema type. I could not match those names to types in the schema.org vocabulary. The closest I know are
TechArticleandAPIReferencefor documentation, andProductwith itsbrandandmanufacturerproperties. Check the vocabulary before using any type name a tool gives you, and do not invent one. - Done when: for each page, you can say what is first-hand about it, and the page says the same thing.
6. Actionability: lead somewhere
- Goal: a reader who has finished the page can see what to do next.
- NovaX checks: whether the page links to next steps, data sheets, pricing pages or official portals. The report’s reasoning is that Copilot sits inside commercial workflows in Microsoft 365 and Edge, so pages that move the reader toward an action score higher than static informational pages.
- Fix: add a short next-steps section that links to the most relevant action for this page: a product page, a data sheet, a contact route, a specification portal. Use descriptive anchor text that says what the destination is, such as “download the data sheet”, and not “click here”.
- A rule I apply: a next step is not a sales pitch. The Claude profile penalises promotional prose, and the Copilot profile rewards action paths, and the two are compatible. A plain list of next steps with accurate anchors is an action path in neutral language.
- Done when: every link in the next-steps section goes somewhere a reader of this page would want to go.
7. Depth and freshness: enough to answer, honestly dated
- Goal: the page holds enough structured substance to synthesize an answer from, and shows when it last changed.
- NovaX checks: a word count of 300 or more, and the freshness score, which includes whether
dateModifiedis present in the JSON-LD. Extractability carries 22% and freshness 12%. - Fix: lead with a direct answer, then support it with lists and tables where facts share attributes. Microsoft’s own guidance for AI search answers recommends concise answers, self-contained phrasing and critical details in HTML and not in PDFs or images. Set
dateModifiedto the real date of the last material edit, and keep the sitemaplastmodconsistent with it. - Rule: change the date only when the content changes. A date that moves without an edit teaches an engine to distrust it. A 300-word threshold is a floor for the check, not a target for the page.
- Done when: the schema date, the visible date, the sitemap date and the real last edit agree.
Ready to see how Copilot sees you?
If you would like the access, schema, source and content gaps scoped across a whole site, that is the work I do in Enterprise Search Advisory, and you can talk to me here.
The schema skeleton for block 4
Adapt this, and do not paste it as it is. Every value must match what is visible on the page. The shared @id values are what link the three nodes.
{
"@context": "https://schema.org",
"@graph": [
{
"@type": "Organization",
"@id": "https://example.com/#organization",
"name": "Your Organization",
"url": "https://example.com/",
"logo": "https://example.com/logo.png"
},
{
"@type": "Person",
"@id": "https://example.com/author/author-name/#person",
"name": "Author Name",
"jobTitle": "Their real job title",
"url": "https://example.com/author/author-name/",
"worksFor": { "@id": "https://example.com/#organization" }
},
{
"@type": "Article",
"mainEntityOfPage": "https://example.com/your-article/",
"headline": "Your article title",
"datePublished": "2026-01-15",
"dateModified": "2026-11-22",
"author": { "@id": "https://example.com/author/author-name/#person" },
"publisher": { "@id": "https://example.com/#organization" }
}
]
}
A Worked Example: One Page, One Report
The same page can score 75 on one engine and 68 on another, and the difference is arithmetic, not a change in the page. Here is a real NovaX Copilot report for one page. It is the same page as in the Claude article: the category archive for Diagnostics and Recovery on this site, which is a listing page and not an article. The specific numbers will differ on your pages. The reading order will not.
The headline: Engine Score 68 (“Good”), Citation Probability 46% (“moderate eligibility”), bingbot allowed, and four of six engine-specific checks passing, the two failures being Article schema and Person plus Organization schema. The five core signals read data extractability 85, freshness and decay 81, structural integrity 75, entity clarity 65 and schema and metadata 50. Two critical issues are listed: no content schema, and missing Person schema. Of the three research-backed signals, official source signals show as missing, actionability as present, and schema completeness as “check”.
| What the report shows | What I make of it | Action | Priority |
|---|---|---|---|
| Last Inspector run was 75 days ago | Every number below describes an older version of the page | Re-inspect first | 1 |
| Same five signal values as the Claude report for this page, but 68 here and 75 there | The signals are identical. The weights differ, so the headline differs | Read the weights before comparing scores across engines | 2 |
| Two critical issues: no content schema and no Person schema | These ask for article-type markup on a listing page | Decide the page type first. Do not add Article to the archive | 3 |
| Official source signals missing, called the highest-impact fix | A primary-source signal on a page that is a list of articles. The panel and the Simulator also disagree about the top fix | Find the pages that are first-hand, and treat the archive as a router | 4 |
| Schema completeness “check”, beside two failed schema checks | The panels do not agree on the state of the markup | View the page’s actual JSON-LD before acting | 5 |
| Actionability present, canonical, single H1 and 300+ words all pass | Blocks 2, 3, 6 and 7 are satisfied | Protect them in any template change | 6 |
| Recommendations: Article and Person schema, 3 images without alt text, a quarterly refresh cadence | The same four items as in the Claude report, partly written for other engines | Do the alt text. Move dateModified only on a real edit | 7 |
Four things in that table are worth saying out loud.
