Why the timing of the title search, not its price, is the variable that protects wealth, defeats fraud, and holds evidence together in an artificial intelligence world.
Three forces, concentrated household wealth, industrialized impersonation fraud, and the collapsing cost of fabricating a document, all reward the same operational change: know the title on day one.
Title insurance is not priced like casualty insurance because it does not behave like casualty insurance. Industry-wide, claims cost just over four cents of the premium dollar; the remainder funds the search, the examination, the curative work, and the operating cost of doing them.1 In 2025 the industry paid $32 per policy in direct losses against $1,367 per policy in net premiums earned.2 That is the signature of a product whose entire value is realized before closing. A buyer, a lender, or a borrower who receives that work at day one rather than at day forty-five receives the same product with materially more optionality.
American households held $49.8 trillion of real estate and $35.8 trillion of owners' equity in it as of the second quarter of 2026, 71.9 percent of the value of the real estate they own.3 For households between the 25th and 50th percentile of net worth, real estate represents roughly 155 percent of net worth.4 A title defect is not a paperwork problem for these households; it is the whole balance sheet.
The share of title firms reporting at least one seller-impersonation attempt rose from 28 percent in a 2024 survey to 59 percent in a 2026 study; the share reporting three or more attempts rose from 4 percent to 23 percent.5 FBI IC3 losses in its "Real Estate" crime category climbed from $145.2 million in 2023 to $275.1 million in 2025.6 Fraud that depends on an owner not being verified is defeated by verifying the owner early, while there is still time to act.
FinCEN has reported that criminals use generative AI to produce falsified identification documents, photographs, and video to defeat customer identification controls.7 Fifty-eight percent of title professionals now rate deepfake image or voice technology as at least somewhat common in impersonation attempts.5 When a document can be manufactured in minutes, the durable asset is the chain of examined, sourced, dated evidence behind it.
Appraisal, environmental, survey, insurance, and credit all begin early because a late answer is an expensive answer. Title alone is routinely ordered last, and then blamed for the delay it was never given time to prevent.
The customary order of operations in a financed real estate transaction was not engineered. It accreted. Title is ordered once a file looks likely to close, on the reasoning that a search performed on a transaction that dies is a search wasted. That made sense when a search was a manual abstract and the closing calendar had slack. It makes far less sense now, when the search is the one workstream capable of revealing the facts that decide transactions, and the calendar has none.
The average purchase loan closed in 36.8 days in March 2026, the fastest since ICE began tracking the metric in 2019. The typical loan moved from application to rate lock in 11 days and from rate lock to closing in 26 more.1 Ordering title at conditional approval places the search inside that final 26-day window, alongside underwriting conditions, appraisal reconciliation, insurance binders, and payoff demands. Anything it finds must be cured inside a period already fully allocated.
The premise that most files are clean does not survive the industry's own measurement. In a 2024 study of 2023 files, title professionals reported that 36 percent of files on average required substantial, nonroutine curative work, and that 62 percent of firms performing curative work addressed four or more issues per transaction.2 A standard file consumed 22.0 hours on average; a difficult file consumed 45.4 hours, 2.1 times longer.2
No public authority publishes a per day cost of a delayed closing, and this paper will not invent one. The closest published anchor is Freddie Mac's 2024 Cost to Originate study, which valued automated collateral evaluation at $179 per loan for 7.3 days of cycle time saved.3 Lenders price the same variable directly: one wholesale rate lock policy charges 2.5 basis points per day to extend a lock.4 Against total loan production expenses of $10,936 per loan in the second quarter of 2026,5 days are not a soft cost.
