hyperexponential

Graded on the Clearwater precedent, which covers a long established platform carrying a genuine embedded artificial intelligence line. Applying the removal test, hx Renew survives intact: the pricing itself is performed by models the carrier's own actuaries write in Python, and no model learns the price. What the company is bought for, rating complex specialty risk, would continue to work with the machine learning stripped out.

The artificial intelligence line is real and shipped rather than a roadmap, which is what separates this from a rejection: a data ingestion library, a virtual actuarial assistant, Portfolio Intelligence, and Triage, which reads submissions and prioritises against appetite with agentic reasoning. That layer sits at ingestion, triage and authoring assistance around a rating core that predates it. The company has repositioned publicly as an agentic underwriting platform, and the grade should be revisited if the agentic products become the reason carriers buy.

The company states that audit trails and governed logic make agent led underwriting explainable, and that agents are called as governed tools inside every workflow. Both are assertions that constraints exist. Neither describes what the constraint is. No confidence threshold, escalation trigger, review queue, approval gate or sampling control on automated output is named anywhere located, and the word governed is doing the work that a mechanism should be doing.

The audit trail is a reconstruction capability after the fact rather than a control in the path. That places this on the middle rung of the index ladder, where a vendor asserts gates exist without describing them. Federato holds the top of this axis by publishing three named automation modes with a random audit on the fully automated tier.

The change control is the strongest in the index and the reason it stops short of the top grade is a distinction worth carrying forward. Every decision, every rule change and every model update is captured end to end, framed explicitly as a path for regulatory review, and the carrier's own actuaries author and own the models rather than receiving them as a black box. Customers hold structured access to all data, configuration and code at any time.

A model risk function could reconstruct how any given price was reached and what changed since the last version, which is what supervisory model risk guidance asks for and what almost nothing else here delivers. What holds it at B is whose models are transparent. The auditability covers the models the customer wrote.

The artificial intelligence layer that hyperexponential itself supplies, described as foundation models, agent frameworks and inference governed for insurance, carries no published validation evidence, no accuracy figure and no account of how the agentic triage reasoning is tested. A platform can make its customers' models fully auditable while disclosing nothing about its own, and that is what has happened here.

Named carriers across three market shapes: Aviva, Markel, Sompo, Convex, Conduit Re, HDI and the specialty managing general agent Novacore. More than 45 billion dollars of gross written premium is contracted through the platform annually. Quantified outcomes are published: quote turnaround reduced by up to 50 percent, an average loss ratio improvement of one to two percentage points across clients, and pricing models deployed up to ten times faster.

The company also reports growing sales tenfold since its 2021 Series A while remaining profitable, which is a harder claim to make loosely than a customer count. Graded on the Rogo pattern, where named tier one institutions combined with published aggregate outcomes carried the top grade. What would strengthen it further is a per client figure attached to a named carrier, which is not published.

Aviva, Markel, Sompo, Convex, Conduit Re and HDI compete with one another and all price risk on the same platform, and pricing models are the most closely held intellectual property a specialty carrier owns. Whether one carrier's models, submission flow or pricing outcomes inform anything served to another is unstated.

The portability commitment, giving customers structured access to their own data, configuration and code, is a genuine stewardship positive and is credited on the integration axis, but it governs a customer's access to its own assets rather than the boundary between customers. This is the third insurance platform assessed in this session with an overlapping client list, and only one of the three states the boundary in public.

A trust centre is operated and no data processing agreement, subprocessor list, retention schedule or lawful basis statement was located in public material. Exposure on this axis is genuinely lower than for most of the index because commercial property and casualty submissions concern businesses, their assets and their loss history rather than individuals, so consumer financial privacy regimes bear lightly on the processing path. The material handled is nonetheless commercially sensitive to a degree that would ordinarily attract a published handling statement, and none exists outside the gated trust centre.

Both of the certifications enterprise insurance buyers actually ask for are held and named by the vendor on its own product pages: a service organisation control type two report and ISO 27001 in its 2022 revision. A dedicated trust centre operates at its own subdomain on a continuously monitored platform.

