How AI Search Is Recommending Home Equity Loans

Home equity lending is becoming an AI-generated shortlist market. Borrowers are not only asking what a home equity loan or HELOC is. They are asking which lender is best, which bank has the best rates, which provider is fastest, which option is safest, and whether online lenders such as Figure can compete with banks, credit unions, aggregators, and mortgage brands.

The supplied LLM Authority Index public benchmark frames the category as a trust-heavy, rate-sensitive, comparison-driven market where the strongest signal is not raw visibility, but shortlist advancement. In the public benchmark, Rocket Mortgage is described as the broad directional leader across a larger mortgage-adjacent dataset.

The uploaded Figure structured dataset gives a more home-equity-specific read. Across the 297-observation Figure file, Bank of America had the strongest raw mention and valid recommendation coverage among the tracked lender set, Rocket Mortgage had the highest modeled monthly captured recommendation value, and Figure showed meaningful home-equity/HELOC specialist visibility but weak rank-one ownership.

Methodology

1. Market studied: Home equity loans, HELOCs, home equity lines of credit, home equity loan rates, home equity lender comparisons, mortgage-adjacent lender discovery, and related borrower shortlisting prompts.
2. Brands/entities included: Figure, Achieve, Bank of America, Bethpage Federal Credit Union, Connexus Credit Union, Discover Home Loans, LendingTree, PNC Bank, Rocket Mortgage, Spring EQ, and TD Bank. The raw observations also surfaced adjacent entities such as Navy Federal Credit Union, Alliant Credit Union, U.S. Bank, Better.com, loanDepot, Chase, Wells Fargo, and New American Funding.
3. Data collection date/window: May 2026 reporting window. The Figure structured extraction was loaded on May 20, 2026.
4. AI platforms tested: ChatGPT, Gemini, Perplexity, Copilot, Google AI Mode, and Google AI Overviews.
5. Number of prompts tested: The uploaded Figure structured dataset contains 297 AI-response observations across 255 unique prompt texts. A narrower home-equity/HELOC subset contains 100 observations across 74 unique prompt texts. The supplied public benchmark references a broader 887-observation Rocket Mortgage dataset, so this report treats the public benchmark as broader category context and the Figure JSON as the structured metric source for the analysis below.
6. Prompt categories: The structured dataset uses three cluster slots, but the labels appear stale and incorrectly reference “Humanoid Robots.” For publication, this draft interprets the actual prompt content as three home-lending intent zones: best lender / HELOC discovery, lender comparisons, and pricing/rate/cost research.
7. Definition of a mention: A company counted as mentioned when it appeared in an AI answer, regardless of whether the answer framed it positively, neutrally, comparatively, or as a valid recommendation.
8. Definition of a valid recommendation: A valid recommendation required positive, shortlist-quality framing. Neutral mentions, source references, factual appearances, and extraction-failed rows were not treated as recommendation credit unless the dataset marked them as valid recommendations.
9. Ranking/scoring metrics used: Raw mention presence, valid recommendation coverage, recommended top-three rate, recommended rank-one rate, average recommended rank, positive/neutral/negative visibility, net sentiment score by mentions, citation/source patterns, and modeled monthly captured recommendation value. Modeled value is a benchmark estimate, not realized loan volume, revenue, or funded originations.
10. Limitations: This is a point-in-time AI discovery benchmark. AI outputs vary across platforms, prompts, borrower geography, lender availability, rate conditions, underwriting requirements, and time. The uploaded structured dataset contains 42 extraction-failed fallback records, about 14.1% of observations. No Ahrefs export was supplied, so this draft does not make organic traffic, keyword ranking, DR, UR, or backlink claims.

Key findings

Bank of America had the strongest broad recommendation coverage in the Figure dataset. Across all 297 observations, Bank of America appeared in 43.1% of responses and received valid recommendation credit in 31.3%. It also had a 20.9% top-three recommendation rate and an 8.4% rank-one rate. In the home-equity/HELOC subset, Bank of America appeared in 48.0% of observations and received valid recommendation credit in 42.0%.

Rocket Mortgage had the strongest modeled value signal in the full structured dataset. Rocket Mortgage appeared in 35.7% of all observations, received valid recommendation credit in 22.6%, and posted the strongest modeled monthly captured recommendation value at $90,369.38. It also had the highest full-dataset rank-one count among the tracked brands, with 36 rank-one placements.

