The 100 Most-Cited Websites in Microsoft Copilot for High-Stakes Insurance Consumer Decisions

LLM Authority Index Research | Microsoft Copilot | Insurance | July-September 2026

Microsoft Copilot uses a highly category-specific source ecosystem when consumers ask consequential questions about insurance coverage, carriers, premiums, policy types, and product comparisons.

Across 1,729 deduplicated Microsoft Copilot responses covering 12 insurance decision verticals, Forbes ranked #1, appearing as a cited domain in 27.59% of eligible responses. U.S. News ranked #2 at 20.19%, NerdWallet #3 at 17.18%, MoneyGeek #4 at 15.85%, and Money.com #5 at 8.62%.

The Insurance source market is much more concentrated than Copilot's broader High-Stakes Consumer Decisions source market. The Top 10 Insurance domains account for 44.0% of domain-response appearances, compared with 30.0% for Copilot overall. The Insurance Top 100 account for 78.1%, compared with 56.7% in the broader Copilot study.

Only 40 of the Insurance Top 100 domains also appear in Copilot's overall Top 100, a 25.0% Jaccard similarity. A strategy based only on broad Copilot source rankings would therefore miss 60 of the 100 domains that matter most in this measured Insurance decision panel.

Copilot also requires an unusually careful technical-source audit. In the deduplicated Insurance slice, 384 of 5,307 captured citation entries, or 7.2%, pointed to Bing, Bing API, Bing network, or OpenStreetMap infrastructure rather than substantive publisher pages. Those entries were excluded from the public ranking.

This is a source-visibility ranking. It is not a trust score, quality grade, recommendation score, or causal ranking. A domain can be frequently cited without causing Microsoft Copilot to recommend a particular insurer, carrier, policy, broker, or comparison service.

Answer Capsule

Which websites does Microsoft Copilot cite most often when answering high-stakes insurance questions?

Forbes leads this measured Copilot Insurance panel with 27.59% response citation coverage, followed by U.S. News at 20.19%, NerdWallet at 17.18%, MoneyGeek at 15.85%, and Money.com at 8.62%. The analysis covers 1,729 deduplicated Copilot responses, 4,923 public-facing citation events, 624 registrable cited domains, and 4,609 domain-response appearances across 12 insurance verticals.

Key Findings

• Forbes ranks #1, cited in 477 of 1,729 eligible Copilot Insurance responses, or 27.59%.
• U.S. News ranks #2 at 20.19%, followed by NerdWallet at 17.18%, MoneyGeek at 15.85%, and Money.com at 8.62%.
• The corrected Copilot Insurance slice contains 4,923 public-facing citation events, 624 cited domains, and 4,609 domain-response appearances.
• 384 technical or navigational entries were excluded: 236 from bing.com, 137 from bing.net, 9 from bingapis.com, and 2 from openstreetmap.org.
• The Top 10 account for 44.0% of response-level domain appearances, while the Top 100 account for 78.1%.
• Only 40 domains overlap between the Insurance Top 100 and Copilot's broader High-Stakes Top 100, a 25.0% Jaccard similarity.
• 37 domains remained in Copilot's Insurance Top 100 in July, August, and September.
• The median of the three monthly Top 100 Jaccard comparisons was 36.1%.
• Forbes ranked #1 in all three months and U.S. News ranked #2 in all three months.
• The final Insurance child studies show Copilot sharing 35 Top 100 domains with Google AI Overviews, 35 with Google AI Mode, 33 with Perplexity, 31 with ChatGPT, and 30 with Gemini.
• Copilot's pairwise Insurance Top 100 Jaccard similarities range from 17.6% to 21.2%, showing a particularly platform-specific source market.
• The parent Insurance study reports 80.2% temporal persistence, 29.0% cross-platform portability, and an approximate 51.2 percentage-point Persistence-Portability Gap at the family level.

Questions This Study Answers

• Which websites dominate Microsoft Copilot citations for high-stakes Insurance decisions, and how does that source hierarchy change by vertical?
• How persistent is Copilot's Insurance source market over time, and how portable is that authority to other AI platforms?
• What should insurance publishers, carriers, brokers, PR teams and CMOs measure before allocating resources to Copilot-focused content, earned media or partnerships?

