Claim 1 of US Patent 12,700,042 does not score a roof once. It scores the same property twice, across a defined time window, and moves money on the difference: the method extracts feature vectors for discoloration, tarp presence, material degradation, missing material, sealing and defects from imagery, computes a condition score at each end of the window, and then increases or decreases the insurance premium based on the detected change (USPTO, granted August 4, 2026). A maintenance signal has become a rating trigger.
Cape Analytics, the geospatial property-intelligence vendor Moody’s acquired in January 2025, filed the application on January 29, 2024. The grant published in the US Patent and Trademark Office’s Official Gazette for the week of August 4, 2026, under Cooperative Patent Classification G06Q 40/08, the insurance-business-method subclass (USPTO Official Gazette, week 31). That classification is itself a tell. It separates this grant from Cape’s own prior patent, US 12,694,669, issued one week earlier on July 28, 2026 under a computer-vision classification, G06V 20/176, for a system that reconciles disagreeing property-attribute readings from multiple data sources into one confident value. The July patent decided what a property looks like. The August patent decides what to charge when that appearance changes.
What the claim actually locks in
The independent claim is specific about mechanism, not just outcome. It determines a set of measurements depicting a property, where the measurements include imagery corresponding to a time window; for each measurement, it derives a semantic segmentation mask, isolates property-feature pixels using that mask, and extracts a feature value vector from only those isolated pixels; for each feature vector, it determines a scaling factor indicating the weight of that visual feature in assessing condition, then applies the scaling factor to produce a scaled feature vector; a machine learning model turns the scaled vectors into a set of attribute values; a condition scoring model turns the attribute values into a condition score for each end of the window; and a final step compares the scores to determine a property analysis indicating a change in condition, before increasing or decreasing the premium based on that change and the score itself (USPTO grant, US 12,700,042 B2). The claim enumerates the semantic attributes by name: discoloration, tarp presence, material degradation, missing material, sealing, or defects.
That is a meaningfully different object than Cape’s existing Roof Condition Rating, which is already approved for ratemaking in more than 40 states across more than 300 filed rating and underwriting uses, and licensed by more than 100 carriers (Cape Analytics). Roof Condition Rating, like most vendor condition scores, is a point-in-time read, typically refreshed at renewal from a single current image. The claim in US 12,700,042 requires two or more reads across a time window and prices the delta between them. Nothing in the claim language ties that window to a policy’s annual term. It is silent on cadence entirely, which means the method as claimed could just as easily run monthly, quarterly, or on any schedule a carrier or vendor chooses to operate it against.
Roof-driven losses already justify the attention
The reason a vendor would bother patenting a change detector, rather than resting on a static condition score, is that homeowners losses are disproportionately roof losses. Wind and hail damage accounted for the largest share of homeowners claims from 2018 through 2022, with 2.8% of insured homes filing a wind or hail claim in that five-year window, roughly one in 35 insured homes in any given year (Triple-I). A roof that is already degrading, an already-missing shingle, a tarp left up after a prior claim, is measurably more exposed to the next wind or hail event than a roof in good repair, which is precisely why roof age and roof condition are already standard rating and underwriting variables in most homeowners programs. What those existing variables do not capture is trajectory. A 12-year-old roof photographed in good condition at last year’s renewal and a 12-year-old roof photographed in deteriorating condition today carry the same age band and, on many current rating plans, the same relativity. The condition-change score in this claim is built specifically to close that gap between an annual snapshot and a continuously accumulating maintenance risk.
The mid-term question the claim does not answer
Because the claim is silent on cadence, it leaves open the ratemaking mechanics question a static, renewal-time variable never has to face: does a detected condition change move premium prospectively at the next renewal, where a filed relativity change is transparent and disclosed on the renewal notice, or does it move premium mid-term, against a policy already in force and an exposure basis already earned? States regulate those two paths very differently. New Jersey’s insurance regulations specifically bar mid-term premium increases and reductions in coverage absent prior written approval from the Commissioner, a restriction other states echo in their own cancellation and mid-term-adjustment rules. A carrier that wants a condition-change score to act between renewals, which the claim's structure permits but does not require, is choosing the more heavily supervised path, one where unearned premium, unilateral mid-term revision, and notice requirements all attach in ways an ordinary renewal-time rate change does not trigger.
