Federato

Federato describes itself as AI native and there is real model work in submission triage, risk scoring, accumulation analysis and agentic quote generation. The platform's own account of its value, though, is consolidation: it bundles analytics, scoring and workflow that underwriters previously stitched together from several vendors into one workbench spanning submission to invoice.

Apply the removal test and a unified underwriting workflow with portfolio visibility survives, which is a substantial product in its own right. That places this alongside the workflow platforms rather than the model native vendors.

This is the most explicit published autonomy design in the index and it should be the reference point for the axis. Federato names three distinct modes rather than gesturing at human in the loop: the system drafts the best triaged deals and explains its reasoning while the underwriter adjusts for nuance and owns the relationship; a defined subset of deals that clearly fit strategy is fully executed by the AI under predefined rules and guardrails; and those automated decisions are randomly audited by underwriters for quality control. Naming which decisions are automated, what constrains them and what the sampling control is, in public, is exactly what almost every other vendor leaves implicit.

Two mechanisms give this substance. Automated decisions are randomly sampled and audited by underwriters, which is an ongoing monitoring control rather than a one off validation, and drafted quotes come with their reasoning exposed so a reviewer can interrogate an individual output.

There is also an unusual external validation loop: the platform produces audit ready evidence of appetite adherence and referral activity that carriers present to reinsurers, so a capital provider with its own money at risk is scrutinising the discipline the system enforces. Formal model documentation, evaluation results and a validation summary are still absent.

Customers are named at the top of the market and their executives speak on the record, including the chief underwriting officer of a global carrier's North American arm describing the platform as the unified core system for the entire policy lifecycle from submission through quote, bind and issue, and the president of a coastal property specialist on wind exposed risk.

Outcome claims are specific and falsifiable, citing hit ratio improvements of nine to twenty four percent and implementation in eight to twelve weeks. The company also publishes original research surveying 500 underwriters and executives across carriers, managing general agents and aggregators rather than recycling industry commentary.

Safety is handled through constraint rather than assertion. Automated execution is bounded by predefined rules and guardrails, drafted quotes carry an explanation of the reasoning behind them, and a governance layer gives carriers visibility over how underwriting operations are actually behaving.

The unaddressed question is competitive rather than technical, and it is specific to this category: risk selection is where carriers compete, so whether one carrier's submissions, pricing and loss experience inform models serving another carrier on the same platform is the central stewardship issue here, and nothing public states the boundary.

The data profile is lighter than most of this index because commercial underwriting works largely with business, property and exposure information rather than bulk consumer personal data, so the consumer financial privacy surface is genuinely smaller and that should count in the vendor's favour. It is not zero: submissions carry named individuals, loss histories and broker relationships, and a data infrastructure product implies aggregation across sources. No published privacy framework, retention schedule or subprocessor disclosure was located in this pass.

This pass surfaced no trust centre, certification list, attestation scope or audit period on the public site. Named customers include a global carrier running the platform as its unified core system across the policy lifecycle, and an insurer of that size would require attestations before placing a vendor in that position, so the published record is very likely behind the actual control environment. The grade records what an outside buyer can verify without entering diligence and should be revisited if a trust surface is published or located.

Federato supplies technology and holds no insurance licence, the expected posture. Its products sit unusually close to regulated artifacts, since insurance is supervised at state level in its principal market and rates and forms are filed instruments, while the platform includes a forms engine and rating and quoting components that operate on exactly those filed materials.

Leadership has publicly identified the regulatory and compliance environment as a central concern and positions the governance layer as the answer. What is not published is how filed rate and form adherence is maintained or evidenced when quotes are generated automatically.

Insurance has moved faster than most financial sectors on governing AI use, and this is the gap that shows. State insurance regulators have adopted a model bulletin setting expectations for insurers using artificial intelligence, several states impose testing duties around unfair discrimination, and the obligation lands on the carrier, which makes vendor documentation a practical necessity for the buyer.

Nothing public addresses that framework, describes testing for unfair discrimination in triage or pricing, or explains what evidence a carrier could hand a regulator asking how the automated decisions were validated. Commercial lines carry a weaker protected class analysis than personal lines, which narrows the exposure without removing it.

Random underwriter audit of automatically executed deals is a genuine accountability mechanism, because it creates a sampled record of whether the automation was right rather than assuming it, and explained reasoning on drafted quotes lets a reviewer challenge an individual output.

What is absent is any vendor side commitment: no accuracy guarantee, no remediation term where an automated bind proves mispriced, and no described route for a broker or insured whose submission was declined by the triage layer.

No public material identifies the model providers behind the triage, scoring and agentic quoting layers, names external data sources feeding exposure and accumulation analysis, or lists subprocessors. The omission has a specific edge in this category: carriers price risk on data, so which third party exposure, catastrophe or property datasets inform a recommendation is a question underwriters would ordinarily ask before trusting a number, and nothing public answers it.

The architectural position is stated plainly and it is the right one: Federato sits on top of the policy administration core and replaces the underwriter desktop rather than attempting to displace the system of record. Native interface integrations reach the three dominant policy administration platforms in this market, which is what makes that position workable at a carrier that cannot rip out its core.

The platform also spans submission intake, pricing, quoting, documentation and referral workflow with a partner portal for brokers, so it covers the full path from submission to invoice, and it does so on an eight to twelve week implementation rather than a multi year replacement programme.

Delivery is cloud hosted software as a service with stated expansion into the United Kingdom, Europe, Latin America and Asia Pacific, which brings several distinct data protection and insurance supervision regimes into scope. No public material identifies hosting regions, residency options, tenancy separation, transfer mechanisms or subprocessors. Tenancy separation deserves particular attention here given that competing carriers underwrite the same risks on the same platform.

No pricing, seat cost or tier structure is published. Two things partially offset that and are worth crediting. A self guided product tour is available without booking a sales meeting, so a buyer can see the software before entering a process.

And the published implementation window of eight to twelve weeks is a genuine commercial disclosure in a market where core underwriting deployments routinely run for years, since implementation cost and risk often exceed licence cost in insurance software.

Within property and casualty the coverage is broad and well differentiated, reaching carriers, managing general agents and underwriters, aggregators, mutuals and reinsurers, with published depth in genuinely hard lines including aviation, commercial trucking, political violence, wildfire and hurricane exposed property, which are the markets where risk selection actually decides solvency. Geographic expansion into the United Kingdom, Europe, Latin America and Asia Pacific is stated. The boundary is the line of business: this is commercial property and casualty underwriting, with nothing addressing life, health, banking, lending or payments.