Why This Research Belongs in Daily AI Search Marketing Decisions

Article 36 sits at the intersection of two commercially useful views of the same source market:

• the Microsoft Copilot platform view, which asks which sources recur across high-stakes consumer decisions on Copilot; and
• the Insurance decision-family view, which asks which sources recur across six AI platform families for one high-stakes commercial category.

For publishers and insurance specialists, that intersection supports a more precise claim than saying a domain is simply "visible in Copilot." A source can document whether it is cited specifically for consequential insurance decisions, whether that visibility persists over time, and whether its authority is broad across multiple insurance verticals or concentrated in one specialist decision journey.

For carriers, brokers, comparison sites, CMOs, PR teams and agencies, the operating question is:

Which sources are already present around the exact insurance decisions our customers ask Microsoft Copilot to help them make?

Five findings matter for daily planning:

• the Top 100 account for 78.1% of domain-response appearances in this Copilot Insurance slice;
• the Top 10 account for 44.0% of domain-response appearances;
• only 40 domains overlap between this Insurance Top 100 and Copilot's broader High-Stakes Top 100;
• 37 domains remained in the monthly Insurance Top 100 in July, August and September, and the median monthly Top 100 Jaccard similarity was 36.1%; and
• the final child-study Top 100 lists show Copilot sharing only 30 to 35 domains with each of the other five platform-specific Insurance Top 100s.

The parent Insurance Citation Authority Study provides the category-level Persistence-Portability Gap interpretation. It reports 80.2% temporal persistence, 29.0% cross-platform portability, and an approximate 51.2 percentage-point Persistence-Portability Gap for Insurance.

Article 36 answers a narrower Copilot-specific question. Its 36.1% median monthly Top 100 persistence should not be subtracted from one pairwise platform overlap to create a Copilot-specific gap. The parent PPG uses category-level summary statistics across the broader family.

The concentration result also should not be confused with stability. Copilot's Insurance source market is concentrated, but its monthly Top 100 still rotated materially. Concentration, persistence and portability are distinct measurements.

Citation visibility should remain separate from brand mentions, share of voice, recommendation rate, ranking, sentiment, citation-recommendation coupling and downstream business outcomes. A source can be frequently cited without causing Copilot to recommend a particular carrier, policy, broker or comparison service.

Commercial Action Matrix: What Publishers and CMOs Can Do With the Findings

The daily operating model is therefore four-layered:

1. Current Copilot Insurance core: maintain the sources that matter now, with the collection date attached.
2. Vertical and prompt-cluster layer: identify which sources dominate the exact auto, health, life, travel, pet, renters or other Insurance decision.
3. Longitudinal check: monitor whether priority sources persist, decline, rotate or change rank.
4. Cross-platform check: determine whether the same sources also matter in Google AI Overviews, Google AI Mode, ChatGPT, Perplexity and Gemini.

How This Study Fits the High-Stakes Consumer Decisions Research Corpus

Article 36 is the intersection of two parent studies:

• The Microsoft Copilot Platform Study
• The Insurance Citation Authority Study

It should also be read with:

• the 2026 AI Citation Authority Study;
• the Persistence-Portability Gap;
• Do AI Platforms Cite the Same Websites?;
• How Stable Are AI Citations?;
• Reddit's Decline in AI Citations for Insurance Decisions; and
• How Reddit Citations Changed in Microsoft Copilot for Insurance Decisions.

Articles 18, 19, 20, 34 and 35 provide the sibling Insurance cuts for Google AI Overviews, Google AI Mode, ChatGPT, Perplexity and Gemini.

What We Studied

The Insurance family contains 12 source vertical labels from the broader LLM Authority Index High-Stakes Consumer Decisions panel:

1. Car Insurance
2. Dental Insurance
3. Disability Insurance
4. Health Insurance
5. Life Insurance Companies
6. Long-Term Care Insurance
7. Medicare Supplement Insurance
8. Pet Insurance
9. Renters Insurance
10. Short Term Health Insurance
11. Travel Insurance
12. Vision Insurance

Study size

Long-Term Care Insurance, Renters Insurance and Vision Insurance do not have July source files in the preserved archive. Those missing months are unavailable, not zero, and are not imputed.