Figure was a strong home-equity specialist but rarely the default answer. In the home-equity/HELOC subset, Figure appeared in 44.0% of observations and received valid recommendation credit in 43.0%, slightly ahead of Bank of America on valid recommendation coverage. But Figure ranked first in only 3.0% of those observations, with an average recommended rank of 3.23 when recommended.

PNC Bank was the strongest rank-one lender in the home-equity subset. In the 100 true home-equity/HELOC observations, PNC Bank appeared in 43.0% of responses, received valid recommendation credit in 39.0%, ranked in the top three in 33.0%, and ranked first in 26.0%. That makes PNC the clearest top-rank winner in the home-equity-specific subset.

Aggregators and publishers were present, but not always treated as lender recommendations. LendingTree appeared in the dataset and had visible source-layer relevance, but its valid recommendation coverage was much lower than direct lenders. This mirrors the public benchmark’s broader warning: some brands become part of the answer without becoming the answer.

What changed in the market

Home equity lending used to be discovered through bank relationships, local credit unions, mortgage brokers, comparison pages, rate tables, search rankings, and paid media. Those still matter. But AI systems are now entering the borrower journey before the application click.

That changes the category structure.

A borrower asking “What bank has the best home equity loan?” is not just gathering background information. A borrower asking “Is Figure the best HELOC?” is asking for validation. A borrower asking “best bank for home equity loan rates” is already in a decision-stage comparison.

These are shortlist-formation prompts.

AI systems compress the research journey by deciding which lenders to surface, which ones to rank first, and which sources to use as evidence. In that environment, a lender needs more than name recognition. It needs AI-readable trust, rate competitiveness, product clarity, review support, comparison visibility, and source-backed differentiation.

What the benchmark found

The benchmark shows a category split between broad lending authority, bank trust, and home-equity specialist relevance.

Rocket Mortgage appears strongest in the broader mortgage-adjacent discovery layer. The public benchmark names Rocket Mortgage as the clear directional leader in a larger 887-observation dataset, with broad recommendation strength across mortgage and home-loan prompts. In the uploaded Figure dataset, Rocket Mortgage also led modeled monthly captured recommendation value and rank-one count across the full prompt universe.

Bank of America appears strongest in broad structured recommendation coverage. It led the Figure dataset in raw mention presence and valid recommendation coverage. That likely reflects bank trust, broad product recognition, and strong retrievability across HELOC, home equity loan, and mortgage-adjacent prompts.

Figure appears strongest as a specialist challenger, not as the default winner. Figure had high valid recommendation coverage in home-equity/HELOC prompts, which suggests AI systems recognize it as a relevant digital home-equity lender. But its low rank-one rate shows the gap: Figure is being included, but not consistently selected first.

PNC Bank appears to be a top-rank home-equity winner in the raw prompt subset. PNC’s strong rank-one rate in the true home-equity subset suggests AI systems often treat it as a credible default lender when the question is tightly focused on HELOCs or home equity products.

Credit unions and specialist lenders had narrower roles. Connexus Credit Union, TD Bank, Achieve, and Spring EQ received some recommendation credit, but not at the scale of Bank of America, Rocket Mortgage, Figure, or PNC Bank. Bethpage Federal Credit Union and Discover Home Loans had no measurable recommendation credit in the structured metrics.

Why visibility is not enough

The Home Equity Loans benchmark makes the visibility problem clear.

A lender can appear in AI answers because it is a large bank, a known mortgage brand, a rate-table source, a comparison marketplace, or a cited publisher. But that does not mean AI systems are recommending it as the best lender.

Figure illustrates the distinction. It had strong home-equity-specific presence and valid recommendation coverage, but weak rank-one ownership. In practical terms, Figure was often credible enough to include, but not consistently strong enough to lead the shortlist.

Rocket Mortgage illustrates a different pattern. It had lower home-equity-specific coverage than some bank competitors, but stronger full-dataset modeled recommendation value because it benefited from broader mortgage-adjacent prompts.

LendingTree illustrates a third pattern. It can appear as a comparison layer or source environment without being treated as the borrower’s primary lender recommendation.

For home equity lenders, the operating question is not “Do AI systems mention us?” It is “Do AI systems recommend us first for the exact loan, rate, and borrower-intent prompts we need to win?”

