How AI Search Is Recommending Mortgage: Monthly Trends

Executive Summary

Rocket Mortgage remains the coverage leader in the October 2026 benchmark, holding 72.5% valid recommendation coverage within the qualified set. That is a 0.2-point step down from its July 2026 baseline of 72.7%, but the more telling number is the 9.7-point jump from September 2026, which marks a second consecutive monthly gain. The gap to second place now stands at 26.9 points, with Veterans United at 45.6% coverage.

loanDepot is the standout mover this month, and the only brand classified as a significant riser. Its coverage reached 35.4% in October 2026, up 8.7 points from July 2026 and up 9.1 points from September 2026, both beyond normal month-to-month variation. The rise is broad: raw mention presence and top-three placement both climbed, up 6.0 points and 5.1 points respectively across the series. The brand has now gained for two consecutive months.

New American Funding is the sharpest decliner and the only significant decliner in the category. Coverage fell 13.9 points from 28.8% in July 2026 to 14.9% in October 2026, its third consecutive monthly decline. The retreat is presence-led: raw mention presence dropped 14.0 points over the same period.

The month is best described as mixed. Two brands moved beyond normal variation in opposite directions, while Rocket Mortgage, Veterans United Home Loans, PennyMac, and Freedom Mortgage were all classified as stable. Against the full four-month series, Rocket Mortgage's leadership has held throughout, and its coverage is now essentially back to where the series opened. Freedom Mortgage holds the lowest coverage figure in the category at 14.1%, a position addressed in detail below.

Each monthly run begins with 800 prompt-surface observations (560 unique questions in October 2026) across the benchmark's defined AI/search surface universe. Of those, all 800 mentioned a tracked brand or competitor; 759 were relevant and 41 were irrelevant. The public metrics use the 619 October 2026 observations that survive both qualification stages. For comparison, July 2026 opened with 800 prompt-surface observations (510 unique questions), 768 relevant and 32 irrelevant, resolving to 678 qualified observations.

Key Findings

AI Response Inconsistency Alerts

Questions This Section Answers

• What factual inconsistencies did AI platforms produce about mortgage lenders this month?
• Where did AI surfaces conflict on the largest FHA lender and PennyMac's credit score requirements?
• What did AI platforms disagree about regarding loanDepot's top 10 ranking?

The benchmark detected 4 critical or high-severity factual inconsistencies across 4 AI platforms: chatgpt, copilot, google_ai_mode, and google_ai_overviews. Each conflict represents a case where two AI surfaces gave materially different answers to the same question with high confidence.

PennyMac

AI platforms provided conflicting information about which lender is the largest FHA lender by loan count. When asked "Who is the largest FHA lender?", google_ai_overviews stated that PennyMac is ranked as the largest FHA lender by overall volume and number of loans issued, citing Scotsman Guide's 2024 Top FHA Lenders rankings, Bankrate, and Yahoo Finance. Copilot stated that United Wholesale Mortgage is the largest FHA lender by loan count, with 56,684 FHA purchase loans in 2025, citing Mortgage Research, Scotsman Guide's 2025 Top FHA Lenders, and MultifamilyBiz. The flagged source excerpts show the Scotsman Guide rankings listing PennyMac at number one and the Mortgage Research list placing United Wholesale Mortgage at number one, both with high confidence. Both claims cannot be true at the same time.

A second conflict on the same topic appeared when the same question was asked of a different surface pair. google_ai_mode stated that PennyMac is the largest FHA lender by total number of individual FHA loans issued, citing HUD.gov, The Motley Fool, and Yahoo Finance. Copilot again stated that UWM is the largest FHA lender by loan count, with 56,684 FHA purchase loans in 2025. The response sets referenced overlapping source pages, including the Scotsman Guide 2025 Top FHA Lenders rankings and the Mortgage Research list, meaning the platforms drew on partly shared evidence but arrived at different conclusions about which lender holds the top position.