And delay is not rare. In the August 2026 REALTORS® Confidence Index, 14 percent of contracts had delayed settlements in the prior three months and 7 percent terminated outright.6 The most recent NAR cause breakdown attributes 11 percent of delayed contracts to titling and deed issues, a 2018 figure, cited with its vintage.7
The engagement in one paragraph A Pre-Policy Intelligence order is placed at day one for $149, on its own, no policy commitment, no executed contract required. An initial response is targeted within 24 hours, covering ownership, liens, encumbrances, judgments and taxes, and the legal description. Close with an Intelligent Title policy and the $149 is credited in full against the premium. If the file doesn't close with an Intelligent Title policy, the fee covers the search and examination, and the findings belong to you.
| Finding | What the initial response covers |
|---|---|
| Ownership | Vested owner of record, chain of conveyance, and the identity questions that follow, entity status, deceased owners, absentee owners, recent transfers. |
| Liens | Mortgages and deeds of trust of record, mechanics' liens, and the payoff and release posture that sets lien position. |
| Encumbrances | Easements, restrictions, covenants, and rights of record that constrain use, access, or collateral value. |
| Judgments and taxes | Docketed judgments, state and federal tax liens, and delinquent property taxes, including assessments not yet on the roll. |
| Legal description | The description that will appear on the instrument, reconciled against the record. |
The trigger differs by transaction type; the principle does not.
| Stage | What happens |
|---|---|
| Day 1 | Letter of intent, term sheet, loan application, or signed contract. Pre-Policy Intelligence is ordered. The property, not the file's probability of closing, justifies the search. |
| Within 24 hrs | Initial response delivered. The parties learn what is on the record before the appraisal is ordered, before the rate is locked, and before a closing date is promised. |
| Days 2–30 | Curative work runs in parallel with underwriting rather than behind it. Payoffs, releases, estate matters, and description corrections are worked on the calendar's slack, not its deadline. |
| Closing | Policy issues, underwritten by Stewart Title Guaranty Company or WFG National Title Insurance Company, with the $149 credited against the premium. |
IRIS prices your policy whenever you ask — by chat, voice, or email. She opens every file the moment the order completes checkout. She examines the title record and communicates its details — ownership, liens, encumbrances, judgments, survey status, lender conditions — in plain language, with every finding cited to its source and search date. She manages the timeline and the updates and remains engaged through closing and policy issuance, alongside licensed title professionals. Cleared title is determined by the licensed examiner and the underwriter; IRIS reports it. IRIS brings the automation and efficiency. The humans bring the licensure.
Thirty-five point eight trillion dollars of American household equity rests on the accuracy of county records. The households with the most at stake are not the wealthiest ones.
As of the second quarter of 2026, U.S. households held $49.79 trillion in real estate at market value and $35.81 trillion in owners' equity in that real estate, 71.9 percent of the value of what they own, against $13.98 trillion of one-to-four-family home mortgages.1 Owners' equity equals roughly 18 percent of total household net worth of $195.87 trillion.1
Aggregates understate the risk. Using 2022 Survey of Consumer Finances data, the Richmond Fed found that for households between roughly the 25th and 99th percentiles of wealth, housing is by far the largest component of assets; only at the very top do equities and business interests outweigh real estate.3
Real estate as a share of net worth, by net-worth percentile, 2022 Survey of Consumer Finances. Shares exceed 100 percent at lower percentiles because mortgage debt is carried as a negative share. Source: Federal Reserve Bank of Richmond.3
A household at the 25th to 50th percentile carries real estate equal to 155 percent of net worth; a household in the top 1 percent carries 23 percent.3 Median net wealth was $396,200 for homeowners against $10,400 for renters, and the median homeowner's net housing value was $201,000 in 2022.2,4 For the great middle of the American balance sheet, a cloud on title is not a diminution of wealth. It is wealth.
If title insurance were a claims-paying product, late discovery would be tolerable, the policy would absorb the loss and the parties would move on. It is not that product. Over 2013–2022, title insurers spent approximately 95 percent of collected premium on loss elimination and other operating costs, with claims costs averaging just over 4 percent. For homeowners insurance over the same period, expenses averaged 27 percent and claims 70 percent.1
Title insurance premium allocation, ten-year average 2013–2022. Source: ALTA / Milliman.1 Note: the 95 percent figure covers loss elimination together with other operating costs, including amounts retained by title agents; it is not a published measure of search-and-examination spend alone.
