USPTO ASAP Pilot Gives Insurance Patent Filers a New Edge

From tracking USPTO filings across the top 20 insurance carriers over the past 18 months, patterns in prosecution timing and claim amendment cycles have become predictable enough to model. The Artificial Intelligence Search Automated Pilot Program, or ASAP, disrupts that predictability. Launched in October 2025 and now extended through June 1, 2026, with the petition fee waived and the intake target doubled to 3,200 applications, ASAP delivers an AI-generated prior art search before an examiner ever reviews the filing. For insurance carriers racing to protect AI innovations in underwriting, claims processing, and pricing, that early visibility into the prior art landscape could reshape filing strategy, prosecution budgets, and the entire timeline from application to allowance.

This matters for insurance specifically because of how concentrated the filing activity is. Three carriers, State Farm, USAA, and Allstate, hold 77% of all insurer AI patents filed since 2014, according to Evident’s Insurance AI Patent Tracker. When a handful of filers dominate a technology class, their claims overlap substantially, and early prior art visibility from ASAP could either accelerate prosecution by enabling preemptive narrowing or delay it by surfacing uncomfortable overlaps that force strategic pivots. The law firm alerts covering the ASAP extension have explained the mechanics. None have analyzed how those mechanics interact with the specific filing patterns, claim architectures, and Section 101 vulnerabilities that define the insurance AI patent landscape.

How the ASAP pilot works

The ASAP pilot subjects eligible patent applications to an automated, AI-driven prior art search before any human examiner begins substantive review. The USPTO’s internal AI tool uses the Cooperative Patent Classification (CPC) designation assigned to the application, together with the specification, claims, and abstract, as contextual inputs. It searches across three databases: U.S. patents, U.S. pre-grant publications, and the Foreign Image and Text (FIT) database of foreign patent documents.

The output is an Automated Search Results Notice, or ASRN. Each ASRN lists up to ten documents ranked in descending order of AI-determined relevance, along with a search string that applicants can enter into the USPTO’s Patent Public Search (PPUBS) tool to explore adjacent references. The ASRN is generated after pre-examination processing and petition approval but before the examiner begins examination.

Several operational details matter for insurance filers:

• No response required. The ASRN is not considered a notification under 35 U.S.C. § 132. Applicants face no mandatory reply obligation, no shortened statutory period, and no presumption that cited references must be addressed.
• File wrapper inclusion. The ASRN becomes part of the official file wrapper upon publication of the application. This means competitors monitoring published applications will see which prior art the AI tool flagged, creating a new information asymmetry (or symmetry, depending on perspective) in the prosecution landscape.
• Examiner discretion preserved. Examiners consider ASRN documents in the same manner as other documents in Office search files. They are not bound by the AI ranking and may conduct independent searches that surface entirely different references.
• Patent face treatment. References from the ASRN appear on the face of an issued patent only if the examiner independently cites them on a PTO-892 form or the applicant formally submits them through a compliant Information Disclosure Statement (IDS).
• Track One compatibility. Applications enrolled in Track One prioritized examination may simultaneously participate in ASAP without affecting their prioritized status.

Eligibility is limited to original, noncontinuing, nonprovisional utility applications filed under 35 U.S.C. § 111(a) in DOCX format. Continuing applications, PCT national-stage entries, reissue applications, plant applications, and design applications are excluded. That exclusion matters for insurance portfolios built on continuation chains, where much of the patent value lives in child applications that refine parent claims. Those continuations cannot enroll in ASAP, even when the underlying technology is identical.

The expansion: doubled targets, waived fees, extended window

The USPTO announced the ASAP extension on April 16, 2026, making three material changes to the original pilot parameters:

Parameter	Original (Oct 2025)	Expanded (Apr 2026)
Program target	~1,600 applications (200 per TC)	≥3,200 applications (400 per TC)
Petition fee	$450 per 37 C.F.R. § 1.17(f)	Waived for filings on/after March 23, 2026
Petition window	Through April 20, 2026	Through June 1, 2026
Closure trigger	200 per Technology Center	400 per Technology Center

The timing of the extension is revealing. As of April 16, 2026, only 169 petitions had been filed across all Technology Centers, with just 76 granted. That is 4.8% of the expanded target and roughly 10.6% of the original 1,600-application goal. The gap between ambition and uptake drove both the fee waiver and the deadline extension. Director Squires has framed ASAP as “the first of many planned AI pilots in patent examination,” and a pilot that fails to attract sufficient participation cannot generate the data needed to justify the next phase.