Responses by month

The Top 25 Microsoft Copilot Insurance Citation Sources

The Full Top 100 Microsoft Copilot Insurance Citation Sources

Citation Concentration

Copilot's measured Insurance source market is substantially more concentrated than its broader High-Stakes source market.

For comparison, the broad Copilot High-Stakes study reports 30.0% Top-10 concentration and 56.7% Top-100 concentration using the same response-level domain-appearance denominator.

This section answers

• How concentrated is Microsoft Copilot's Insurance citation market?
• How much of the measured source layer is captured by the Top 100?
• Is Insurance more concentrated than broad Copilot source behavior?

Insurance Specialization vs. Copilot Overall

Only 40 domains appear in both the Copilot Insurance Top 100 and Copilot's broader High-Stakes Top 100.

That is a 25.0% Jaccard similarity because the two 100-domain sets share 40 domains across a 160-domain union.

The result is not that the broad Copilot leaderboard is wrong. It answers a different question. Insurance changes both the source mix and the concentration pattern.

This section answers

• How different is Insurance from Copilot's broad High-Stakes source market?
• Would a broad Copilot Top 100 capture most category-specific Insurance sources?
• Does the category change which publishers rise to the top?

The Most-Cited Sources by Individual Insurance Vertical

The category-wide Top 100 is useful for finding recurring Copilot sources, but the source hierarchy changes materially by insurance decision type.

Several examples show why a category-wide list is not enough:

• MoneyGeek leads Car Insurance in this Copilot panel.
• Money.com leads Dental Insurance.
• NerdWallet leads Disability Insurance and Life Insurance Companies.
• Forbes leads Health Insurance, Medicare Supplement Insurance and Short Term Health Insurance.
• U.S. News leads Pet Insurance, Renters Insurance and Travel Insurance.
• eMedicare leads Vision Insurance.
• SeniorSimple leads Long-Term Care Insurance, although that vertical has only August and September source files in the preserved archive.

This section answers

• Do the same domains lead every Insurance vertical?
• Which sources lead the major Insurance decision types?
• Why should publisher outreach and content planning be organized at the vertical or prompt-cluster level?

Breadth vs. Specialization

Some Copilot sources recur across many Insurance verticals, while others earn high rank through specialist strength.

Forbes and MoneyGeek each appear in at least one response across 11 of the 12 Insurance verticals, while VetX reaches #7 overall despite appearing only in the Pet Insurance vertical. Squaremouth ranks #23 overall and is concentrated in Travel Insurance.

That distinction matters commercially. A broad source can be useful for cross-category Insurance authority, while a specialist source can be more relevant for a specific high-intent decision.

Monthly Persistence and Citation Drift

Copilot's Insurance source hierarchy changed materially across the three monthly snapshots.

The median of the three monthly Top 100 Jaccard comparisons was 36.1%.

Only 37 domains remained in the Copilot Insurance Top 100 in all three months.

The top of the ranking was more stable than the perimeter:

Forbes and U.S. News held the same #1 and #2 positions in every month. By contrast, HealthcareInsider fell from #5 in July to #21 in September. A three-month aggregate therefore captures recurring authority, but it should not be presented as a fixed current ranking.

This section answers

• How stable was Copilot's Insurance Top 100 from July through September 2026?
• Which leading domains were persistent at the top?
• Why is a one-time citation snapshot insufficient for ongoing source strategy?

Platform Differences: Copilot's Insurance Source Market Is Not Interchangeable With Other AI Engines

The final Insurance child-study Top 100 lists show meaningful cross-platform fragmentation.

No validated sibling platform shares more than 35 of Copilot's Insurance Top 100 domains.

The ranking order also differs materially. Copilot begins with Forbes, U.S. News, NerdWallet, MoneyGeek and Money.com. Google AI Mode begins with U.S. News and NerdWallet. ChatGPT puts Forbes even farther ahead. Gemini puts Money.com first. Perplexity begins with NerdWallet.

