How AI Search Is Recommending Flood Insurance: Monthly Trends

Executive Summary

Chubb remains the coverage leader in AI-powered flood insurance recommendations in October 2026, holding a 67.2% valid recommendation coverage rate. Allstate sits second at 56.8%, leaving a 10.4-point gap at the top that has narrowed from 22.3 points in July. The leader was stable across the series, and no brand recorded a significant movement between September and October.

Allstate was the only brand the benchmark classifies as a significant riser across the series, climbing from 46.6% coverage in July to 56.8% in October, a gain of 10.2 points that exceeds normal month-to-month variation. That marks three consecutive months of coverage gains for the brand. Its raw mention presence also climbed, from 78.0% in July to 89.6% in October, an 11.6-point increase alongside the coverage gain.

Aon Edge was the sharpest decliner, falling from 11.3% coverage in July to 5.8% in October, a drop of 5.5 points classified as significant. The brand's top-three placement rate fell from 5.8% to 1.5%, and its rank-one rate moved from 1.4% to 0.0% across the same span. FEMA NFIP and Hiscox were the other significant decliners across the series.

Against September, no brand recorded a movement large enough to be classified as significant. Chubb's coverage moved from 62.4% in September to 67.2% in October, and Allstate's moved from 49.0% to 56.8% — both changes within the benchmark's normal range for a single month. The more durable pattern remains the three-month view: Allstate's steady climb from the July baseline, and the three brands whose coverage has significantly declined since July.

Each monthly run begins with 800 prompt-surface observations (615 unique questions in October; 585 in July) across the benchmark's defined AI/search surface universe. Of those, 800 mentioned a tracked brand or competitor in each month; 356 were relevant and 444 were irrelevant in October, against 416 relevant and 384 irrelevant in July. August and September sat between those figures. The public metrics use the 259 qualified observations in October and 363 in July that survive both qualification stages.

Key Findings

AI Response Inconsistency Alerts

Three critical or high-severity factual inconsistencies were detected across four AI platforms (ChatGPT, Copilot, Gemini, and Google AI Mode), spanning two companies.

Allstate

AI platforms provided conflicting information about whether Allstate is currently writing new homeowners policies in California. In a high-severity availability conflict, Google AI Mode stated that Allstate "has submitted a filing to resume writing new homeowners policies across nearly the entire state, effectively moving to end its multi-year freeze," when asked who still sells homeowners insurance in California. Copilot, answering the same question, stated that Allstate remains closed to new business. Google AI Mode cited the San Francisco Chronicle, Insurance Business, and Latent Insurance; Copilot cited a list of California homeowners insurers still writing in 2026, Insurysis, and CNBC Select. A flagged source on the Copilot side, Latent Insurance, states that Allstate "has been paused since November 2022," while a flagged source on the Google AI Mode side reports that the company has asked California regulators for permission to start writing new policies again.

A second high-severity availability conflict involved a different pairing and question. When asked what companies are insuring homes in California, ChatGPT listed Allstate among major California property insurers offering homeowners coverage, citing the California Department of Insurance residential contact list, Farmers, and a State Farm newsroom update. Gemini, answering the same question, stated that Allstate "has largely paused or heavily restricted new standard homeowners applications," citing the Latent Insurance list as its source. The two claims cannot both be accurate.

Neptune Flood

AI platforms provided conflicting information about Neptune Flood's maximum building coverage limit. In a high-severity factual conflict, Google AI Mode stated that "Neptune Flood can offer building coverage up to $15,000,000" when asked who has the best flood insurance, citing AG Beck Insurance, Strange Insurance, and CNBC Select. Gemini, answering the same question, stated that Neptune Flood offers coverage limits up to $4 million for buildings, citing Insurance.com and Insurance Business Magazine. Two flagged sources sit on opposite sides: a Flood Insurance Guru post describing "higher coverage limits up to $15 million for dwellings" and a separate Flood Insurance Guru post stating the product is "ideal for mid-range homeowners needing fast, flexible coverage up to $4 million."