The fee waiver eliminates the $450 cost barrier, but as Nixon Peabody noted in its April analysis, attorney time spent reviewing ASRN references and evaluating whether to file preliminary amendments represents a real cost even at zero petition fee. For solo practitioners filing a single application, that overhead may not justify participation. For institutional filers with dozens of pending applications, the economics look different.

Why participation has lagged, and what it signals

The 169-petition count deserves scrutiny. The USPTO receives roughly 650,000 utility patent applications per year, so ASAP’s target of 3,200 represents less than 0.5% of annual volume. Even this modest target has proven difficult to reach.

Several factors explain the gap. First, the original October 2025 launch required applicants to file the petition on the same day as the application, a logistical constraint that excluded applicants who learned about the program after their filing date. Second, the $450 fee, while modest relative to total prosecution costs, added friction to a program whose benefits were unproven. Third, and most consequentially, early participants reported that the AI tool’s results were mixed in quality. Anonymous practitioner feedback cited in IPWatchdog described references that were tangentially relevant or structurally dissimilar to the claims, undermining confidence in the tool’s practical utility.

For the USPTO, low participation is a data problem, not just an optics problem. The Office needs statistically meaningful results across Technology Centers to evaluate whether AI-assisted pre-examination search improves examination quality, reduces pendency, and yields better prior art identification than examiner-only workflows. Without sufficient enrollment, the pilot cannot answer those questions, and the planned follow-on pilots lack evidentiary support.

For insurance filers specifically, the participation gap creates an information vacuum. If few insurance-relevant applications enroll in ASAP during the remaining window through June 1, the pilot’s findings will be shaped disproportionately by other technology sectors. The insurance patent community would then inherit an AI search framework calibrated on semiconductor, pharmaceutical, or software-general applications rather than on the pricing engine, fraud detection, and underwriting automation claims that define insurance AI.

How early prior art visibility changes prosecution strategy

The strategic value of ASAP lies not in the ASRN itself, which is advisory and non-binding, but in the timing of the information it delivers. Ordinarily, an applicant files a patent application and waits 12 to 24 months (sometimes longer) before receiving a first office action that reveals which prior art the examiner found most problematic. During that gap, the applicant has no visibility into the examiner’s search results, no ability to refine claims preemptively, and no signal about whether the prosecution path will be smooth or contested.

ASAP compresses that information gap. By delivering the ASRN before examination begins, the program enables several tactical moves that were previously unavailable or impractical:

Preliminary amendments before examination. An applicant who receives an ASRN identifying close prior art can file a preliminary amendment narrowing claims before the examiner writes the first office action. If the narrowed claims clear the references the examiner would have cited, the first action may be an allowance rather than a rejection. That eliminates one or more rounds of office actions, each of which costs $3,000 to $8,000 in attorney time for a typical insurance AI prosecution. For carriers filing 10 to 20 applications annually, the cumulative savings could reach six figures.

Informed foreign filing decisions. Most patent applicants must decide whether to pursue foreign filings within 12 months of the U.S. priority date, long before the first office action. Early ASRN results allow applicants to assess whether the prior art landscape supports broad international claims or whether the investment in foreign prosecution, often $50,000 to $100,000 per jurisdiction, is likely to produce narrow claims of limited commercial value. Insurance carriers with global operations, such as AIG and Chubb, face this decision on every new filing.

Portfolio-level resource allocation. When a patent team receives ASRNs across a batch of filings, patterns emerge. If the AI tool consistently flags the same cluster of prior art references across multiple applications in underwriting automation, that signals a crowded field where marginal filings may not produce defensible claims. Resources can shift toward less saturated technology areas, such as agentic AI coordination or parametric trigger validation, where the prior art landscape is thinner and the claim space is more open.