Discoloration is exactly what regulators already flagged
The sharper problem sits in the claim’s own attribute list. The Connecticut Insurance Department issued a bulletin on March 19, 2024, after a wave of consumer complaints about carriers nonrenewing or cancelling homeowners policies based on aerial imagery showing apparent roof damage. The bulletin drew a specific evidentiary line under Connecticut General Statutes Section 38a-689: mere discoloration, streaking, or other cosmetic issues that do not affect structural integrity or increase loss risk are not valid grounds for a nonrenewal, and if imagery does not definitively show material damage, the carrier must obtain a physical inspection or a licensed contractor’s report before acting. By July 2026, at least 20 states had adopted comparable aerial-imagery underwriting bulletins, several naming roof streaking or discoloration explicitly as insufficient standalone grounds for adverse action (Nearmap regulatory tracker, updated July 14, 2026).
Discoloration is not an incidental example in US 12,700,042. It is one of exactly six named semantic attributes the claim requires the model to extract and score. A rate filing and a nonrenewal notice are legally distinct instruments, and most states review them under separate statutes with separate evidentiary standards. But a premium increase driven by a detected discoloration score functions, from a policyholder’s vantage point, almost identically to the practice the Connecticut bulletin and its 19-plus peer states were written to stop: an adverse financial consequence triggered by a cosmetic imagery signal regulators have already said cannot stand alone. Whether a state treats a filed condition-change rating factor as exempt from those aerial-imagery bulletins, because it is priced and actuarially supported rather than an outright cancellation, or as the same disguised practice wearing a rating-plan label, is an open question this grant does not resolve. It is also the first question a carrier's regulatory and pricing teams need to answer, together, before filing this kind of variable.
None of that forecloses a lower-friction path to the same underlying signal. The claim covers increasing or decreasing a premium, but nothing about the underlying condition-change score requires a carrier to act on it through pricing at all. A carrier could route the same detected-discoloration or new-tarp signal into loss-control outreach instead, a policyholder notice offering a free roof inspection or a maintenance credit for prompt repair, which sidesteps the aerial-imagery bulletins entirely because no adverse pricing or underwriting action occurred. That is a materially different actuarial posture: the condition-change score becomes a loss-mitigation trigger with a claims-frequency payoff downstream rather than a rating variable with a rate-filing burden upstream, and it is available to any carrier licensing this data regardless of which state it operates in.
Disparate impact sits underneath the same feature list
The NAIC’s Model Bulletin on the Use of Artificial Intelligence Systems by Insurers, adopted December 4, 2023, requires a documented governance program that tests AI-driven outputs, including third-party vendor scores, for the potential to produce unfair discrimination against protected classes. As of July 2026, 25 states had adopted the bulletin in full or substantially similar form (Quarles Law Firm). Deferred-maintenance signals, a lingering tarp, spreading discoloration, missing material left unrepaired, correlate with household income and, in some markets, with historically disinvested neighborhoods where deferred maintenance tracks financial constraint rather than a homeowner’s individual risk-taking. That correlation is exactly the pattern consumer advocates and regulators describe as a modern proxy for protected-class status, distinct from but adjacent to the aerial-imagery evidentiary rules above. A carrier filing this score would need disparate-impact testing run against the change-detection output specifically, not only against whatever static condition variable it has already tested, since a delta score built from six named cosmetic and structural features can encode a very different demographic signature than a single point-in-time value.
What Moody’s now owns, and the pattern it extends
Moody’s announced the Cape acquisition on January 13, 2025, terms undisclosed, folding Cape’s geospatial property intelligence into its own catastrophe risk models. "By combining our CAT risk models with CAPE’s AI-powered property risk intelligence, we will provide our customers with the most advanced property risk analytics available in the industry, enhancing insights and decision-making across the insurance lifecycle," said Rob Fauber, Moody’s president and CEO (Moody’s, January 13, 2025). US 12,700,042 is the second Cape-authored patent to publish within eight days, both now Moody’s intellectual property, both built on the same underlying aerial-imagery pipeline: one governs how disagreeing attribute readings get reconciled into a single confident value, the other governs how a change in that value gets priced. Together they extend the aerial-imagery data-moat pattern actuary.info has tracked across EagleView, Cape, and Zesty.ai: the durable advantage is not the underlying image-recognition model, which is now table stakes across vendors, but the proprietary logic layered on top of it, in this case a specific claimed method for turning a detected change into a priced outcome.