Benchmark Context

Questions This Section Answers

• How many prompt observations qualified for flood insurance analysis in October, and how does that compare to July?
• Why did the qualified observation count fall from 363 to 259?
• Did the number of AI surface families represented in the qualified set change across the series?

The raw collection universe is larger than the qualified analysis set. Brand-level recommendation percentages are calculated within the qualified benchmark set only.

The qualified surface breadth held at six AI/search families across both months, and the research-stage figures above summarize how the qualified set performed at the benchmark level.

Benchmark-Level Metrics

The qualified observation count fell for a fourth straight month, from 363 in July to 259 in October. August sat at 346 and September at 298, so the decline has been gradual rather than abrupt. The recommendation-shaped answer share moved in the opposite direction, rising from 38.6% in July to 54.1% in October, with September between them at 39.6%. Fewer qualified observations produced proportionally more recommendation-shaped answers in October.

AI Recommendation Trend

Questions This Section Answers

• Which brand led flood insurance coverage across the tracked months, and by how much?
• How much of the Chubb-Allstate coverage gap closed between July and October?
• How far does the third-place brand trail the coverage leader in October?

A Stable Leader With a Widening Gap Below It

Chubb has led the category in valid recommendation coverage throughout the tracked months, holding 67.2% in October against 68.9% in July, a change within normal variation. Allstate's 10.2-point gain across the same span is the only movement that exceeds the benchmark's normal range at the coverage level; it has narrowed the Chubb-Allstate gap from 22.3 points in July to 10.4 points in October. At the same time, the gap below Allstate has widened: Neptune Flood, in third place, trails Allstate by 40.6 points in October versus 27.0 points in July. No other brand exceeds 16.2% coverage, and Aon Edge, FEMA NFIP, Neptune Flood, and Wright Flood all eased from their July positions within normal variation.

What Changed This Month

Questions This Section Answers

• Which flood insurance brand rose significantly across the series, and how did its coverage and mention presence move?
• What happened to Allstate's top-three and rank-one rates despite its coverage gains?
• Why was Chubb's rank-one placement decline notable even though its leadership held?

Allstate: Three Consecutive Coverage Gains Build a Clear Second Position

Allstate's coverage rose to 56.8% in October from 46.6% in July, a gain of 10.2 points classified as significant, and up 7.8 points from September. This is the third straight monthly increase in the series. Earlier months contributed 1.1 points (July to August) and 1.3 points (August to September), making October the largest single-month increase within Allstate's three-month streak.

Raw mention presence climbed from 78.0% in July to 89.6% in October, an 11.6-point gain that moved in step with the coverage increase. That places Allstate in AI responses in nearly nine of every ten qualified observations. Its valid recommendation count stood at 147 in October against a qualified denominator of 259.

The distinction to notice is that Allstate's rise is happening at the coverage and presence layers, not at the top of recommendations. Its top-three rate held flat at 12.4% in July and October, and its rank-one rate slipped from 4.7% to 2.7%. The evidence suggests Allstate is being recommended more often overall without becoming more prominent in the highest positions.

Highest-priority diagnostic: Which prompt types are driving the presence and coverage gains, and why has the brand not converted that growth into top-three placements?

Chubb: Leadership Steady, Depth of Placement Easing

Chubb's coverage moved from 68.9% in July to 67.2% in October, a change of 1.7 points that sits within the benchmark's normal month-to-month variation. The series dipped to 62.4% in September before settling at 67.2% in October; both the July-to-October span and the September-to-October change fall within normal variation for the brand.

Raw mention presence eased from 86.0% in July to 83.4% in October, and the top-three rate fell from 28.6% to 25.1%. The rank-one rate also declined from 4.7% to 2.3%, with rank-one counts falling from 17 to 6. Chubb's valid recommendation count was 174 in October against 250 in July, on a smaller qualified denominator.

The distinction to notice is that Chubb's headline leadership is unchanged even as its top-of-response presence narrows. The brand remains the most frequently recommended option, but AI systems are placing it first less often in the current data.

Highest-priority diagnostic: Which prompts account for the decline in Chubb's rank-one placements, and did those placements shift toward Allstate or another brand?