The citation layer

The citation layer is central to AI-generated home equity recommendations.

In the uploaded Figure dataset, frequently cited sources included Bankrate, CNBC, Bank of America, Rocket Mortgage, Money.com, NerdWallet, Forbes, Reddit, The Mortgage Reports, PNC, LendingTree, Connexus Credit Union, U.S. Bank, Better.com, Credit Karma, HUD.gov, TD Bank, and Figure.com. In the home-equity/HELOC subset specifically, Bankrate, Reddit, Connexus, The Mortgage Reports, CNBC, PNC, Navy Federal, U.S. Bank, Money.com, Bank of America, Forbes, Rocket Mortgage, TD Bank, NerdWallet, and Figure appeared repeatedly.

That pattern matters because AI systems are not only reading lender websites. They are synthesizing lender pages, editorial reviews, rate guides, comparison sites, bank product pages, credit union pages, government explainers, and consumer discussion sources.

Citation frequency is not endorsement. But it does reveal the public evidence layer AI systems can retrieve and summarize.

For home equity lenders, the citation layer is now part of the competitive funnel. A lender that is poorly represented in rate guides, comparison articles, product explainers, and trusted third-party sources may still appear in AI answers but lose top-rank recommendation credit.

What brands need to fix

Home equity lenders need to build for recommendation-stage confidence.

First, they need clearer product distinction. AI systems need to understand whether the lender is strong for HELOCs, fixed-rate home equity loans, fast digital approvals, high loan amounts, low fees, low introductory rates, broad state availability, or bank-relationship discounts.

Second, brands need better rate and fee transparency. Home equity borrowing is rate-sensitive. If AI systems cannot easily compare rates, APRs, closing costs, draw periods, repayment terms, and eligibility criteria, they may favor lenders with clearer public evidence.

Third, lenders need stronger third-party validation. Bankrate, NerdWallet, Forbes, CNBC, Money.com, The Mortgage Reports, and similar sources appear repeatedly in the citation layer. Stronger framing in those environments can influence whether a lender becomes a recommendation rather than a passing mention.

Fourth, specialist lenders need to convert inclusion into rank-one confidence. Figure’s current pattern suggests meaningful eligibility but limited default ownership. The likely growth lever is not generic visibility alone; it is better evidence that helps AI systems rank Figure above banks and broader mortgage brands for specific home-equity use cases.

Finally, lenders need prompt-level monitoring. “Best HELOC lender,” “best home equity loan rates,” “best bank for home equity loan,” “Figure vs bank HELOC,” and “home equity loan vs HELOC” are different competitive moments. Each one has a different shortlist.

How CiteWorks Studio helps

1. Map AI recommendation visibility. Track prompts, platforms, company presence, valid recommendations, top-three and rank-one performance, framing, and citation sources.
2. Identify the sources shaping AI answers. Find the editorial, review, forum, government, directory, owned, and search-visible sources that influence brand framing.
3. Build the citation architecture plan. Strengthen the public evidence layer so AI systems have more accurate, consistent, and persuasive source material to synthesize.

Commercial takeaway

Home equity loan discovery is being compressed into AI-generated lender shortlists.

The broader public benchmark points to Rocket Mortgage as a strong directional leader across mortgage-adjacent AI discovery. The uploaded Figure dataset shows a more nuanced home-equity picture: Bank of America had the strongest overall recommendation coverage, Rocket Mortgage captured the most modeled value across the full prompt universe, PNC Bank showed strong rank-one performance in the HELOC/home-equity subset, and Figure was frequently included but rarely selected first.

For home equity lenders, the growth opportunity is not simply to appear in AI answers. It is to become the lender AI systems can confidently recommend first for high-intent borrower prompts around rates, speed, trust, eligibility, product fit, and comparisons.

That requires stronger citation architecture, clearer public evidence, and more consistent lender framing across the sources AI systems use to construct borrower shortlists.

CTA

Want to know how AI systems are recommending your home equity loan brand?

CiteWorks Studio can map where your lender appears, where competitors are recommended instead, which prompts carry the most commercial risk, and which sources are shaping AI-generated borrower shortlists.

Request an AI Visibility Audit or Citation Architecture Review to see how your brand performs across recommendation-stage visibility, HELOC prompts, home equity loan rate prompts, lender comparison prompts, and the public evidence layer AI systems use to form lender recommendations.