The operating implication is to maintain three layers: a cross-platform Insurance core, a Copilot-specific source layer, and a vertical-specific specialist layer.

This section answers

• Do the same publishers lead every AI platform for Insurance decisions?
• Which engines share the most Copilot Insurance sources?
• Should an insurance brand use one universal outreach list for all AI engines?

Prior Research and Why the Rankings Can Look Different

Ahrefs: broad all-topic U.S. Copilot queries

Ahrefs' September 2026 Copilot citation study tracks a broad U.S. all-topic query population of more than 3 million queries. Amazon leads its current Top 50 with 16.8% mention share, followed by Walmart at 12.6% and Wikipedia at 7.6%. Forbes ranks #19 at 1.4% mention share. Ahrefs: The 50 Most-Cited Websites in Copilot

That is very different from Article 36, where Forbes, U.S. News, NerdWallet, MoneyGeek and Money.com lead. The prompt population, category, metric and denominator are different.

Conductor: Copilot is one part of a much larger Insurance citation market

Conductor's August 2026 Insurance benchmark analyzes more than 3.6 million Insurance brand citations across more than 178,000 prompts from January through May 2026 across seven answer engines. In that corpus, Copilot generated 2.1% of measured Insurance citations and averaged 3.1 citations per answer. Conductor: 2026 Insurance AI Search Benchmarks

Conductor's denominator is citation volume across a large Insurance brand corpus. Article 36 ranks cited domains by response-level coverage inside a narrower high-stakes decision panel. The studies should not be merged numerically, but both show why Copilot should be measured separately from the other engines.

Microsoft: Copilot web grounding uses Bing search

Microsoft documents that, when web search is enabled, Copilot can generate search queries from the user's prompt and send those queries to the Bing search service to retrieve current web information. Microsoft Learn: Data, privacy, and security for web search in Microsoft Copilot

That architecture helps explain why raw Copilot exports can contain Bing-owned technical and navigation URLs. It does not mean every Bing-owned URL should be counted as an external publisher citation.

Tinuiti: platform and category mix change source behavior

Tinuiti's Q3 2026 AI Citation Trends Report tracks ChatGPT, Perplexity, Google AI Mode, Google AI Overviews, Gemini and Microsoft Copilot across multiple commercial categories. Its current public summary emphasizes substantial platform and category differences in source behavior. Insurance is not one of Tinuiti's tracked categories, so its percentages should not be imported into this Insurance ranking. Tinuiti Q3 2026 AI Citation Trends Report

Studies can disagree because of:

• query selection
• commercial intent
• geography
• time window
• ranking metric
• citation-event versus response-level counting
• source-type classification
• technical-asset exclusions
• domain normalization
• platform mode and UI behavior

The relevant question is not which ranking is universally correct. It is which ranking matches the decision environment and measurement contract a publisher or brand actually cares about.

Why This Study Is Different

This article combines several analytical layers that are often separated:

1. a defined high-stakes Insurance decision family;
2. a platform-specific Microsoft Copilot slice;
3. response-level domain coverage rather than raw-link volume as the primary ranking metric;
4. registrable-domain normalization;
5. explicit Bing and navigation-infrastructure exclusions;
6. monthly persistence analysis;
7. vertical-level source leaders;
8. direct Top 100 comparison with the other five AI platform families; and
9. direct comparison with the broader Copilot source market.

The goal is not to produce a universal list of websites Copilot trusts. It is to map the visible source environment for a specific, commercially consequential decision set.

Why AI Citation Studies Can Disagree

AI citation research can produce materially different rankings without either study being wrong.

A broad all-topic query set can elevate Amazon, Walmart and Wikipedia. A high-stakes Insurance panel can elevate Forbes, U.S. News, NerdWallet, MoneyGeek and Money.com. A brand-citation study can emphasize carrier domains. A response-level source study can elevate publishers that recur across many answers.

The measurement contract must therefore be stated before a leaderboard is interpreted.

For Article 36, the contract is:

• U.S.-oriented high-stakes consumer Insurance decisions from the preserved LLM Authority Index research corpus;
• July through September 2026;
• Microsoft Copilot observations only;
• exact-record deduplication before analysis;
• response citation coverage as the primary ranking metric;
• registrable-domain normalization;
• explicit technical-infrastructure exclusions; and
• no claim that citation presence proves trust, quality, recommendation or causality.