A third inconsistency for PennyMac concerns a minimum credit score requirement. When asked "Who is the best FHA lender?", google_ai_mode implied that PennyMac accepts FHA loans with a 580 FICO score by listing it among lenders accepting 580, citing NerdWallet, Credible, and CNBC. Copilot stated that PennyMac's minimum credit score is 620 or higher, citing Mortgage Info, CNBC, and Forbes Advisor. The flagged NerdWallet excerpt shows PennyMac listed with a minimum credit score of 580 and a minimum down payment of 3.5%. A 580 minimum and a 620 minimum cannot both describe the same lender's FHA eligibility requirement.

loanDepot

AI platforms provided conflicting information about loanDepot's top 10 ranking inclusion. When asked "Who are the top 10 mortgage lenders in the US?", ChatGPT stated that loanDepot is ranked sixth among the top 10 mortgage lenders, citing Bankrate, Polygon Research, and The Wall Street Journal. Copilot stated that loanDepot is not included in the top 10 list, citing Bankrate, The Truth About Mortgage, and CNBC. The flagged source excerpts show Bankrate and CNBC both listing loanDepot within the top 10 by loan origination volume, and Polygon Research listing loanDepot at position six. loanDepot cannot both be the sixth largest mortgage lender and absent from the top 10 list in the same ranking.

Benchmark Context

Questions This Section Answers

• How was the qualified benchmark set defined for October 2026, and how did it change from July 2026?
• What happened to the recommendation-shaped answer share and valid recommendation shortlist share across the series?

The report separates the raw collection universe from the qualified analysis set. Brand-level recommendation percentages are calculated within the qualified benchmark set.

The following category-wide metrics summarize how the qualified benchmark shifted between the two months.

Benchmark-Level Metrics

The recommendation-shaped answer share dipped modestly across the series, while the valid recommendation shortlist share rose 1.5 points. The September 2026 qualified pool was 642 observations, between the July and October figures.

AI Recommendation Trend

Questions This Section Answers

• Which mortgage lenders held, rose, or fell in AI recommendation coverage from July to October 2026?
• Which brands moved beyond normal month-to-month variation, and which were classified as stable?

The category leader held its position with a two-month recovery back to the level it opened the series at, while the middle of the field reordered as loanDepot rose into a clear third and New American Funding settled at the bottom.

Coverage by Brand

loanDepot's 8.7-point rise and New American Funding's 13.9-point fall both exceeded normal month-to-month variation. The remaining brands moved within variation, though PennyMac's 4.7-point decline and Freedom Mortgage's 3.4-point decline were each classified as stable and should be read as directional rather than decisive.

What Changed This Month

Questions This Section Answers

• What drove loanDepot's significant rise in AI recommendation coverage?
• What explains New American Funding's third consecutive monthly decline?
• How did Rocket Mortgage recover to its July 2026 baseline, and what does Freedom Mortgage's lowest-in-category coverage mean?

loanDepot: The category's only significant riser

loanDepot's valid recommendation coverage rose 9.1 points from 26.3% in September 2026 to 35.4% in October 2026, one of the largest single-month moves this period, trailing only Rocket Mortgage's 9.7-point gain. Across the series it is up 8.7 points from 26.7% in July 2026. Both movements were classified as significant, and the brand has now gained for two consecutive months.

The rise is driven by both presence and placement. Raw mention presence climbed 6.0 points from 36.3% in July 2026 to 42.3% in October 2026, a significant move. Top-three placement rose 5.1 points from 11.7% to 16.8%, also significant. Rank-one frequency was essentially flat at 2.1%, up 0.5 points from 1.6%.

The distinction: loanDepot is being surfaced more often and recommended in the top three more often, but it is not yet taking the first position at a meaningfully higher rate. Its gains are concentrated in mid-shortlist placement rather than top-of-list wins. With 219 valid recommendations in October 2026, up from 181 in July 2026, the count movement reinforces the percentage trend.

Highest-priority diagnostic: Which prompts and surfaces account for loanDepot's rising presence and top-three placement, and why has that rise not converted into corresponding rank-one gains?

New American Funding: Third straight decline, now the lowest in the field by placement

New American Funding's coverage fell 2.1 points from 17.0% in September 2026 to 14.9% in October 2026, extending a now three-month declining streak. Across the full series the brand dropped 13.9 points from 28.8% in July 2026, the steepest decline of any tracked brand, and it is the only significant decliner this month.

