Earnix AIOS ensures insurers get 'anchored intelligence' for native AI tools & services
Earnix is an AI company for insurance decisioning.
The company claims to offer “production-grade” AI and decisioning technology into the workflows and decisions that shape insurance performance across insurers’ pricing, underwriting, claims, customer engagement, retention etc.
The company has this week launched AIOS, an “AI Orchestration System” for insurance.
AIOS puts AI into areas such as risk evaluation, underwriting, claims and customer engagement with explainability and visibility for the regulatory control.
Insurers are under mounting pressure to act with greater speed and precision as risk conditions accelerate; customer expectations increase, margins tighten, and regulatory scrutiny intensifies.
The software “anchors intelligence” in the insurer’s decisioning layer, rather than isolating it within generic AI tools or in data from core systems.
Insurance-specific context
Operating above and across these environments, integrating insurance-specific context, governance, and human oversight, enabling insurers to scale AI without the costly rip-and-replace of existing systems or change management projects.
“Insurers are entering a new phase of AI, where value will be measured by business performance rather than experimentation,” said Robin Gilthorpe, CEO of Earnix. “The greatest returns will come from AI purpose-built for insurance and applied at the point where decisions determine growth, profitability, risk, and outcomes.”
Gilthorpe says that in a constantly changing market, effective AI strategies are not built around static data or disconnected analysis, he thinks they must be built around “dynamic intelligence” that informs decisions as they are made.
4 billion transactions served
The company points to its ability to handle more than 4 billion transactions (processed annually) and the 25+ AI agents it has already deployed in live insurance workflows.
AIOS its services through an architecture that combines decision orchestration, AI agents, workflow automation, model management, governance controls, and human-in-the-loop review for insurance operations.
“As insurers seek to respond more quickly to changing market conditions, they need capabilities that help connect data, analytics, and business processes across the organisation. Announcements like this reflect the broader trend toward embedding intelligence into core insurance workflows while balancing speed, transparency, and governance,” said Harry Huberty, senior analyst at Celent.
Earnix says that users can orchestrate decisions across teams, personas, workflows, and lines of business in real time, connecting functions that have historically operated independently and enabling faster, more consistent, governed action through certified, insurance-specific apps and agents.
AIOS for developers
AIOS coding plugin is a Claude Code plugin we built to automatically apply Earnix’s coding standards across AI-assisted development sessions. Rather than relying on developers to manually configure Claude for each project, the plugin handles it at the session level so Claude follows team conventions from the moment a session starts.
Setup is a one-time install via a single script (Mac, Linux, or Windows). After that there’s nothing to maintain. Standards update silently in the background, so the whole team stays in sync without any manual steps.
The plugin uses a workflow routing table to determine which skill-sets to load based on what the developer is doing: writing a new feature, reviewing code, authoring tests, and so on. Each skill-set is structured in layers: hard constraints, practical how-to guidance, and reference docs loaded only when needed. Size limits are enforced to keep Claude’s context window lean and focused.
The plugin is designed for distributed team ownership. CI enforces that each MR touches a single skill-set, and CODEOWNERS automatically routes reviews to the right people. Teams across the organization are actively authoring their own rules and standards within the shared structure.
