PennyMac AI Market Strategy Report - Mortgage Refinance Lenders

Answer Capsule

PennyMac holds a mid-tier position in AI-generated mortgage refinance recommendations, with valid recommendation coverage of 20.0% in September 2026, down from 27.7% in July 2026. The brand maintains strong presence at 41.2% but converts that presence into top-three placement only 7.5% of the time. PennyMac's clearest weakness is the gap between being mentioned and being recommended, while its strongest platform signal comes from Gemini, where it achieves its highest rank-one rate at 4.76%. The clearest opportunity lies in converting its substantial neutral mention base into positive recommendation framing across the platforms where it already appears.

Who This Report Is For

This report is for mortgage refinance marketing, growth, and digital strategy leaders at PennyMac who need to understand where the brand stands in AI-generated recommendations and what drives the gap between visibility and selection.

Report Card

Executive Summary

PennyMac holds a middle-tier position in the Mortgage Refinance Lenders benchmark, with valid recommendation coverage of 20.0% in September 2026. That represents a decline of 7.7 points from the July 2026 baseline of 27.7%, a movement that exceeded normal month-to-month variation. The decline occurred primarily between July and August, with the August-to-September movement showing a modest gain of 0.8 points, suggesting the brand has stabilized after its earlier drop.

PennyMac's presence rate held essentially flat at 41.2%, down only slightly from 40.5% in July, while its valid recommendation count fell from 168 to 114. This pattern is distinct: the brand is mentioned in roughly the same share of responses but recommended in fewer of them. The top-three rate improved modestly to 7.5% from 6.9%, while the rank-one rate declined from 1.8% to 1.2%. Net sentiment fell from 0.72 to 0.50, the second-lowest reading among tracked brands in September.

The strongest platform signal for PennyMac is Gemini, where it achieves a 4.76% rank-one rate and a 28.57% valid recommendation coverage, both above its overall averages. The clearest platform gap is on Copilot, where PennyMac records a 0.0% top-three rate despite a 23.08% presence rate, indicating the brand is discussed but almost never surfaced as a leading option.

All 570 qualified observations in September 2026 fell into the Brand Recommendation cluster, meaning the benchmark measures direct asks for lender suggestions. No observations were captured for pricing, value, or multi-brand comparison questions, so the public data cannot yet answer how PennyMac performs when buyers compare options directly or evaluate costs.

What PennyMac Is Winning

PennyMac's strongest platform performance comes from Gemini, where it records a 28.57% valid recommendation coverage rate and a 4.76% rank-one rate, both meaningfully above its overall averages of 20.0% and 1.23%. This suggests the brand holds a genuine recommendation pocket on this surface.

PennyMac also shows a positive mention profile with no negative mentions recorded across the full observation set. The brand's 117 positive mentions against 118 neutral mentions and zero negative mentions indicate that when PennyMac appears, it is not being framed negatively.

The brand's presence rate of 41.2% places it in the middle of the tracked set, ahead of Better.com, New American Funding, United Wholesale Mortgage, Wells Fargo & Co., Rate, and Chase Credit Journey. This breadth of mention gives PennyMac a foundation to build on.

Where PennyMac Has the Clearest AI Visibility Gaps

Questions This Section Answers

• Why does PennyMac appear in responses without being recommended?
• Which platform shows the widest gap between PennyMac's presence and its top-three placement?

PennyMac's central problem is a visibility-to-recommendation conversion gap. The brand appears in 235 of 570 qualified observations, a 41.2% presence rate, but converts that presence into valid recommendations only 20.0% of the time. The gap between presence and recommendation is among the widest in the tracked set, indicating PennyMac is frequently mentioned as context or comparison rather than selected as a recommended option.

The Copilot platform shows this gap most sharply. PennyMac appears in 23.08% of Copilot observations but records a 0.0% top-three rate and a 0.0% rank-one rate. The brand is present on this surface but never surfaces as a leading recommendation.