The same page scored 75 for Claude and 68 for Copilot, and nothing about the page changed. Claude’s profile puts 45% on structure, where this page scores 75, and 30% on entity clarity. Copilot’s profile puts 38% on schema, where the page scores 50, and gives it more weight than any other signal. The same page looks better to one engine than to the other. That is a reason to read a score as a statement about a profile, and not about the page alone.
Check a projected gain against the weights. The Simulator projects +16 for Article schema and +12 for Person schema, which is +28 together, or 96. If the schema signal went from 50 to 100 and nothing else moved, a 38% weight would add 19 points, which is about 87. The projection is larger than that, so it must be counting effects on other signals, or it is not a plain weighted average. I cannot see which from the report. Treat the projected figures as direction, not as a forecast.
The report’s panels can disagree about the top fix. The Overview lists Article and Person schema as the top actions. The research-backed section says official source signals are the highest-priority fix and the largest score driver. Yet the four weighted components already reproduce the headline score on their own, and official source signals are not one of them. So I cannot see where that signal enters the score. Both panels are useful, but they are describing different things, and the second is a claim I could not verify against the arithmetic.
Official source is a claim about who you are, and a category archive is not a source. An archive lists articles written elsewhere on the same site. It is not the first-hand home of a specification, a price or a research finding. So the missing signal is probably a page-type artifact, in the same way as the schema issues. The useful question is which pages on the site are first-hand for something, and whether those pages say so. For this site, that means the pages that define its frameworks and describe NovaX, and not its listings.
The result of working that list is a re-inspected page, a clearer decision about which pages deserve this audit, and a short fix list for the pages that do. It is not a guarantee of citations, which is the subject of the next section.
What NovaX Measures vs What Microsoft Documents
A NovaX score is a model of what tends to go with eligibility. It is not Microsoft’s own weighting, and nobody outside Microsoft has that. I built NovaX, so you should read my claims for it with that in mind. In the optimize article I noted that Microsoft does not document how Copilot chooses which retrieved pages to cite, and that applies to NovaX as much as to any other tracker.
So the weights in the report, such as the 38% on schema, are NovaX’s modelling choices. Here is how each part of the template stands against what Microsoft itself says, as I read it from the sources I cited in the optimize article and from Microsoft Advertising’s October 2025 guidance on AI search answers.
| Template block | Microsoft’s position or the outside evidence | NovaX’s treatment | My read |
|---|---|---|---|
| 1. bingbot, sitemaps and IndexNow | Documented: Microsoft’s settings govern use, not crawling. Accurate lastmod and IndexNow are documented | Pass or fail check on bingbot | Do first. The strongest evidence of any block |
| 1. Grounding preferences | Documented: Allow, DisallowExtended and Disallow, per page, and a legal call | Not checked | Decide it on purpose. A page scanner cannot see it |
| 2. Canonical | Not covered in Microsoft’s October 2025 guidance. The de-duplication reasoning is NovaX’s | Pass or fail check | Standard hygiene. Do it |
| 3. Single H1 and headings | The guidance says the H1 acts as the headline and should reflect the page title, and that H2 and H3 headings mark where ideas begin and end | Weighted 28% | The closest block to documented practice |
| 4. Article, Organization and Person schema | Reported secondhand from a Microsoft talk in March 2025 that schema helps its LLMs understand content. The guidance says schema helps machines interpret content, and does not claim it affects selection | Weighted 38%, the highest | Good hygiene. Treat the weight, and the Knowledge Graph validation story, as hypotheses |
| 5. Official source signals | Not documented. The guidance mentions authority and relevance only in general terms | Research-backed signal, outside the weights | A good reason to look at which pages are first-hand. The penalty on aggregators is NovaX’s claim |
| 6. Actionability | Not documented | Research-backed signal | Cheap and useful to readers. Do it in plain language |
| 7. Depth and snippable structure | The guidance recommends concise answers, lists, tables, Q&A blocks and self-contained phrasing | 300+ words, weighted 22% | Aligned in spirit. The word threshold is NovaX’s |
| 7. Freshness | The guidance says content should be fresh and current, without a definition. Accurate lastmod is documented | Weighted 12% | Cheap and honest if the date is real |
For the March 2025 point, I am relying on a secondhand report of a talk, so I would not quote it as a Microsoft publication.
What a score can and cannot do follows from that table. It can show you whether a page is complete against a defined checklist, order your fixes by projected payoff, catch regressions after a template change, and compare pages of the same type on the same site. It cannot prove that a fix caused a citation, show which Copilot product used the page, or see your robots.txt, your sitemap or your Bing Webmaster Tools account.