Presence is the dominant driver. Raw mention presence fell 14.0 points from 30.8% in July 2026 to 16.8% in October 2026, a significant decline. Placement moved less: the top-three rate fell 2.4 points from 6.9% to 4.5%, and rank-one fell 0.9 points from 1.2% to 0.3%. Valid recommendations fell from 195 in July 2026 to 92 in October 2026.

The distinction: this remains a presence problem more than a placement problem. When New American Funding does appear, its placement quality has eroded less than its presence, though at 92 valid recommendations against a qualified set of 619, small count swings move the percentage visibly. The brand is now fifth by coverage but suffers the weakest top-three conversion of any tracked lender.

Highest-priority diagnostic: Which prompts, surfaces, or evidence sources stopped surfacing New American Funding across the four-month series, and which competitor is absorbing the recommendations it lost?

Rocket Mortgage: Leadership recovered to baseline

Rocket Mortgage's valid recommendation coverage rose 9.7 points from 62.8% in September 2026 to 72.5% in October 2026, a significant gain and its second consecutive monthly improvement. Against the July 2026 baseline of 72.7%, the brand is essentially flat, down 0.2 points across the series.

The recovery is supported by placement gains. Top-three placement rose 11.1 points from 51.6% in July 2026 to 62.7% in October 2026, and rank-one frequency rose 7.0 points from 31.1% to 38.1%, both significant. Raw mention presence was slightly lower at 96.3% versus 97.9% in July 2026, a non-significant change.

The distinction: Rocket Mortgage's presence has been near-universal throughout, so the series movement reflects recommendation conversion rather than visibility. The brand appears in nearly every qualified response and is now converting that presence into top-three and first-position placements at a higher rate than it did in September 2026.

Highest-priority diagnostic: Which surfaces and prompt types account for the gap between Rocket Mortgage's stable near-universal presence and the swing in its recommendation conversion across the series?

Veterans United Home Loans and PennyMac: Stable but diverging underneath

Veterans United Home Loans held second place at 45.6%, down 1.3 points from 46.9% in July 2026 and up 3.4 points from September 2026, both within normal variation. Its top-three rate improved 2.3 points to 31.3%, while rank-one slipped 1.6 points to 17.0%.

PennyMac's coverage slipped 4.7 points from 29.3% in July 2026 to 24.6% in October 2026, a move within normal month-to-month variation, best read as directional rather than decisive. Its raw mention presence fell 9.9 points from 45.1% to 35.2%, a significant decline, while top-three and rank-one placement were roughly flat. PennyMac is visible less often despite broadly stable placement when it appears.

The distinction across these two: neither posted a coverage swing beyond normal variation, so the more useful question is which sub-signals moved, not the headline rate. For each, modest coverage changes track a mix of presence and placement shifts rather than a single clear driver.

Highest-priority diagnostic: For each brand, check whether the sub-signal movement reflects the qualified denominator or a genuine shift in how AI systems describe and recommend the lender.

Freedom Mortgage: Lowest coverage in the category, with a widening gap to the leader

Freedom Mortgage recorded the lowest valid recommendation coverage in the category at 14.1% in October 2026, down 3.4 points from 17.5% in July 2026 and down 0.7 points from September 2026, both classified as within normal variation. Its rank-one rate fell to 0.3%, down 1.0 point from 1.3% in July 2026, on a small base of 87 valid recommendations.

Rocket Mortgage records 72.5% on valid recommendation coverage, while Freedom Mortgage stands at 14.1%. The more important issue is what that gap means for Freedom Mortgage's recommendation position: even against the more comparable mid-table brands, Freedom Mortgage trails PennyMac's 24.6% and New American Funding's 14.9%, meaning it now holds the single lowest coverage figure of the six tracked lenders, and its already-thin rank-one rate moved from 1.3% to 0.3% across the series. A brand converting this rarely into a first-choice answer, while also holding the weakest overall coverage, has an unresolved competitive question that the stable classification does not make less pressing: stability at the bottom of the category is not the same as safety.