Aon Edge: Three-Month Coverage Slide Now Classified as Significant

Aon Edge's coverage fell from 11.3% in July to 5.8% in October, a decline of 5.5 points classified as significant, and down 2.3 points from September. The brand has now declined for three consecutive months. This is the sharpest series-level drop among brands still actively tracked, after Hiscox's instrument change.

Raw mention presence fell from 13.0% in July to 8.1% in October, a 4.9-point drop that tracked the same direction as the coverage decline. The top-three rate fell from 5.8% to 1.5%, and the rank-one rate moved from 1.4% to 0.0%. The brand held 15 valid recommendations in October against 41 in July.

The distinction to notice is that the decline is broad rather than concentrated. Aon Edge is being mentioned less, recommended less, and placed first not at all. Net sentiment remained comparatively strong at 0.8, so the pattern is not driven by negative framing. Small counts apply here: 15 valid recommendations out of 259 qualified observations is a thin base for month-to-month comparison, and the trail lower warrants inspection rather than a firm conclusion.

Highest-priority diagnostic: Which surfaces stopped surfacing Aon Edge, and which brands appear in response positions the brand previously held?

FEMA NFIP: Frequent Appearance, Rare Recommendation

FEMA NFIP's coverage fell from 7.2% in July to 3.5% in October, a decline of 3.7 points classified as significant, and down 2.5 points from September. The brand's top-three rate rose modestly from 1.9% to 3.1%, and its rank-one rate from 1.1% to 1.9%, so the decline is not in placement quality but in how often the brand enters a valid shortlist.

Raw mention presence eased from 26.7% in July to 23.2% in October, a move within normal variation. The divergence between presence and coverage is the story: FEMA NFIP appears in roughly one in four responses but converts to a valid recommendation in only 3.5% of qualified observations. Net sentiment in October was 0.2, the lowest in the tracked set.

The distinction to notice is that FEMA NFIP is visible but rarely recommended. Only 9 valid recommendations sit behind the 3.5% figure, so the rate rests on a very small base and should be read as a directional signal.

Highest-priority diagnostic: What role does FEMA NFIP play in AI responses when it appears without being recommended, and which brand occupies the recommendation position instead?

Neptune Flood: Strongest Rank-One Rate in the Category

Neptune Flood's coverage eased from 19.6% in July to 16.2% in October, a 3.4-point decline within normal variation and down 0.2 points from September. The brand holds the third position by coverage and the strongest rank-one rate in the tracked set at 10.8% in October, down from 11.6% in July.

Its top-three rate moved from 17.4% to 14.7%, and its raw mention presence from 23.1% to 20.8%, both within normal variation. The brand held 42 valid recommendations in October. Average recommended rank of 1.4 indicates that when Neptune Flood appears in a recommendation, it tends to appear at or near the top.

The distinction to notice is that Neptune Flood competes differently from the coverage leaders. It enters fewer responses than Chubb or Allstate, but when it does enter, it frequently takes the first position. That is a placement-driven profile rather than a breadth-driven one.

Highest-priority diagnostic: Which prompt types produce Neptune Flood's rank-one placements, and which surfaces suppress its broader coverage?

Second-Tier Brands: Small Movements, Small Bases

Wright Flood held 11.2% coverage in October, down 2.6 points from July and down 2.6 points from September. Its top-three rate rose slightly from 6.3% to 6.6%, while its rank-one rate eased from 0.8% to 0.4%. The brand recorded 29 valid recommendations.

Palomar rose from 6.6% in July to 7.3% in October, a 0.7-point gain within normal variation. Its top-three rate held near flat at 3.9%, its rank-one rate moved from 0.3% to 0.0%, and its net sentiment improved from 0.8 to 1.0. The brand recorded 19 valid recommendations.

Assurant moved from 5.0% to 4.6% coverage, a 0.4-point decline within normal variation, with 12 valid recommendations. Hiscox Usa held 0.4% coverage with a single valid recommendation. The Flood Insurance Agency held 0.8% coverage with two valid recommendations and a rank-one count of zero.

