Allstate Patents a Pricing Engine That Auto-Rejects Risks

Allstate's newly granted US Patent 12,711,555 B2 runs a flagged policy through a second premium model, one built with different modeling techniques than the filed rating plan, and when that alternative price and the standard price diverge past a set threshold, the system rejects the policy and fires the notification itself.

The U.S. Patent and Trademark Office issued the grant on August 18, 2026, with 17 claims covering the mechanism. It is a pricing and risk-selection engine, not a claims chatbot, and it places an automated accept-reject decision squarely inside actuarial rate adequacy.

Key Takeaways

• Claim 1 requires no human review between the score comparison and the rejection notice. The underwriting module is configured to "automatically reject the policy and transmit a rejection notification."
• The score threshold moves by market. Claim 3 sets it from the ratio between the policy premium and the unrestrained premium, and from monitoring underwriting factors within a geographical region.
• The 2024 parent patent claimed the mechanism for a "new homeowner insurance policy." This continuation drops that limitation and claims the same architecture for "a policy."
• State Farm, USAA and Allstate hold 77% of insurer AI patents since 2014, but most of that portfolio sits in claims and service. This one sits on the accept-reject line.

Inside Claim 1: Three Modules and a Threshold

The patent, titled "System and Network for Tiered Optimization," names three Allstate engineers, David MacInnis, Jennifer Jabben and Teresa J. Dalenta, and assigns to Allstate Insurance Company of Northbrook, Illinois (Google Patents). Claim 1 builds the system from three communicating modules. An inspection module retrieves inspection information tied to a physical inspection of the insured asset, a plurality of "rejection conditions," and runs an inspection optimization algorithm that decides whether to waive mitigation of any condition it finds.

When the algorithm determines that a condition requires mitigation to keep the policy in force, the unrestrained premium module generates an "unrestrained rating plan premium" from premium components "using first modeling techniques different from second modeling techniques used to generate the policy premium." The underwriting module then computes a score from that premium and the standard one, compares the score to a threshold, and, on failure, is configured to "automatically reject the policy and transmit a rejection notification to a computing device associated with the policy."

Two comparisons do the work. The inspection module weighs a "lifetime premium with loss" against a "lifetime premium after mitigation," waiving the mitigation requirement outright when the loss scenario is the cheaper of the two. The underwriting module then checks the filed rating-plan premium against the unrestrained premium the second module built independently. Claim 2 confirms both models use "at least one multiplicative factor and at least one additive factor," the standard building blocks of a rating algorithm.

Priority tier	Trigger	System action
Priority 1 (high)	At least one priority-1 rejection condition present	Immediate rejection; rejection letter lists all applicable conditions, no premium comparison run
Indicator-based (medium)	No priority-1 condition present	Underwriting module requests the unrestrained premium, calculates the score, and compares it to the threshold; a score that clears the threshold accepts the policy even with a low-priority condition present
Priority 3 (low)	Score fails to meet the threshold	System checks for priority-3 conditions; if present, rejects and lists them; if absent, rejects anyway based on the medium-priority condition alone

The lineage widened the claim rather than the idea. Application 18/585,776, filed February 23, 2024, is a continuation of application 15/245,343, filed August 24, 2016 and granted in March 2024 as US 11,928,736 B2 under the same title. The parent claimed the mechanism specifically for a "new homeowner insurance policy." This grant drops that limitation from its independent claims.

What "Unrestrained" Means to an Actuary

A standard rating plan is the filed set of factors, relativities and caps a regulator has approved. It is, by construction, restrained by whatever the state's prior-approval or file-and-use process allows. An unrestrained premium, built with a distinct modeling technique, is not bound by those same filed constraints, so it can register a fuller cost indication for a given risk than the filed plan is permitted to charge.

Comparing the two and declining on the gap is a rate-adequacy test automated at the point of underwriting. If the risk the model actually prices sits too far above what the filed plan can legally charge for it, the system declines to write the business rather than accept an inadequate premium. That is a solvency check running policy by policy, in real time, with no rate filing in between.

It runs the opposite direction from the price optimization regulators moved against a decade ago, when several states restricted the use of a policyholder's price elasticity or willingness to pay, factors unrelated to the cost of risk, to set premium above a cost-justified rate. The NAIC's Casualty Actuarial and Statistical Task Force finalized its white paper on price optimization in November 2015, warning that rating plans built around demand elasticity could violate unfair-discrimination standards. Allstate's mechanism tests whether the filed premium is too low, not whether it could be pushed above cost.

It is also a narrower category than most of the sector's AI filings. State Farm, USAA and Allstate together account for 77% of insurer AI patents filed since 2014, with 326, 218 and 136 respectively, and P&C insurers hold 89% of every AI patent the sector has filed in that span (Insurance Journal, citing Evident). Generative filings rose from 4% of insurer AI patents in 2014 to 31% by October 2025, nearly all of it aimed at claims handling and service chat.

The Audit Trail Behind Every Algorithmic No

The NAIC's Model Bulletin on the Use of Artificial Intelligence Systems by Insurers, adopted December 4, 2023, names underwriting explicitly among the lifecycle stages its AI Systems Program requirements cover. It defines an "Adverse Consumer Outcome" as a decision "subject to insurance regulatory standards enforced by the Department that adversely impacts the consumer in a manner that violates those standards," and ties the Department's oversight to the Property and Casualty Model Rating Law's requirement that rates not be "excessive, inadequate, or unfairly discriminatory."

A system whose entire claimed function is producing automated rejections generates exactly that kind of consumer-impacting decision. Insurers are expected to be able to reconstruct why a given policy was declined, what data and thresholds drove it, and whether the outcome was tested for disparate impact.

None of the independent claims requires a person to see the file before the system acts. The wind pool and inspection modules offer conditional paths, accept with conditions or waive mitigation, that soften a straight reject, but the documentation burden rests entirely on records the system itself generates and preserves: the unrestrained premium's inputs, the threshold in effect for that policy's geography, and the rejection condition, if any, that triggered the outcome.

The book this would govern is growing into that requirement. Homeowners new premiums written rose 8.1% to $4.8 billion in the second quarter of 2026, and the line's combined ratio improved to 94.6 from an unprofitable 102 a year earlier, with property-liability underwriting income of $2.0 billion, up 56.7% (Insurance Journal). A carrier that cannot reconstruct those inputs for a market-conduct examiner is exposed regardless of whether the underlying rejection was actuarially sound.