The 2025 statutory results tell the same story in different units. The title industry wrote $18.6 billion of direct premium, issued 13.58 million policies, and recorded a net loss ratio of 4.4 percent against an expense ratio of 98.5 percent and a combined ratio of 102.8 percent. Direct losses paid came to $32 per policy.2 An industry that pays out thirty-two dollars a policy and spends nearly all of its premium preventing claims is telling you plainly where its value is created.
A policy indemnifies. It does not restore a closing date, a rate lock, a 1031 deadline, or a seller's patience. And a meaningful share of loss arises from matters the public record never showed: ALTA's 2025 complexity study reports that nearly 30 percent of losses on purchase transactions and 40 percent on refinances arise from issues not discoverable in the public records, with an average residential claim cost of $143,000 for fraud and forgery in that category.3
Indemnity is what remains after prevention fails. Pre-Policy Intelligence buys the interval in which prevention is still possible.
Average title and settlement charges on a large sample of purchase loans were $2,408, roughly 1 percent of purchase price and about a third of average total mortgage cost of $7,207.4 The CFPB describes title insurance as typically 0.5 to 1 percent of purchase price, paid once.5 Set against $35.8 trillion of equity and a median household whose home is more than its entire net worth, the argument for buying the diligence earlier, not more of it, earlier, is not a close call.
Impersonation fraud is a timing attack. It depends on the seller never being verified against the record until the calendar is too compressed to stop the closing.
ALTA's 2026 seller impersonation study found that 59 percent of title firms experienced at least one attempt in the prior calendar year, against 28 percent in its 2024 survey. Firms reporting an attempt in the single month before the survey rose from 19 to 45 percent, and those reporting three or more attempts rose from 4 to 23 percent. Among firms with attempts, one in four paid a claim, and half of those disclosing average costs reported costs above $100,000.1
Left: FBI IC3 losses in its "Real Estate" crime category, 2023–2025.2 Right: share of title firms reporting at least one seller-impersonation attempt in the prior calendar year, 2024 survey (783 respondents) versus 2026 study (245 respondents).1 The two ALTA samples differ in size and composition.
The federal series moves the same direction. IC3 recorded 12,368 complaints and $275.1 million in losses in its Real Estate category in 2025, up from $173.6 million in 2024 and $145.2 million in 2023. Business email compromise, reported separately, accounted for 24,768 complaints and $3.05 billion.2 These are distinct categories and should never be added together; a closing wire loss may land in either.
What makes impersonation defeatable is that its preconditions are visible in the record before anyone signs. ALTA's respondents identify absentee owners, properties owned free and clear, properties associated with recently deceased owners, sellers who avoid meetings or calls, mail-away signing requests, and seller-chosen notaries as the recurring red flags. Vacant land remains the top target, with vacation homes, rental properties, agricultural land, and primary residences all rising as targets since 2024.1
Every one of those signals except the behavioral ones is a search result. Ownership vesting, absentee mailing address, lien-free status, and a deceased owner of record all surface in the initial Pre-Policy Intelligence response, inside 24 hours of day one, weeks before wire instructions exist to be spoofed.