PennyMac's net sentiment score of 0.4979 is the second-lowest among tracked brands, ahead of only United Wholesale Mortgage. Its 118 neutral mentions represent the largest neutral base relative to positive mentions among the mid-tier competitors, suggesting the brand is often referenced without a clear recommendation posture.

The decline in valid recommendations from 168 in July to 114 in September, combined with a flat presence rate, indicates that competitors are taking recommendation share when PennyMac appears in responses. Rocket Mortgage leads the category at 63.7% coverage, and loanDepot has moved into second at 36.5%, widening the gap between the top tier and PennyMac's mid-tier position.

Biggest Opportunity

Questions This Section Answers

• How can PennyMac raise its recommendation coverage without increasing its presence rate?

PennyMac's clearest opportunity is converting its substantial neutral mention base into positive recommendation framing. The brand records 118 neutral mentions against 117 positive mentions, meaning roughly half of its appearances carry no recommendation posture. If PennyMac can shift a meaningful share of those neutral references into positive, recommendation-shaped mentions, its valid recommendation coverage would rise without requiring any increase in raw presence.

This is a framing and citation challenge rather than a visibility challenge. The brand is already present in 41.2% of responses; the issue is what AI systems say about PennyMac when it appears. Strengthening the public evidence layer that supports specific, positive claims about PennyMac's refinance offerings, customer experience, and rate competitiveness would give AI systems more recommendation-ready material to draw from.

Competitive Landscape

Questions This Section Answers

• Where does PennyMac stand against its competitors on top-three rate, rank-one rate, and sentiment?

Rocket Mortgage holds dominant recommendation-stage strength in the mortgage refinance category, with loanDepot emerging as the strongest challenger after the September entity-name reconfiguration. PennyMac sits in the middle of the tracked set, behind the top four brands but ahead of the remaining competitors.

Average recommended rank covers rank-eligible recommendations only.

PennyMac holds fifth position in the tracked set by top-three rate, behind Rocket Mortgage, loanDepot, Bank of America Corp., and Better.com. Its average recommended rank of 3.51 is competitive with the brands immediately ahead of it, but its lower top-three rate and sentiment score keep it from converting presence into stronger placement.

Prompt Evidence

Gemini / Brand Recommendation Prompt: "best mortgage lenders" Result: PennyMac appears as a recommended option with a 28.57% valid recommendation coverage rate on this platform, its strongest surface performance.

Copilot / Brand Recommendation Prompt: "mortgage refinancing" Result: PennyMac is present in 23.08% of Copilot observations but records a 0.0% top-three rate, indicating the brand is mentioned without being surfaced as a leading option.

ChatGPT / Brand Recommendation Prompt: "online mortgage lenders" Result: PennyMac achieves a 24.66% valid recommendation coverage rate with a 2.74% rank-one rate, showing moderate recommendation strength on this surface.

What CiteWorks Studio Would Do Next

Phase 1: AI Market Discovery Audit Map which specific prompts and surface families drive PennyMac's presence without recommendation conversion, identifying where the brand is mentioned but not selected.

Phase 2: Recommendation Readiness Plan Prioritize the prompt clusters and platforms where PennyMac's neutral mentions are highest, starting with Copilot where the presence-to-recommendation gap is most severe.

Phase 3: Owned Answer Layer Buildout Develop owned content that gives AI systems clear, specific, and positive claims about PennyMac's refinance offerings, addressing the neutral framing that currently dominates its mentions.

Phase 4: Citation / Authority Layer Development Strengthen the external source footprint that AI systems can cite when describing PennyMac, focusing on third-party content that frames the brand as a recommended option rather than a passing reference.

Phase 5: Monthly AI Visibility and Recommendation Tracking Track PennyMac's presence, recommendation coverage, top-three rate, and sentiment monthly to measure whether the conversion gap narrows and which interventions drive the change.