Cape and Moody’s are not alone in patenting a continuously updated signal that can move premium outside the annual renewal cycle. State Farm has pursued the same idea from a different data source, patenting IoT and smart-building sensor networks that feed automated risk-mitigation and premium-adjustment logic for commercial property. Where State Farm’s claims run on first-party sensor data the carrier controls directly, Cape’s runs on externally licensed aerial imagery a carrier buys as a vendor feed, a distinction that matters for who owns the underlying model-validation burden when a regulator asks how the score was built.
The classification also matters for how durable this specific grant is likely to prove. G06Q 40/08 covers insurance and finance business methods, a category the Federal Circuit has treated with more skepticism under Section 101 than image-recognition claims filed under G06V, because a claim that merely applies a computer to an abstract business practice, "adjust a price based on new information," is the textbook pattern courts have invalidated since Alice v. CLS Bank. What likely got this claim past that bar is its specificity: it does not claim "price insurance using AI." It claims a particular sequence, semantic segmentation masks, pixel isolation, per-feature scaling factors, a named six-item attribute list, and a scoring model applied at each end of a defined window, that ties the abstract idea of condition-based pricing to a concrete technical implementation. actuary.info’s coverage of the Section 101 reset at the USPTO found the same pattern across other recent insurance AI grants: narrow, implementation-heavy claims are clearing examination where broad "apply machine learning to X" claims are not.
| Patent | Granted | CPC class | What it determines |
|---|---|---|---|
| US 12,694,669 B1 | July 28, 2026 | G06V 20/176 (image recognition) | A combined confidence metric reconciling disagreeing sources into one property attribute value |
| US 12,700,042 B2 | August 4, 2026 | G06Q 40/08 (insurance business methods) | A condition-change score across a time window that directly increases or decreases the premium |
The filing and validation work ahead
A carrier that wants to file this kind of score needs three things a static attribute never demanded. First, a disclosed time window and refresh cadence: since the claim leaves the interval undefined, a filing has to specify it precisely enough for a regulator, and a future market-conduct exam, to reproduce the score from the same imagery inputs. Second, a clean separation between the filed rate factor and any accompanying underwriting-guideline trigger, given the discoloration tension above; a single vendor score cannot simultaneously function as an unremarkable pricing input in a rate filing and as evidence supporting a nonrenewal letter without independently satisfying both bodies of state law. Third, disparate-impact testing run on the change-detection output itself, not inherited from whatever testing a carrier already ran on Cape’s existing static condition variable. That is model-validation and rate-filing consulting work that did not exist as a defined task before August 4, and the more than 100 carriers already licensing Cape’s Roof Condition Rating are the most exposed to it, since a change-detection variant is the natural upsell path Moody’s is now positioned, and patented, to offer them.
The claim gives Moody’s exclusive rights to a specific method for turning a sequence of property photographs into a premium decision. Whether any given state treats that decision as a defensible, filed rate factor or as the same aerial-imagery problem regulators spent 2024 through 2026 writing bulletins against is a filing-by-filing question the grant itself leaves open.
Further Reading
- The AI Patent Race in Insurance: Complete Guide - Hub page tracking carrier and vendor AI patent strategy across the industry.
- Cape Analytics Patents Confidence Scoring for Property Data - The prior Cape grant this patent builds on, one week earlier.
- EagleView Horizon’s Agentic Geospatial Engine Reframes Property Imagery as a Carrier Data Moat - How aerial-imagery vendors are positioning data as a competitive moat.
- State Farm’s Smart-Building IoT Patent for Commercial Property Risk - A first-party sensor approach to the same continuously updated premium signal.
- AI Regulation in Insurance 2026: The NAIC Model Bulletin and State Adoption - The governance framework a filed condition-change score has to satisfy.
Sources
- US Patent 12,700,042 B2, "System and method for property condition analysis" (USPTO, granted August 4, 2026)
- USPTO Official Gazette, week 31 2026
- Moody’s: Moody’s to Acquire CAPE Analytics (January 13, 2025)
- CAPE Analytics: Roof Condition Rating milestone data
- NAIC Model Bulletin on the Use of Artificial Intelligence Systems by Insurers (adopted December 4, 2023)
- Quarles: Nearly Half of States Have Now Adopted NAIC Model Bulletin on Insurers’ Use of AI
- Connecticut Insurance Department: Notice on Aerial Imagery and Underwriting (March 19, 2024)
- Nearmap: Insurance Regulations by State, Aerial Imagery Underwriting (updated July 14, 2026)
- Triple-I: Facts + Statistics, Homeowners and Renters Insurance
- New Jersey Department of Banking and Insurance: Bulletin 01-21