No federal agency publishes a national count of fraudulent deed filings. That absence is frequently mistaken for evidence that the problem is small. The jurisdiction-level record suggests otherwise. In a 2025 NAR survey, 63 percent of state and local association respondents were aware of title fraud or deed theft in their area in the prior twelve months, 92 percent in the Northeast, and 62 percent of the most recent known instances involved vacant land.1 New York City, which operates a statutory recorded-deed notification program, logged 56 deed-fraud complaints and referrals between November 2022 and October 2023.2
In June 2026 the FBI issued a public service announcement on parcel owner impersonation, describing criminals who "impersonate the owners of vacant property parcels and attempt to illegally sell the property parcels without the owner's knowledge or consent," using fabricated driver's licenses and passports, disposable email addresses, and VoIP numbers.3
IC3 recorded 2,473 Real Estate complaints and $123.7 million in losses from complainants aged 60 and over in 2025, plus 4,566 BEC complaints and $568.0 million in losses in that age band.4 This is the same cohort that holds $14.92 trillion of housing wealth.5 IC3's own 2025 case file includes a Missouri senior who received spoofed wire instructions purportedly from the title company and wired over $1.3 million.4
Firms that assumed federal reporting would supply a fraud-deterrence layer should note where that stands. FinCEN's Residential Real Estate Rule, published at 89 FR 70258 on 29 August 2024, was scheduled to take effect 1 December 2025, postponed by exemptive order to 1 March 2026, took effect on that date, and was then vacated in its entirety on 19 March 2026 by the U.S. District Court for the Eastern District of Texas, which found FinCEN exceeded its authority under the Bank Secrecy Act. FinCEN states that while the order remains in force, reporting persons need not file Real Estate Reports; FinCEN and the Department of Justice have appealed.6
What this means operationally As of the date of this paper, there is no enforceable federal beneficial-ownership reporting requirement for non-financed residential transfers. Private diligence is not supplementing a federal regime right now; in this area it is the regime. Statements about compliance obligations in this space must carry a date, and any party relying on them should re-verify status before acting.
Detection timing is measurable. In ALTA's 2024 study, 46 percent of companies said identifying and preventing fraudulent transactions before closing was at least somewhat common, against 26 percent after closing.7 Ninety-four percent of firms in the 2026 study use multiple fraud-detection tools, averaging 5.3 per firm.8 Tools help. But the cheapest control available is asking who owns the property, on the first day, from the record, and having enough calendar left to act when the answer does not match the person signing.
When a convincing document costs nothing to produce, the document stops being evidence. What survives is provenance: who examined what record, sourced from where, on what date.
FinCEN's November 2024 alert is the clearest official statement of the change. Its analysis of Bank Secrecy Act data found that "criminals have used GenAI to create falsified documents, photographs, and videos to circumvent financial institutions' customer identification and verification and customer due diligence controls," including altered or generated images used for driver's licenses and passports, and GenAI images combined with stolen or wholly fabricated personal information to create synthetic identities. FinCEN observed plainly that "GenAI tools have greatly reduced the resources required to produce high-quality synthetic content."1
The alert publishes no counts, it is qualitative, and this paper will not attach a number to it that FinCEN did not publish. The quantified federal signal comes from IC3, which began breaking out an AI nexus in 2025: adjusted losses exceeding $893 million across complaints with an AI nexus, of which $30.3 million in BEC and $2.70 million in the Real Estate category.2 IC3 itself cautions that these figures understate the problem, noting that overall investment scam losses exceeded $8 billion and that "many victims do not realize the extent AI may be involved."2 Treat the AI-nexus dollars as a floor, not a measure.
Fifty-eight percent of respondents to ALTA's 2026 study rate deepfake image or voice technology as at least somewhat common in seller impersonation attempts, and 87 percent say the same of spoofed contact information.3 In the 2024 study, 43 percent reported fake notary credentials as at least somewhat common, 31 percent reported legitimate notary credentials used without the notary's consent, and 26 percent reported in-person paper notarization in which the fraudster defeated the notary with a fake identification document.4
The notary findings are worth sitting with. The notarial act is the identity control that the entire recording system leans on. When it can be defeated by a generated identification document, the recorded instrument inherits the defect, and the record, which everyone downstream treats as ground truth, now contains a lie with a stamp on it.
We build with AI and we are candid about its limits. A 2026 study reviewing 200 residential title files reported that a public-record-only AI search missed at least one meaningful matter in 40.8 percent of searchable files, with a fail rate above 36 percent in high-risk categories including involuntary liens, and could not search 16 files at all for lack of title plant data or comparable normalized datasets.1 The study is vendor-sponsored and its extrapolated liability figures are explicitly illustrative; we cite the file-level fail rate, not the extrapolation.
The reason is structural rather than technological. Title examination is not a document-retrieval task. It is a judgment about whether a set of instruments, read together and in order, conveys what it appears to convey. ALTA's 2026 complexity study found that 82 percent of purchase transactions require careful review of eleven or more documents, 21 percent require more than fifty, half of respondents rely on nine or more document sources, and 27 percent must obtain documents in person.2 Retrieval breadth is where machines are strong. Reconciling contradictory instruments across sources of varying reliability is where they are not.