The distinction to notice with these brands is that the counts are small. Hiscox Usa's one valid recommendation and The Flood Insurance Agency's two valid recommendations are valid signals for a niche vertical, but single-digit counts move percentages sharply. Assurant and Palomar show only net sentiment divergence worth watching, with Assurant at 0.5 and Palomar at 1.0.

Highest-priority diagnostic: Whether the smaller coverage bases for Hiscox Usa and The Flood Insurance Agency reflect genuine category presence or incidental mentions within broader response sets.

Buyer-Intent Interpretation

Questions This Section Answers

• Which buyer-intent cluster did every qualified flood insurance observation fall into?
• What can this benchmark not yet measure about price and brand comparison for flood insurance buyers?

Every qualified observation in the series, including all 259 in October, fell into the Brand Recommendation cluster. The benchmark found no qualified observations in the Pricing & Value or Multi-Brand Comparison clusters for this vertical.

The practical consequence is that this benchmark can measure which brands AI systems name when a buyer asks for a recommendation, but it cannot yet answer the commercial questions that sit behind a flood insurance purchase. Price competitiveness, value positioning, and head-to-head comparison outcomes are not represented in the qualified set. For a category where cost is a central buyer consideration, the public benchmark's read is limited to recommendation frequency and placement.

Brand Opportunity Summary

Questions This Section Answers

• Which flood insurance brands are the significant riser and decliner, and what signals define their positions?
• Which brands carry the strongest rank-one rate or the lowest sentiment in the October set?
• Which brands have such small recommendation bases that their percentages should be read as directional?

The benchmark identifies where attention is warranted based on coverage and movement; a company-level analysis is needed to explain why these shifts occurred and what they mean for each brand's AI visibility strategy.

Evidence Behind the Benchmark

The aggregate metrics are built from prompt-level observations, capturing the query, the AI surface, the recommendation outcome, the rank, sentiment, and citations where exposed. Company-level analysis can go deeper into which specific prompts a brand wins or loses, which competitors appear when a brand is absent, and which external sources shape AI responses. Source presence is not automatically treated as proof of causation in these findings.

About This Benchmark

This report is part of the CiteWorks Studio AI Visibility Industry Market research program. Explore the underlying methodology and standards through these research resources:

• AI Visibility Industry Market Methodology
• AI Visibility Industry Market Metrics
• AI Visibility Industry Market Standards

Report-Specific Interpretation Notes

Small-count movements are valid signals. Hiscox Usa's one valid recommendation and The Flood Insurance Agency's two valid recommendations in October are meaningful even though the counts are small. Percentages for these brands are calculated against the qualified benchmark set of 259 observations, so a single observation moves the rate by roughly 0.4 points.

The qualified denominator of 259 observations in October differs from the raw collection of 800 prompt-surface observations. The denominator also fell from 363 in July, 346 in August, and 298 in September, meaning percentage movements across the series reflect both brand-level changes and a smaller qualified set. Figures in this report are based on the qualified set unless explicitly stated otherwise.

Directional analysis identifies changes worth investigating. A movement between months, whether significant or not, does not by itself establish the cause of that change. The benchmark highlights where to look; a company-level audit explains why. The Hiscox to Hiscox Usa change was an instrument adjustment in September, not a market signal, and Hiscox Usa's October figures should be read on that basis.

Next Step

The Public Benchmark Shows Where a Brand Is Winning or Losing. A Company-Level Audit Shows Why.

Beneath the aggregate coverage percentages lie specific questions: Which high-intent prompts does a brand win, and which does it lose? When a brand is not recommended, which competitor takes its place? What attributes do AI systems associate with each option, and which external sources shape those associations? For flood insurance, where Allstate's coverage gain, Aon Edge's series decline, and the conflicting availability claims about Allstate all point to questions the aggregate layer cannot answer, these details determine whether a brand's AI visibility is an asset or a liability in the buying journey.

A company-specific AI visibility audit maps the prompt, surface, competitor, ranking, sentiment, and evidence-source patterns into a prioritized visibility strategy. Where the benchmark identifies movement, the audit explains the mechanism and defines the response.

Request an AI visibility audit