Insurance AI for insurers and MGAs: which tool for pricing, underwriting or claims

Which tool for which job?

Your job	Documented example	The reference to read	What it documents
Underwrite medical stop-loss for self-funded employers	Gradient AI	Evolution Risk Partners	Evolution Risk Partners, an MGU, uses Gradient AI's SAIL to assess group risk where claims data is thin and to speed underwriting decisions on hard cases. 4
Analyze group health risk for brokerage clients	Gradient AI	The Provant Group	The Provant Group, a national brokerage, uses SAIL to quantify group health risk and tailor policies for each client. 5
Assess workers' comp claims and underwriting risk	Gradient AI	The Builders Group	The Builders Group uses Gradient AI so adjusters can assess claims earlier and underwriters can evaluate workers' compensation risk. 6
Analyze commercial applications at intake	Sixfold	AXIS	AXIS integrated Sixfold into its automated clearance process after a dashboard pilot and analyzed over 15,000 applications in the first month. 7
Require AI input on every underwriting quote	Sixfold	Generali	Generali's global head of operations and IT says the company moved from testing Sixfold to requiring its input for every quote. 8
Review disaster-surge claims estimates and photos	Shift Technology	Tokio Marine & Nichido Fire Insurance	Tokio Marine & Nichido Fire introduced Shift's solution to handle claim surges during disasters, standardizing review of estimates and photos. 9
Deploy pricing models under governance guardrails	Akur8	Hellas Direct	Hellas Direct's pricing team goes from insight to deployment with Akur8 Deploy, under governance guardrails and without an engineering conversation each time. 10

What does the public record show about insurance AI tools?

Six tools recur among insurers, MGAs and brokers. Akur8 is actuarial software for pricing, reserving and life modelling. hyperexponential sells an underwriting and pricing platform for commercial and specialty lines. Gradient AI predicts risk for underwriting and claims. Shift Technology covers claims fraud, subrogation and coverage decisions. Indico Data automates submission intake. Sixfold reads underwriting documents and scores risk with cited signals. The named examples below are a scoped selection, not a map of the market.

As of 2026-11-03, usedby shows Akur8 for 103 of the 367 Insurance & InsurTech companies it tracks, hyperexponential for 19, Gradient AI for 16, Shift Technology for 12 and Indico Data for 10. These are public customer lines, many of them logos on a vendor's own pages. They are not market share and say nothing about which tool performs better. Written task detail exists for only a handful of the names.

811121314151617

Which tools are documented for pricing and actuarial work?

Akur8 has three lines in this set that go beyond a name. Portage Mutual is backed by a press release on AI-driven pricing modernization across Canada. Hellas Direct is quoted on Akur8 Deploy: its pricing team goes from insight to deployment under governance guardrails, without an engineering conversation each time. Cypress's VP of Product Management describes running pricing and profitability models for several lines of business and locations. Those lines tell you the job is ratemaking. They do not tell you how many models, which lines or what changed.

RB Jones is the most detailed pricing story in the set. As a start-up MGA with no trading history, it routes underwriters from its policy admin system into hyperexponential at the point of quoting, pulls external data such as Moody's climate modelling through APIs, and captures broker submissions, including declined ones, through a data ingestion agent. The project began in October 2025 and five pricing and underwriting models were live across three underwriting cells at launch. Its chief actuary wanted her own team to own and update the models. Here pricing and underwriting are one workflow.

One caveat for anyone on Guidewire. hyperexponential's site states that Guidewire has signed an agreement to acquire it, and Akur8's site announces a Guidewire Marketplace accelerator for its Deploy rating engine. Both vendors therefore sit near the same platform. We cannot say what either change means for a given customer, so ask each vendor about integration and roadmap before committing.

110111218

How do underwriters use Sixfold, Gradient AI and Indico Data?

AXIS describes the clearest Sixfold workflow. Underwriters had been doing heavy manual research and classifying each applicant's business. The pilot began as an underwriter-facing dashboard, then moved into AXIS's automated clearance process through an API, using one internal IT resource and about six weeks including testing. AXIS analyzed over 15,000 applications in the first month. The case study states the industry-code accuracy figure two different ways (over 75% and over 90%), so we omit it. Generali's operations head says Sixfold's input is now required for every quote.

Gradient AI's documented cases sit in medical stop-loss, group health and workers' compensation, mostly at MGUs and brokers. Evolution Risk Partners uses its SAIL solution where group-specific data is thin and says it speeds decisions on hard cases. The Provant Group, a brokerage, uses SAIL to analyze group health risk for clients. Bardon's chief underwriting officer says the solution lets it price more confidently, after Bardon evaluated competing tools. All three are vendor-published press releases from 2022 to 2025, so treat the benefits as the companies' own statements.

Indico Data's homepage describes ingesting emails, loss runs and statements of values, screening submissions against appetite and routing work through APIs. Allstate, Markel, HDI and others appear as OBSERVED customers, but the sources reviewed tie no task to any of them, and the homepage testimonials are not attributed to named insurers. For Indico, the intake step is a described capability, not a documented customer workflow here.

2457815

Which tools show documented claims work?

Assurant has worked with Shift Technology on claims fraud since 2018, inside its Special Investigations Unit. After a trial where Shift identified dozens of fraud cases, Assurant selected it. Shift customized the investigation software and refined models for specialty vehicles, extreme weather and multi-claim cases. The case study reports a higher case acceptance rate since 2018, without giving a figure. It was published in December 2022.

Generali France used Shift's Force to raise detection of fraudulent property claims, covering more than 100,000 claims a year, with fewer false positives, per a 2020 case study. Tokio Marine & Nichido Fire adopted Shift's solution for disaster claim surges, standardizing review of estimates and photos. The Builders Group, a workers' compensation fund, uses Gradient AI so adjusters can assess claims earlier. Its integration with the Origami Risk platform was still being implemented when the release was written.

36919

What do the US peers show, and at what proof level?

Three US names have CONFIRMED status with a written workflow. RB Jones uses hyperexponential for pricing and underwriting. Assurant uses Shift Technology for claims fraud investigation. Bardon Insurance Group uses Gradient AI's SAIL for medical stop-loss underwriting and pricing.

The other five US peers are OBSERVED, meaning a vendor shows them as a customer, with at most a quote or a case-study title. Cypress has a pricing-oriented Akur8 quote. Guardian has a Sixfold-site quote from its underwriting transformation head about a 50% cut in review time, a vendor-presented figure. Redline Underwriting appears under a Federato case-study title about securing Lloyd's coverage. Allstate (Indico Data) and Newfront (Adaptional) show adoption only. None of this shows a preference among US insurers.

123810152021

What should a smaller insurer check before copying a workflow?

Editorial recommendation, not a vendor claim: compare your starting point with the documented one. RB Jones was a greenfield MGA with no legacy data. AXIS needed one IT person for an API link. Assurant's fraud results came after years of model refinement with its investigators. Check which system of record the tool must plug into (policy admin, Origami Risk, Guidewire), who will own the models afterwards, and who reviews outputs before a quote or a claim decision. No verified price appears in our material, so pricing is unknown.

13710