Competitive intelligence on publication. Because the ASRN becomes part of the file wrapper at publication (typically 18 months after filing), competitors can eventually see which prior art the AI flagged for a rival’s application. This creates a new channel for competitive monitoring. A carrier like USAA, which has filed 218 AI patents, could systematically track what references the ASAP tool identifies against State Farm’s or Allstate’s applications, mapping the overlap space and identifying freedom-to-operate gaps.

Version 2 and the claims-search upgrade

Director Squires revealed at the University of San Diego Patent Law Conference that the current ASAP AI tool searches the specification only, not the claims. This is a material limitation. In patent prosecution, claims define the scope of protection, and prior art relevance is assessed against claims, not the specification. A search that processes the specification and abstract as contextual input but does not directly analyze claim language may miss references that are highly relevant to the actual scope of the application while surfacing references relevant to disclosed but unclaimed embodiments.

An upcoming Version 2 of the tool, which Squires described as forthcoming, will search claims directly. For insurance AI applications, where claim language often recites specific data processing sequences (loss development factor computation, credibility weighting adjustments, telematics score generation), a claims-based search should produce substantially more targeted results. The gap between Version 1 and Version 2 may partially explain the practitioner dissatisfaction that contributed to low participation: references flagged based on specification context alone may have felt tangentially relevant because they did not address the particular steps recited in the claims.

The transition to Version 2 has practical implications for timing. Applicants filing in the remaining ASAP window through June 1, 2026 will receive Version 1 ASRNs. If Version 2 launches in a subsequent pilot or becomes a permanent feature later in 2026 or early 2027, those applicants will not retroactively receive updated search results. The strategic question for insurance filers is whether the Version 1 ASRN, with its specification-only search, delivers enough actionable signal to justify enrollment, or whether waiting for a claims-based tool is the better play.

The Section 101 gap: what ASAP does not fix

ASAP addresses prior art under 35 U.S.C. §§ 102 and 103, the novelty and obviousness requirements. It does not address the abstract-idea barrier under § 101 that has historically been the primary obstacle for insurance AI patent claims. This distinction is critical because the two rejection types require entirely different remedial strategies.

A § 102/103 rejection means the examiner found prior art that anticipates or renders obvious the claimed invention. The remedy is straightforward: narrow the claims to distinguish over the cited references. ASAP accelerates this by surfacing likely references earlier. A § 101 rejection means the examiner concluded that the claims are directed to an abstract idea, such as a method of organizing human activity or a mathematical concept, without an inventive concept that transforms the claim into patent-eligible subject matter. No amount of prior art visibility helps with that determination.

For insurance AI applications, § 101 remains the dominant rejection ground. Claims directed to “using machine learning to predict claim severity” or “applying a neural network to optimize premium pricing” routinely draw abstract-idea rejections, regardless of how novel or nonobvious the specific implementation is. The § 101 landscape shifted materially in late 2025 with the Ex parte Desjardins precedential decision, the Kim Memo on mental process limitations, and the Subject Matter Eligibility Declaration (SMED) procedure. Those developments created new pathways for overcoming § 101 rejections, but they operate independently of ASAP.

The practical implication is that ASAP is most valuable for insurance AI applications that have already cleared, or are likely to clear, the § 101 hurdle. Applications with technically grounded claims (specific model architectures, defined data transformation pipelines, concrete hardware-software integrations) will benefit most from early prior art visibility, because their prosecution bottleneck is distinguishing over prior art, not surviving an abstract-idea challenge. Applications with broadly functional claims (“a system for automating underwriting decisions using AI”) face a § 101 wall before prior art even becomes relevant, and ASAP does nothing to help scale that wall.

Impact on insurance carrier patent portfolios

The ASAP pilot intersects with insurance carrier filing activity at a moment when the concentration of AI patent ownership is at its historical peak. According to Evident’s Insurance AI Patent Tracker, State Farm has filed 326 AI-related patents since 2014, USAA has filed 218, and Allstate has filed 136. Together, those three carriers account for 77% of all insurer AI patents. The remaining 23% is split among roughly 27 other carriers and a growing cohort of vendors and insurtechs.

