Milliman v. Gradient AI: Six-Patent Actuarial AI Dispute Settled in Massachusetts
Milliman Inc., Milliman Solutions, and Vigilytics jointly sued Gradient AI Corp. and two individual defendants over six US patents protecting group risk scoring software. Filed in May 2021 before Judge Nathaniel Gorton in Massachusetts, the case settled in February 2024 — roughly 1,000 days after filing — without costs and without prejudice.
Multi-patent actuarial AI dispute settles after nearly three years in Boston federal court
Filed on 25 May 2021 in the US District Court for the District of Massachusetts (Case No. 1:21-cv-10865), this action was brought by Milliman Inc. and two affiliated entities — Milliman Solutions, LLC and Vigilytics, LLC — against insurtech AI firm Gradient AI Corp. and two individual defendants, Stanford A. Smith and Samuel Chase Pettus. The plaintiffs asserted infringement of six US patents, all directed at actuarial group risk scoring software used in insurance underwriting and predictive analytics. The case was assigned to Chief Judge Nathaniel M. Gorton.
On 12 February 2024, the court was advised that the parties had reached a settlement. Judge Gorton entered an order dismissing the action without costs and without prejudice, with a 45-day window for any party to reopen proceedings if the settlement agreement was not ultimately consummated. The ‘without prejudice’ designation means the dismissal does not constitute an adjudication on the merits — either party could theoretically refile if the settlement collapsed within that window.
At roughly 1,000 days from filing to settlement notification, the case ran longer than many patent disputes that resolve early, yet avoided full trial. The involvement of two named individual defendants alongside the corporate entity — suggesting possible trade secret or employment-related dimensions beyond the patent counts — may have complicated and ultimately accelerated settlement calculus. The specific financial terms and any licensing arrangements remain confidential and are not reflected in the public court record.
Filing to filing in 994 days
Days from filing to settlement (May 2021 – Feb 2024)
Full party and counsel information
| Role | Name | Type | Detail |
|---|---|---|---|
| Plaintiff | Milliman, Inc. | Company | Actuarial analytics firm and IP affiliates — holders of six group risk scoring patentsSearch in Eureka ↗ |
| Defendant | Gradient A.I., Corp. | Company | Gradient AI Corp. — insurtech startup applying AI to commercial insurance underwritingSearch in Eureka ↗ |
| Plaintiff counsel | Christopher Centurelli | Attorney | Counsel for Milliman, Inc.Search in Eureka ↗ |
| Defendant counsel | Paul D. Popeo | Attorney | Counsel for Gradient A.I., Corp.Search in Eureka ↗ |
| Presiding judge | Judge Nathaniel M. Gorton | Chief Judge | Massachusetts District Court — Chief JudgeSearch in Eureka ↗ |
Stipulation of dismissal — official text
The court’s order — dismissing ‘without costs and without prejudice’ on notification of settlement — is a standard conditional settlement dismissal. It creates no legal precedent on infringement or validity of the six asserted patents. The 45-day reopening clause is a procedural safeguard, not an indication of dispute; it lapses automatically if the settlement closes. Both parties exit without any public admission of liability, and Milliman’s patents survive fully intact and enforceable against third parties.
Six Milliman Patents — Actuarial Group Risk Scoring Software
The six patents asserted in this case — US9665685, US10109375, US9323892, US10886012, US9118641, and US9965651 — collectively cover methods and systems for generating group-level risk scores, a core function in commercial insurance underwriting and actuarial analytics. The application lineage spans from 2009 (US12/827745, the likely foundational filing) through to 2018 (US16/131094), suggesting a deliberate continuation prosecution strategy designed to extend claim coverage as the underlying technology and competitive landscape evolved. The patents sit at the intersection of statistical modeling, data processing, and predictive analytics applied to insurance risk.
For the insurtech and actuarial software sector, this portfolio represents a significant IP barrier. Milliman — one of the world’s largest actuarial consulting firms — appears to have systematically patented the algorithmic underpinnings of group risk scoring rather than relying solely on trade secrecy. The breadth of six distinct patent numbers across multiple application families means that a competitor designing around one grant may inadvertently infringe another. Gradient AI, which applies machine learning to commercial insurance underwriting, sits squarely in the commercial territory these patents appear to protect.
Should your AI underwriting platform run an FTO against Milliman’s risk-scoring patents?
Any company building software that generates group-level risk scores for commercial insurance — whether for health, workers’ compensation, or employer stop-loss products — should treat this six-patent family as a priority FTO target. The patents cover methods, not just specific implementations, which means feature-level similarity may be sufficient to trigger infringement analysis regardless of underlying architecture. The fact that Milliman actively litigated these patents against a well-funded insurtech startup signals genuine enforcement intent.
PatSnap Eureka’s FTO Search Agent can map your product’s technical feature set against the independent and dependent claims of all six Milliman patents simultaneously, flagging overlap risk across the full family in a fraction of the time required for manual review. Claim monitoring alerts will notify your team if Milliman files continuation applications that expand claim scope into adjacent technical territory — a critical early-warning capability given the portfolio’s multi-decade prosecution history.
Run a freedom-to-operate analysis on US9665685B1 to assess your product’s exposure
Run FTO in Eureka →Similar patent cases in actuarial AI, insurance software, and risk scoring
PatSnap Eureka tracks related litigation across truck body equipment, vehicle accessories, and comparable infringement actions in the Georgia district system.
What this case signals for the insurance AI and actuarial software IP landscape
Milliman’s six-patent assertion against an insurtech AI startup sets a clear enforcement precedent in predictive risk scoring.
Milliman is actively enforcing its actuarial AI patent portfolio in federal court
This action confirms that Milliman and its IP affiliates are willing to pursue costly federal litigation — including naming individual employees — to protect group risk scoring technology. Insurtech firms building on similar actuarial AI methodologies should treat Milliman’s portfolio as an active enforcement risk, not a dormant filing strategy.
Six-patent suits raise the cost of litigation for both sides — settlements become rational quickly
Multi-patent assertions in technically complex domains like actuarial AI dramatically increase discovery costs, claim-construction hearings, and expert fees. The ~1,000-day timeline here, without trial, is consistent with a case that settled once both sides had conducted sufficient discovery to assess relative risk. Early FTO analysis can prevent reaching that costly inflection point.
Milliman v Gradient — key questions answered
Milliman and its affiliates asserted six US patents: US9665685, US10109375, US9323892, US10886012, US9118641, and US9965651. All relate to group risk scoring software used in actuarial and insurance underwriting applications. The case was filed in the District of Massachusetts in May 2021.
The case settled. On 12 February 2024, the court was notified of settlement and entered an order dismissing the action without costs and without prejudice, with a 45-day window to reopen if settlement was not consummated. No judgment on patent infringement or validity was entered.
A dismissal without prejudice means the court made no ruling on the merits of infringement or patent validity. Milliman’s six asserted patents remain valid and fully enforceable against third parties. Milliman retains the right to assert them in future actions against other defendants.
Stanford A. Smith and Samuel Chase Pettus were named alongside Gradient AI Corp. While the public record does not specify the exact basis for individual liability, naming individuals in patent suits alongside a corporate defendant is consistent with claims of inducement of infringement, and may also signal potential trade secret or employment-related allegations. The specific causes of action against individuals are not confirmed in the available case data.
Milliman and its affiliates were represented by K&L Gates LLP (Massachusetts), with Christopher Centurelli listed as plaintiff’s agent. Gradient AI and the individual defendants were represented by Choate, Hall & Stewart LLP, with Paul D. Popeo listed as defense agent.
PatSnap Eureka searches patents and litigation data to answer instantly.