AI Regulation Tracker / Regulatory sandbox
Brazil's SUSEP makes its insurance sandbox permanent, ending fixed application windows
Brazil's insurance regulator dropped the fixed deadline on its regulatory sandbox and now takes proposals continuously, giving insurers and insurtechs a supervised route to test AI-driven products before firm rules exist. Six firms already hold full insurer authorization.
Brazil's insurance regulator has quietly removed one of the biggest constraints on testing new insurance technology. The Superintendencia de Seguros Privados (SUSEP) has made its regulatory sandbox continuous, ending the fixed application windows that governed earlier rounds. In its current third edition, the program accepts proposals on a rolling basis under two stated themes, ecological transformation and technological innovation, and places participants under direct SUSEP supervision rather than the full weight of ordinary insurance rules.
The shift matters because the sandbox has become the practical route through which supervised experimentation, including AI-assisted products, reaches the Brazilian market. Instead of writing detailed AI rules first and permitting products second, SUSEP is letting a controlled set of firms operate under relaxed conditions while it watches how the technology behaves. The regulator describes a program that "no longer had a determined deadline, receiving proposals continuously."
What the regulator confirmed
SUSEP's March 2026 notice reports that roughly 20 companies have been admitted to the sandbox across its three editions. Six of those firms now hold definitive insurer authorization, the full license that lets a company operate outside the sandbox as a regulated insurer, and seven more are in the process of converting. The most recent addition was 88i Seguradora Digital S.A., whose definitive authorization was approved on February 26, 2026 and published on March 4, 2026.
Those numbers describe a program that is doing more than running pilots. The conversion of sandbox entrants into fully authorized insurers is the point of the exercise, and SUSEP is now able to show a pipeline of firms that entered under flexibilized rules and graduated into the regulated market.
The AI angle, and its limits
The sandbox has been associated with AI-driven insurance work, including claims regulation, indemnity payment, fully digital customer journeys, and contracts written in accessible language. According to the sandbox program materials, these are among the documented uses that firms have tested under the relaxed regime. It is worth being precise here. SUSEP's March 2026 authorization notice itself does not name artificial intelligence or describe specific AI systems. It confirms the continuous structure, the themes, and the authorization counts, but the specific AI use cases come from the broader sandbox record rather than from that single announcement.
What the sandbox does is narrow. It is a supervised testing environment, not a blanket permission to deploy AI in insurance. Participants operate under conditions set by SUSEP, on a limited scale, and remain under the regulator's direct monitoring. It does not exempt a firm from oversight, and it does not substitute for the definitive authorization that a company must still obtain to operate as a full insurer. Firms outside the sandbox gain no new latitude to automate claims from these developments.
Why a US reader should care
For US insurers and insurtechs, the contrast with the domestic direction is the story. As US state regulators tighten expectations on insurer use of AI, largely through the National Association of Insurance Commissioners model bulletin on the use of artificial intelligence systems, Brazil is running a different play. Rather than issuing governance expectations and leaving deployment to the market, SUSEP is green-lighting AI-assisted claims and underwriting inside a supervised sandbox, then using what it learns to shape the eventual rules. It is a test-then-rule approach set against the US pattern of rule-then-test.
The practical takeaway for a US firm operating in Brazil is that the sandbox is the supervised on-ramp for AI-driven products, and entering it means accepting close SUSEP monitoring in exchange for relaxed rules. For firms watching from the US market, Brazil offers an early read on how AI claims automation performs under a regulator's eye before binding standards are written, which is the kind of evidence base US regulators have so far lacked.
Frequently Asked Questions
What changed with the SUSEP sandbox?
SUSEP made its regulatory sandbox continuous in its third edition, ending fixed application deadlines and accepting proposals on a rolling basis under themes of ecological transformation and technological innovation, with participants supervised directly by the regulator.
Who does this affect?
Licensed insurers, insurtech startups seeking authorization, and technology vendors building AI-assisted claims, indemnity, and underwriting tools for the Brazilian insurance market.
Does the sandbox let firms deploy AI in claims freely?
No. It is a supervised, limited-scale testing environment under relaxed rules and direct SUSEP monitoring. It does not grant a general exemption, and firms must still obtain definitive insurer authorization to operate fully.
How many firms have graduated from the sandbox?
As of the March 2026 SUSEP notice, about 20 companies have been admitted across three editions, six hold definitive insurer authorization, and seven more are converting. The latest was 88i Seguradora Digital S.A., published March 4, 2026.
Sponsored Training
Browse the full AI Regulation News tracker
Informational analysis for working professionals, not legal advice. Confirm how any rule applies to your situation with qualified counsel.