What This Means for Insurance Publishers

For insurance publishers, the most defensible opportunity is not to claim that a citation proves Copilot endorsement. It is to use the data as a source-market benchmark.

A publisher can ask:

• Do we appear in the Copilot Insurance Top 100 at all?
• Which Insurance verticals cite us?
• Are we part of the persistent monthly core or a one-month entrant?
• Which competitors or adjacent publishers appear in the prompts where we are absent?
• Does our Copilot visibility carry over to Google AI Search, ChatGPT, Perplexity or Gemini?

The 37-domain three-month persistent core is especially useful for separating durable visibility from one-time appearance. But persistence does not establish causality. A publisher still needs prompt-level review to understand which pages, claims, formats and topics are actually being cited.

What This Means for Brands and CMOs

For carriers, brokers, marketplaces and other insurance brands, citation measurement should be one layer in a broader AI-search measurement system.

At minimum, keep these measurements separate:

1. Citation visibility: which third-party and owned domains Copilot cites.
2. Mentions/share of voice: which brands Copilot names.
3. Recommendations: which brands Copilot actually recommends for a use case.
4. Rank: where a brand appears inside a recommendation set.
5. Sentiment: how the brand is framed.
6. Citation-recommendation coupling: whether a cited source is associated with a recommendation outcome in the same response.
7. Downstream outcomes: traffic, leads, applications, quotes, policies or other business results.

A high-citation publisher is not automatically the best place to spend money. It becomes more commercially interesting when the publisher repeatedly appears in the same high-value decision clusters where the brand wants to improve mentions, recommendations or recommendation rank.

Methodology

Research source

This platform-category study uses the same frozen source archive as the 2026 AI Citation Authority Study and the parent Insurance Citation Authority Study.

Platform filter

Only observations whose preserved platform label is copilot are included in this article.

Category assignment

The 12 source labels listed above are assigned to the Insurance family for category-level analysis.

These labels are treated as source memberships, not as proof that every prompt is perfectly classified. Prompt-level review is still required for narrower causal or vertical-specific claims.

Exact file deduplication

The preserved source files can contain overlapping records across vertical exports. A valid observation is counted once when all of the following match:

1. report month;
2. exact raw platform surface;
3. normalized prompt text; and
4. exact set of original citation URLs.

All originating vertical memberships are preserved on the surviving analytical record.

Prompt normalization

Prompt matching uses Unicode NFKC normalization, case folding, whitespace collapse and trimming. Wording is otherwise preserved.

Extraction failures

The Copilot Insurance source files contain 1 observation explicitly marked as an extraction failure. That observation is excluded from valid-response analysis rather than treated as an ordinary zero-citation response.

Registrable-domain normalization

Citation URLs are parsed to hostnames and consolidated to the registrable domain using the Public Suffix List.

This prevents subdomains from being counted as separate publishers when they belong to the same registrable domain.

Infrastructure exclusions

Before the public publisher ranking was built, the corrected Copilot Insurance slice contained 5,307 captured citation entries.

The analysis excludes 384 obvious technical or navigational entries:

• bing.com: 236
• bing.net: 137
• bingapis.com: 9
• openstreetmap.org: 2

The remaining 4,923 public-facing citation events form the secondary citation-event denominator used in this article.

Microsoft's web-search documentation explains why Bing-related retrieval surfaces can appear around Copilot's web-grounded responses. The exclusion is analytical: retrieval, image, thumbnail, search or navigation infrastructure should not be treated as an external publisher merely because it appears as a captured URL.

Primary metric

The primary metric is response citation coverage:

response citation coverage = deduplicated Copilot Insurance responses citing domain / all eligible deduplicated Copilot Insurance responses

A domain counts no more than once per response for this metric.

Secondary metric

Raw citation events are reported as a secondary metric. If one response cites several URLs from the same domain, those can contribute several citation events even though the domain receives only one response-level appearance.

Monthly ranks

July, August and September Top 100 lists are constructed independently using the same response-coverage metric and filtering rules. Missing vertical-month files are not imputed or treated as zeros.