In a world where any document can be generated, the question is no longer what a document says. It is who looked, where they looked, and when.
Because the method is the product. A lender relying on a title conclusion is relying on a chain of evidence it cannot independently reconstruct. When fabrication is cheap and retrieval is automated, the differentiator among title providers stops being turn time and becomes the integrity of that chain, whether the search was real, whether the sources were named, whether the date is recent, and whether a person with liability attached to their judgment signed off. Delivering that chain at day one, rather than at day forty-five, is the whole of what Pre-Policy Intelligence offers.
Nine constituencies. One operational change. Each of them is protecting something different with it, a guarantee, a rate lock, a calendar, a fee, a lien position, a file, a home, a structure, or a portfolio.
The order trigger differs because the first moment of commitment differs. An SBA lender commits at the letter of intent; a mortgage banker commits when the borrower signals intent to proceed after the Loan Estimate; a hard money lender commits at the term sheet. In each case the principle holds: order the search at the first moment the outcome starts to matter, not at the last moment before it is required.
7(a), 504, and Express files run against a guarantee timeline with no slack. Title is an eligible project cost, and lien position on real estate is generally evidenced by a title insurance policy under SBA's own standard operating procedure.1 Knowing the vesting, the liens, and the description at LOI protects the authorization calendar rather than testing it.
Conforming lenders, portfolio banks, and credit unions are managing two calendars at once: the rate lock and the post-closing delivery file. With production expenses at $10,936 per loan2 and extensions priced by the day,3 a title surprise inside the lock window is a measurable loss, not an inconvenience.
Speed to close is the product. A lender competing on a seven-to-fourteen-day close cannot absorb a curative item discovered on day ten. Ordering at the term sheet converts title from the gating item into the first item finished, and protects the pricing premium that speed commands.
The business relationship is the asset, and it is spent down every time a borrower is told the closing has moved. Fourteen percent of contracts had delayed settlements in the three months to August 2026.4 Early title gives the originator something to say on day two rather than an apology on day thirty.
A 504 project is a two-lien structure by regulation: at least 10 percent borrower contribution, a CDC debenture of up to 40 percent on a second lien, and a third-party loan on a first lien.1 Two lenders must agree on lien position over one record. Establishing that record at intake keeps the debenture cycle from slipping a month.
Counsel carries the file's legal risk and, in many states, the closing itself. A search delivered at engagement gives counsel the record before advising on structure, contingencies, and seller representations — rather than after the contract has committed the client to terms the record will not support.
One property at a time, to live in, rent, lease, or hold. This is the buyer whose home is more than their entire net worth.2 For them, the value of early title is not cycle time. It is finding out that the seller is not who the deed says, or that a lien exists, while every decision is still ahead of you.
Subject-to, seller-financed, wraparound, and lease-option structures are built on assumptions about the underlying loan, the vesting, and the due-on-sale posture. Those assumptions are record questions. A structure designed before the record is read is a structure designed on hope. The common thread Nine different parties, nine different things being protected, one shared dependency: a factual answer about the record, early enough to act on. None of these constituencies needs a different search. They need the same search, sooner.
Rental, reposition, wholesale, portfolio, and fund buyers underwrite on thin margins and repeat volume. An early title record supports acquisition decisions, holding-cost planning, and confidence in the next offer. At $149 per property, early title is cheaper than a week of carry.