This concentration has several consequences for ASAP’s impact:

Overlap density. When three filers dominate a technology class, their claims inevitably overlap. State Farm and USAA both file extensively in telematics-based risk scoring. Allstate and State Farm both file in claims automation. The prior art most likely to appear in an ASRN for a new State Farm application is an existing USAA or Allstate patent, and vice versa. ASAP could surface these intra-carrier conflicts earlier, forcing filers to decide whether to narrow around a competitor’s existing claims or pursue a different claim strategy entirely.

Filing velocity and volume economics. State Farm’s 326 AI patents represent roughly 30 to 40 new filings per year. At that volume, even a modest prosecution efficiency gain per application, one fewer office action, one less round of examiner interviews, compounds to meaningful budget savings. The zero-fee ASAP pilot is essentially a free option on prosecution efficiency for high-volume filers. For a carrier filing two or three AI applications per year, the overhead of reviewing ASRNs may not justify the effort.

Generative AI and agentic claim architectures. Generative AI patents surged from 4% to 31% of insurer AI filings between 2014 and late 2025, with agentic AI patents representing the newest and least-tested claim category. USAA, in particular, has filed agentic AI patents with multi-agent coordination and feedback loop architectures that are structurally different from traditional ML claims. For these newer filing types, the prior art landscape is genuinely sparse, and an ASRN that confirms thin prior art coverage could embolden filers to pursue broader claims than they otherwise would.

Vendor and insurtech exposure. EXL, with its 10-patent AI portfolio spanning document extraction, knowledge graphs, and a domain-specific insurance LLM, represents the services-company filing model. Vendors like EXL file for different strategic reasons than carriers: their patents are often defensive, designed to protect platform architectures from infringement claims by competitors. For vendors, ASAP’s early prior art signal is particularly valuable because it clarifies whether a filing will produce a broad platform patent or a narrow method claim, a distinction that directly affects licensing negotiation leverage.

Duty of candor and the ASRN disclosure question

The ASRN creates a nuanced duty-of-candor consideration that law firm alerts have flagged but not fully analyzed for institutional filers. Under 37 C.F.R. § 1.56, patent applicants and their counsel have a duty to disclose to the Office all information known to be material to patentability. The question is whether references surfaced in an ASRN trigger a disclosure obligation beyond what the applicant would have independently.

The USPTO has stated that the ASRN does not require a response and is not a § 132 notification. Applicants are not required to file an IDS citing ASRN references. However, as Nixon Peabody has noted, counsel who review an ASRN must still assess each reference’s materiality under existing disclosure obligations. If a reference in the ASRN is material to patentability and counsel is aware of it, the duty of candor arguably requires disclosure regardless of whether the ASRN formally compels it.

For insurance carriers with large patent teams managing dozens of active prosecutions, this creates an administrative question: should ASRN references be routinely submitted via IDS as a defensive measure, even when no submission is required? The conservative approach, submitting everything, adds prosecution cost but eliminates inequitable conduct risk. The aggressive approach, submitting nothing unless independently material, saves cost but creates litigation exposure if a competitor later argues that known ASRN references should have been disclosed.

Patterns we have observed in insurance patent prosecution suggest that most major carrier patent teams will default to the conservative approach, particularly for high-value filings in underwriting and pricing. The cost of an IDS filing is trivial compared to the cost of defending an inequitable conduct challenge during enforcement litigation.

Director Squires and the broader AI examination roadmap

ASAP is not an isolated experiment. Director Squires, who took office as the 60th USPTO Director on September 22, 2025, has positioned the program as the opening move in a comprehensive strategy to integrate AI into every phase of patent examination.