Highest-priority diagnostic: Why does Freedom Mortgage convert into valid recommendations and first-choice placements less often than every other tracked lender, including brands with comparable or declining presence, and which prompts or surfaces would close the largest part of that gap?

A category gap that widened every month

Between loanDepot and New American Funding, the coverage gap moved from a 2.1-point deficit for loanDepot in July 2026 to a 20.5-point lead by October 2026, widening in each of the four tracked months. The gap between loanDepot and PennyMac also widened every month, from 2.6 points in July 2026 to 10.8 points in October 2026.

This is a category-observation about competitive repositioning, not a single brand's problem. Two brands that started the series clustered together have separated, with loanDepot rising as New American Funding and PennyMac held flat to lower. The benchmark shows the separation; it does not establish what caused it.

Highest-priority diagnostic: Which prompts and surfaces drive the divergence between these brands, and is the separation driven by one brand's gains, another's losses, or both?

Buyer-Intent Interpretation

Questions This Section Answers

• Which buyer-intent clusters survived qualification in October 2026?
• What can the benchmark currently say about how AI systems characterize mortgage pricing and head-to-head lender comparisons?

In October 2026, as in every prior month of the series, all 619 qualified observations fell into the Brand Recommendation class. No Pricing & Value or Multi-Brand Comparison observations survived qualification, meaning the public benchmark cannot yet answer questions about how AI systems characterize mortgage costs, rate competitiveness, or head-to-head lender trade-offs.

The practical implication: recommendation coverage remains the only commercially readable signal in this vertical at present. A lender winning on coverage is winning the discovery and consideration conversation, but that tells readers nothing about how AI would represent their pricing or comparative positioning. Those questions require a separate, company-level investigation.

Brand Opportunity Summary

Questions This Section Answers

• Which mortgage lender has the strongest Oct 2026 coverage signal, and which has the weakest?
• What highest-priority diagnostic applies to each tracked brand's current AI recommendation position?

The benchmark identifies where attention is warranted based on recommendation coverage movement; a company-level analysis is needed to explain why those movements occurred.

Evidence Behind the Benchmark

Questions This Section Answers

• What data does the benchmark use to calculate coverage, and what are its limitations for explaining movement?
• How does the smaller October 2026 qualified pool affect the interpretation of percentage changes?

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 prompt, competitor, surface, and evidence patterns to identify the specific drivers behind coverage changes. Source presence in a response is not automatically treated as proof of causation.

Report-Specific Interpretation Notes

• Two brands moved beyond normal variation across the July-to-October series, one up (loanDepot) and one down (New American Funding), while the remaining four were classified as stable. This mixed pattern means the category headline masks opposing movements underneath.
• The October 2026 qualified pool of 619 observations is the smallest in the four-month series, below July 2026's 678 and September 2026's 642, which affects how percentage movements translate into absolute counts.
• Movement identifies changes worth investigating; it does not establish cause. A coverage change can stem from prompt mix, surface behavior, evidence sources, or a combination, none of which the public benchmark can isolate.
• Small absolute counts matter: New American Funding's 92 valid recommendations and Freedom Mortgage's 87 are based on a qualified set of 619, so single digits of movement in count translate into visible percentage changes.
• The benchmark separates the raw collection universe from the qualified analysis set; brand-level percentages are calculated within the qualified set, not the raw prompt count.

About This Benchmark

This report is part of the CiteWorks Studio AI Visibility Industry Market research program.

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

Next Step

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

The aggregate coverage percentage tells you which lenders AI systems recommend, but not why. Beneath the October 2026 numbers sit unanswered questions: which high-intent prompts are being won, which competitor takes the recommendation when a brand loses, what attributes AI systems associate with each lender, and which external sources shape those outputs.

A company-specific AI visibility audit maps those prompt, surface, competitor, ranking, sentiment, and evidence-source patterns into a prioritized visibility strategy. That analysis can distinguish a presence problem from a recommendation problem, and can trace each movement to the underlying driver.

Request an AI visibility audit