Nine different parties, nine different things being protected, one shared dependency: a factual answer about the record, early enough to act on. They need the same search, sooner.
| Constituency | The order trigger, and what day one protects |
|---|---|
| SBA Lenders | Letter of intent, protects the guarantee timeline and a documented lien position. |
| Institutional Lenders | Loan Estimate and intent to proceed, protects the rate lock and the post-closing delivery calendar. |
| DSCR & Hard Money Lenders | Term sheet, gives the file weeks of clearance runway inside a days-not-weeks window. |
| Mortgage Bankers & Brokers | Loan Estimate and intent to proceed, the file closes on the calendar the borrower was quoted. |
| Certified Development Companies | Project intake, protects the SBA guarantee and the CDC's program standing. |
| Attorneys & Closing Counsel | Engagement, counsel advises from the record, not around it. |
| Real Estate Buyers & Owners | Signed contract, or before it; the search needs no policy and no contract. |
| Creative Finance | The offer, before structure is committed; these are the deals where the search is routinely never run. |
| Real Estate Investors | The offer, before earnest money is at risk, including auction and off-market acquisitions. |
| Decision | Title ordered at conditional approval | Title ordered on day one |
|---|---|---|
| Pricing the deal | Priced before the encumbrances are known; revisited later | Priced against the actual record, including easements and liens |
| Setting the closing date | Promised on assumption, then moved | Promised against a known title clearance list |
| Locking the rate | Locked into a window that title clearance work may exceed | Locked with title clearance work already scoped or complete |
| Verifying the seller | Verified once wire instructions and signing logistics are already in motion | Verified against the vesting early in the transaction |
| Planning next steps | Appraisal, legal, and inspection work is already underway when the title record arrives | The title record informs the next commitment of time and resources |
The right-hand column is not a faster process. It is the same process with the information moved to the front of it.
Policies are underwritten by Stewart Title Guaranty Company, established in 1893 and licensed in all fifty states, and by WFG National Title Insurance Company, established in 1975 and part of Williston Financial Group. The intelligence is ours; the indemnity sits with underwriters whose balance sheets and regulatory standing are a matter of public record.
Title is priced differently depending on the state's rate framework, and the difference is not cosmetic. In promulgated states, the regulator sets the rate and every provider charges it, Texas premiums derive from the Basic Premium Rate table effective 1 March 2026 under Commissioner's Order 2025-9697, with simultaneous issue and reissue treatment governed by Rate Rules R-5, R-6 and R-8.1 In filed-rate states such as Florida, New Mexico, New York and Pennsylvania, rates are filed with or promulgated by the regulator or a rating bureau.2 In negotiable states such as California, rates are filed but competitively set.3 IRIS surfaces the applicable regime and the resulting figure rather than quoting a national average that applies nowhere.
Order the search on the first day. Everything else in this paper follows from that one change.
Every figure in this paper was taken from a primary publication fetched during its preparation and is quoted as published, with its period stated. Federal sources include the Federal Reserve's Financial Accounts of the United States, the Survey of Consumer Finances, FBI IC3 annual reports, FinCEN alerts and rulemaking records, the eCFR, SBA standard operating procedures, and CFPB research. Industry sources are attributed to their publishers, ALTA, NAIC, MBA, NAR, ICE, Freddie Mac and Fannie Mae, with sample sizes and survey dates given where they materially affect interpretation.
Several statistics circulate widely in title marketing that we could not trace to a primary source, and we excluded them rather than repeat them:
No national count of fraudulent deed filings exists. No current NAR breakdown of delay causes exists; the most recent is from 2018. CFPB publishes total loan costs without a title breakout. SBA publishes its own review turnaround times, not application-to-funding averages. Where a number was unavailable we said so rather than substituting an estimate.
Pre-Policy Intelligence℠ and IRIS℠ are service marks of American Commercial Title LLC. Intelligent Title is a brand of American Commercial Title LLC. This paper is for informational purposes and is not legal advice. Original research edition: September 2026. Language and presentation updated October 26, 2026; cited research periods are unchanged.
A paper arguing for sourced, dated intelligence has to practice it. Five widely repeated industry statistics were considered for this paper and refused, including the claim that one in three transactions has a title defect, and the misquoted claim that 95 percent of the premium goes to search and examination. Each refusal, with its reason, appears in the methodology section above and on page 16 of the designed edition.
The sixteen-page edition, cover artwork, commissioned illustrations, and the complete footnote apparatus with live source URLs.
Read & download the PDF Share via email