The roadmap, drawn from Squires’s public remarks and subsequent reporting, includes several planned deployments:

• Generative AI tools for examiners (July 2026). Examining attorneys are expected to receive access to an internal generative AI tool after all guardrails, rules, and policies are in place. This tool will assist with claim construction, invention summaries, and claim-to-citation mappings. Unlike ASAP, which is applicant-facing, these tools will operate within examiner workflows and directly influence how rejections and allowances are formulated.
• Narrow pilot projects (12-month horizon). The patent side of the USPTO plans to conduct examiner-designed pilot projects in isolated environments covering claim construction, invention summaries, claim-to-citation mappings, and eligibility analyses. The eligibility pilot is the most consequential for insurance AI, because it could introduce AI-assisted § 101 analysis into the examination workflow.
• Permanent search integration. If ASAP generates positive results, the pre-examination AI search could become a permanent feature of the application processing pipeline, applied to all utility applications rather than a self-selected subset. That would fundamentally alter the information environment for every patent filer, not just pilot participants.

Squires has stated that the USPTO is “open for business, not only for the technologies of today, but especially for those of tomorrow.” That message, combined with the Ex parte Desjardins precedential decision and the SMED procedure, signals a directional shift toward facilitating patent protection for AI innovations rather than gate-keeping against them. For insurance carriers whose AI investments are accelerating, from Allstate’s ALLIE agentic AI platform to Travelers’s 10,000-seat Anthropic deployment, the patent office environment is becoming more receptive at the same time that the technology itself is becoming more patentable.

Filing strategy for the remaining ASAP window

The ASAP pilot accepts petitions through June 1, 2026, or until each Technology Center reaches 400 accepted applications, whichever comes first. Given the current pace of 169 total petitions filed as of mid-April, the June 1 deadline is far more likely to bind than the per-TC capacity limit. That leaves approximately 12 days from publication of this analysis for eligible filers to decide whether to participate.

For insurance carriers and vendors weighing enrollment, several factors favor participation:

• Zero marginal cost. The fee waiver eliminates the $450 per-application cost. The only incremental expense is counsel time reviewing the ASRN, which is modest for applications that would require a prior art review anyway.
• Free option value. If the ASRN surfaces useful references, the filer gains prosecution leverage. If the references are irrelevant, the filer has lost nothing except the time spent reviewing them. The asymmetric payoff favors enrollment.
• Shaping the tool’s development. Participation generates training data for the USPTO’s AI models. Insurance-specific applications that enroll in ASAP contribute to the tool’s calibration in insurance-relevant technology classes. Non-participation means the tool evolves based on other sectors’s filing patterns.
• Track One stacking. Applications already enrolled in Track One can simultaneously participate in ASAP, combining expedited examination with early prior art visibility. For time-sensitive filings, such as claims tied to a specific product launch or regulatory deadline, the combination is particularly attractive.

Factors that argue against participation include the Version 1 limitation (specification-only search, not claims-based), the uncertain quality of results reported by early participants, and the administrative overhead of managing ASRN review workflows. For filers with a small number of pending applications, the overhead may not justify the effort.

Why this matters for actuarial practice

The ASAP pilot sits at the intersection of two trends that actuaries are increasingly asked to evaluate: the value of intangible assets in insurance M&A due diligence, and the governance frameworks required for AI systems deployed in regulated insurance markets.

Patent portfolio value is an input to M&A pricing. When a carrier acquires a company with AI patents, the quality and enforceability of those patents affect the acquisition premium. ASAP changes the information available during prosecution, which in turn affects the breadth and defensibility of issued claims, which ultimately flows through to portfolio value. Actuaries involved in acquisition modeling or intangible asset valuation will increasingly encounter patent portfolio assessments where ASAP participation, or non-participation, is a variable.

On the governance side, the USPTO’s pivot toward AI-assisted examination creates a parallel to the NAIC’s own AI evaluation pilot for rate filing review. Both regulators and patent examiners are deploying AI tools to evaluate insurance-related submissions, and both are grappling with the same fundamental questions: how much to trust AI-generated outputs, how to maintain human oversight, and how to calibrate the tools over time. The convergence is not coincidental. It reflects a broader institutional recognition that AI-intensive industries require AI-assisted regulatory frameworks.

For the insurance AI patent landscape specifically, ASAP is a transitional tool. It provides a preview of a future where machine-assisted prosecution is the default, not the exception. The carriers and vendors that engage with this preview, learning its limitations and shaping its development through participation, will be better positioned when AI search becomes permanent. Those that sit out the pilot risk facing a permanent search framework designed without their input.