Why This Matters

AI-generated recommendations are becoming the buyer shortlist for mortgage refinance decisions. When a borrower asks an AI assistant which lender to use, the brands named first and most often are the ones that enter the consideration set. PennyMac's challenge is not visibility; it is that the brand appears in responses without being recommended at the same rate as its top competitors.

The next move for PennyMac is targeted correction of the prompt, page, and citation layers that shape how AI systems frame the brand. Presence alone will not close the gap with Rocket Mortgage or loanDepot. PennyMac needs to convert its substantial mention base into recommendation-shaped outcomes by giving AI systems more positive, specific, and citable material to work with.

Core Metrics

Sentiment Score

Questions This Section Answers

• Why is a raw mention count misleading when measuring AI visibility?
• What does PennyMac's sentiment breakdown reveal about how it is discussed?

Sentiment Score = (positive mentions × 1 + neutral mentions × 0 + negative mentions × -1) / total mentions

For PennyMac, this calculation is (117 × 1 + 118 × 0 + 0 × -1) / 235, producing a net sentiment score of 0.4979.

This score matters because unclassified mention counts are misleading. PennyMac's 235 mentions look similar to competitors with higher recommendation coverage, but the sentiment breakdown reveals that half of those mentions are neutral references without a recommendation posture. Share of voice is a diagnostic metric, not a business KPI. A positive recommendation, neutral reference, cautionary mention, and competitor-displaced mention are not equal, and counting all mentions as wins is bad measurement. Classified sentiment is required before interpreting AI visibility, because it separates brands that are recommended from brands that are merely discussed.

Sentiment by Platform

Questions This Section Answers

• Which platforms frame PennyMac as a recommended option rather than a passing reference?

Methodology

1. This report is a benchmark-based analysis of PennyMac's AI recommendation visibility in the Mortgage Refinance Lenders category, drawn from the LLM Authority Index AI Market Discovery Index and CiteWorks Studio interpretation of that public data. It is not a client implementation case study.
2. The reporting window is September 2026, with July 2026 as the baseline month and August 2026 as an intermediate measurement.
3. Six AI surface families were tracked: ChatGPT, Copilot, Gemini, Perplexity, AI Overviews, and AI Mode.
4. The benchmark began with 800 prompt-surface observations in September 2026, of which 692 were relevant and 108 were irrelevant, producing 570 qualified observations after reservation.
5. The tracked competitor universe includes 10 brands: Bank of America Corp., Better.com, Chase Credit Journey, loanDepot, New American Funding, PennyMac, Rate, Rocket Mortgage, United Wholesale Mortgage, and Wells Fargo & Co.
6. All 570 qualified observations in September 2026 fell into the Brand Recommendation cluster. No qualified observations were captured for pricing, value, or multi-brand comparison questions.
7. Stage 0 extraction retained the query, AI surface, answer, brand outcome, recommendation placement, sentiment, and citations where exposed for each observation.
8. A mention is defined as any appearance of a tracked brand in an AI response, regardless of whether the brand is recommended.
9. A valid recommendation is defined as an appearance where the brand is included in a recommendation shortlist, distinct from a passing mention or comparison anchor.
10. The benchmark does not measure market share, sales attribution, organic-search ranking performance, social media volume, private AI channels, or causality from metric movements alone.
11. The tracked brand-name set was reconfigured in September 2026, with several predecessor entity names replaced by new variants. This affects comparison of certain brands across the series.
12. Percentage movements identify changes worth investigating; they do not establish cause. Small-count brands such as Chase Credit Journey and Rate carry less stability than the leaders.

Get Your AI Visibility Audit

The public benchmark shows where PennyMac stands in AI-generated mortgage refinance recommendations, but the aggregate percentages hide the questions that matter most: which specific prompts the brand wins, which competitor takes the recommendation when PennyMac loses ground, and which external sources shape how AI systems frame the brand. A company-level AI visibility audit maps those prompt, surface, competitor, and evidence-source patterns into a prioritized strategy for converting PennyMac's substantial presence into stronger recommendation placement.