Pairwise monthly overlap uses Jaccard similarity:

Jaccard similarity = shared domains / unique domains across both Top 100 lists

Vertical views

For the 12 vertical rows, every category record retains its originating vertical memberships. Response-level domain coverage is then calculated within each vertical membership set.

Because overlapping vertical records can share a common analytical record, vertical sample sizes should not be summed and treated as independent category observations.

Platform overlap

The corrected Copilot Top 100 is compared with the final published Top 100 lists for Google AI Overviews, Google AI Mode, ChatGPT, Perplexity and Gemini.

No causal interpretation

This study measures visible citations. It does not prove that a cited source caused Copilot to mention, rank, recommend or describe any specific insurance company in a particular way.

Limitations

1. The study covers July through September 2026. Copilot behavior can change after the measurement window.
2. The prompt corpus is designed around high-stakes commercial consumer decisions and is not representative of every Copilot query.
3. Three Insurance source labels lack July files: Long-Term Care Insurance, Renters Insurance and Vision Insurance.
4. Missing months are not imputed, but their absence means month-to-month samples are not perfectly composition-matched.
5. The public ranking is domain-level. It does not show which individual URL, author or page format drove every citation.
6. Registrable-domain normalization can combine multiple subdomains with distinct editorial functions.
7. Technical-infrastructure filtering is rule-based. Edge cases can remain.
8. Response citation coverage measures reach across responses, not the number of links in one response.
9. Citation frequency is not a quality score, factual-accuracy score or editorial endorsement.
10. Cross-platform overlap compares 100-domain sets and does not imply that shared domains have equal rank or equal citation intensity.
11. Category-level vertical assignment is a practical research taxonomy, not a universal classification system.
12. This article does not estimate the causal effect of publisher coverage, advertising, affiliate placement, PR or content changes on Copilot recommendations or business outcomes.

Commercial Disclosure

LLM Authority Index is an AI-search measurement and research property. Related businesses may provide AI-search tracking, research, publishing, optimization or consulting services.

Commercial relationships do not alter the frozen ranking logic in this study. The Top 100 is based on the preserved source corpus and stated response-level coverage methodology.

Advertising, affiliate relationships, sponsorship, consulting or outreach should not be interpreted as a way to purchase inclusion, ranking, recommendation or position in this research.

References

1. LLM Authority Index: The Most-Cited Websites in AI for High-Stakes Insurance Consumer Decisions
2. LLM Authority Index: The 100 Most-Cited Websites in Microsoft Copilot for High-Stakes Consumer Decisions
3. LLM Authority Index: The Persistence-Portability Gap
4. LLM Authority Index: Do AI Platforms Cite the Same Websites for High-Stakes Consumer Decisions?
5. LLM Authority Index: Reddit's Decline in AI Citations for High-Stakes Insurance Consumer Decisions
6. Ahrefs: The 50 Most-Cited Websites in Copilot, September 2026
7. Conductor: 2026 Insurance AI Search and AI Overviews Benchmarks
8. Microsoft Learn: Data, privacy, and security for web search in Microsoft Copilot
9. Tinuiti: Q3 2026 AI Citation Trends Report

Related LLM Authority Index Research

Parent studies and measurement framework

• 2026 AI Citation Authority Study
• Insurance Citation Authority Study
• Microsoft Copilot Platform Study
• Persistence-Portability Gap
• Do AI Platforms Cite the Same Websites?
• How Stable Are AI Citations?

Sibling Insurance platform studies

• Google AI Overviews
• Google AI Mode
• ChatGPT
• Perplexity
• Gemini

Insurance Reddit trend studies

• Insurance Reddit parent study
• Google AI Overviews Reddit trend
• Google AI Mode Reddit trend
• ChatGPT Reddit trend
• Perplexity Reddit trend
• Gemini Reddit trend
• Microsoft Copilot Reddit trend

Article 36 is the Microsoft Copilot x Insurance intersection study. The parent articles explain the broader platform and decision-family markets. The sibling platform studies show portability across engines. The Reddit studies provide one longitudinal source example inside the same Insurance decision family.