The table also shows where Copilot differs from the other engines in this series. Of the seven blocks, the crawler and indexing controls in block 1 and the heading guidance in block 3 have a Microsoft source. The schema weight, the official source penalty and the actionability reward are the profile’s own reasoning. They are inexpensive, and for the most part good practice anyway. But the engine that has a first-party citation report is also the one where it is easiest to check a claim, so use that report to do it.
The Fix-and-Verify Loop
Run the loop on one page first, then on the pages that matter most, and keep each round small enough to attribute. Changing seven things at once tells you nothing about which one worked.
- Inspect. Run the page through NovaX and open the Copilot report. Note the Engine Score, the Citation Probability, the page type and the date of the inspection.
- Check access by hand. Confirm the robots.txt rules for bingbot, the page’s noindex status, the site’s status in Bing Webmaster Tools, the sitemap
lastmodvalues and any grounding preference. The report cannot see them. - Simulate. Use the Simulator to see which fixes carry the highest projected impact, and for each one check whether it fits the page type and whether the projection is plausible against the weights.
- Fix in template order. Work through blocks 1 to 7, cheapest and most concrete first. Fix site-wide problems, such as an Organization node missing from every page or a second H1 from the theme, in the template and not page by page.
- Ask for a recrawl. Submit the changed URLs through IndexNow or Bing Webmaster Tools, and keep the sitemap
lastmodaccurate. Microsoft says IndexNow can prompt a recrawl and guarantees nothing, so allow time. - Re-inspect. Run the page again and compare. Did the blocks you touched improve, and did anything else regress?
- Measure over time. A score changing is not the same as being cited more. This is where Copilot differs from the other engines: the AI Performance report in Bing Webmaster Tools shows citation counts, a sample of grounding queries and page-level citation activity, with the surfaces combined. Read it alongside repeated prompt sampling, using the method in How to Measure Your Presence in Copilot, and see AI Search Visibility Share of Voice for the wider picture.
Two supporting reads help at the writing stage: How to Write a Blog Post for AI SEO and Citations and AI Content Structure for Enterprise Visibility. For the schema side in more depth, see Schema Markup for AI Search: What Actually Matters?
Where to Go From Here?
Start with the AI Visibility Inspector on one page you care about, and work the seven blocks in order. When you need Copilot tracked across your pages and against named competitors over time, NovaX covers it. And if you want the access, schema, source and content gaps scoped for your organization, contact me.
This is not a promise that a higher score earns more citations. It is not a replacement for the access checks, which no page scanner can do for you. It is not a claim that Microsoft uses NovaX’s weights, because it does not publish its own. It is not a way to become an official source by adding markup, since that status depends on who produced the information. And it is not a reason to put an Article on a listing page, because a page described as something it is not has become a less accurate page.
It is a way to inspect one page against a defined checklist, fix the cheap and concrete gaps first, and confirm afterwards, in the report Microsoft provides, that something changed.
FAQ
It is the engine profile that scores a page on commercial and action-oriented signals: whether it is a primary source, whether its schema bundle is complete, and whether it leads the reader to a next step. It scores page-level signals. It does not tell you which Copilot product cited you.
The Engine Score shows how completely a page’s signals match the Copilot profile. The Citation Probability is a more cautious eligibility estimate. They can differ, as a 68 score beside a 46% probability shows, and the gap is a reason to read the detail.
Per Microsoft, a robots.txt rule that blocks bingbot blocks crawling, and Copilot’s web grounding runs through Bing. So allowing bingbot is the first check. Microsoft’s separate grounding preferences govern how content is used, and not whether it can be crawled.
NovaX treats them as a single bundle and gives schema the largest weight. Microsoft has said, as reported secondhand, that schema helps its models understand content, and its guidance says schema helps machines interpret a page. Neither says it decides selection. I add the bundle to article pages because it is cheap and accurate, and I hold the weight loosely.
Because the profiles weight the same signals differently. In the example, the five signal values were identical in both reports. The Claude profile weights structure and entity clarity, and the Copilot profile weights schema most heavily, so the headline numbers differ.
Not on their own. The signal is about whether the page really is the first-hand source of its information. Markup can state who published a page, and it cannot make a page the primary source. Check any schema type name the report suggests against the schema.org vocabulary before using it.
They are different panels. The Simulator projects changes to the weighted signals. The research-backed section describes a signal that is not one of the four weighted components. I could not reproduce the headline score with it, so I read the Simulator as the guide to the number and the official source section as a prompt to look at which pages are first-hand.
After any material change, and before acting on a report that is old. In the worked example the last inspection was 75 days earlier, so the first step was to run it again.
No. You can promise a baseline, a fixed set of changes and repeated measurement, and for Copilot you can also point to the citation counts in Bing Webmaster Tools.
This article was researched and drafted with the assistance of AI tools and reviewed and edited by author prior to publication.