That combination clears all four parts of the index credential test and places hyperexponential in the strongest fifth of the index on an axis where most vendors sit at C. Worth recording for internal consistency: this is a stronger security position than Cytora, built the same day, which publishes signed certificate documents but names no service organisation control report.

hyperexponential is a technology supplier to carriers and holds no insurance licence, which is the correct posture and carries no penalty. The grade turns on a presentation issue that the index credential test is built to catch.

The company displays a single badge row reading service organisation control type two, ISO 27001, Prudential Regulation Authority, National Association of Insurance Commissioners, Bermuda Monetary Authority and Lloyd's, and states elsewhere that the platform is designed for those regimes. Two of those six are audited certifications. The other four are supervisory bodies that have certified, approved and admitted nothing.

Designed for is a statement of build intent, not a credential, and placing four supervisor names beside two real certifications invites a buyer to read approvals that do not exist. Setting the non credentials aside, as the standing rule requires, leaves no formal admission, no supervised regulator test and no programme enrolment, which is a C. The underlying claim is not false and the regimes named are genuinely those of its buyers, which is why this is a presentation finding rather than a integrity one.

Nothing addressing fairness, differential outcomes or bias testing was located. Commercial property and casualty specialty lines weaken the protected class analysis that drives this axis in consumer credit and in life and health underwriting, since the insured is usually a business, so the exposure here is lower than for most of the index.

It is not absent: appetite based triage decides which submissions receive attention first, and a prioritisation model trained on historical writing patterns will reproduce whatever the book has historically favoured, including by geography and industry class. No account of that risk is published.

No error rate, remediation commitment, liability position or correction path is published for the artificial intelligence layer. The audit trail gives a carrier the ability to reconstruct after the fact how a price or a triage decision was reached, which is more than most vendors here offer and is credited on the model risk axis, but reconstruction is not recourse.

The party carrying the consequence of a mis prioritised submission is the broker whose client waits, and the insured whose risk is deprioritised on an appetite judgement neither of them sees, and neither has any stated route to challenge it.

The platform layer is described as foundation models, agent frameworks and inference, governed for insurance. That names the class and identifies no supplier, model family, version or country of processing. The disclosure gap is narrower than usual in one respect, because the pricing models themselves are written by the carrier's own actuaries and are therefore fully known to the buyer.

It is the ingestion, triage and assistant layer where third party models sit, and where a buyer cannot determine whether broker submission content, loss runs and financials reach an external provider.

The integration posture is explicit and the partner list is named. The platform runs inside underwriting workbenches, policy administration systems, broker portals, application programming interfaces and scheduled jobs, and the company states plainly that the carrier's policy administration system remains the system of record.

Named integration partnerships exist with Send for its underwriting workbench, with Cytora for risk digitization, and with Supercede, and the company maintains pre built integrations and code snippets so customers can plug into alternative data sources and algorithms at the point of pricing. The unusual addition is portability: customers retain structured access to all data, configuration and code at any time, stated as a commitment to transparency and operational continuity. Very little in this index offers an exit path in writing.

No region list, residency commitment, single tenant option or self hosted path was located, and the platform serves customers across the United Kingdom, Europe, Bermuda and the United States, several of which impose insurance data requirements a buyer would need answered. The counterweight, and it is a real one, is the exit position: customers hold structured access to all data, configuration and code at any time, stated as a commitment to portability and operational continuity. That answers the question of whether a carrier can leave, which most of this index does not, while leaving unanswered the question of where the data sits while it stays.

No pricing, tier structure, billing basis or indicative range is published, and the route to a number is a demo request. This is the index norm and is measured against Sumsub, which publishes per verification rates on a public page.

Outcome claims are stated with precision while cost is absent: an average loss ratio improvement of one to two percentage points and a halving of quote turnaround are published with nothing indicating what the platform costs, so a carrier cannot construct the return calculation the outcome figures are plainly designed to support.

Depth in one segment rather than breadth across several. Commercial property and casualty is served thoroughly, including specialty and excess and surplus lines, reinsurance through a dedicated reinsurance pricing model, and managing general agents across admitted, non admitted and reinsurance business on one foundation. Geographic reach spans the United Kingdom, Europe, Bermuda and the United States. Held off the top grade because life and health is absent entirely, and coverage on this axis is measured against vendors serving both halves of the insurance market.