Earnix, a Boston-headquartered insurtech, has launched Agent Hub, bringing more than 25 insurance-specific AI agents and apps into its AI Orchestration System (AIOS).
The new catalogue brings agentic AI directly into Earnix's pricing and rating, underwriting, and customer engagement solutions. The company positions the technology as a way for insurers to move from AI that informs decisions towards AI that can act within the workflows behind them.
The launch comes as insurers face increasingly volatile risk and market conditions, shortening the useful life of individual decisions and making growth, profitability, and portfolio performance harder to manage. As AI increases the speed and scope of what insurers can analyse and recommend, Earnix argues that the next step is enabling intelligence to operate within the business processes where those decisions are made, while maintaining the governance and accountability required for high-stakes insurance activity.
Agent Hub is designed to operate within insurers' existing technology environments, including policy administration systems, data platforms, underwriting workbenches, and customer portals. The agents work within defined permissions and traceability controls, with human oversight retained for decisions requiring judgement and accountability. Earnix said this allows insurers to use specialised agents across high-value workflows to accelerate execution, improve consistency and auditability, increase practitioner capacity, and support faster, better-informed decisions.
The Agent Hub catalogue spans pricing, underwriting, modelling, customer engagement, data, and technology, with 14 agents being demonstrated within Earnix solutions at its Excelerate London event. Examples include Model Feature Mapper, which connects model features to the appropriate data variables to help pricing and actuarial teams work with trusted inputs, while improving transparency and auditability around model-driven decisions. Product Expert Advisor provides real-time answers to product questions using approved product information and expertise, helping users access information faster while reducing reliance on manual escalation to domain experts. Premium Explainer, designed for customer-facing use, provides personalised explanations of premiums based on an individual's policy, reducing the need for customers to search for answers or escalate questions to specialists.
The agents sit within Earnix AIOS, which brings together AI, data, models, business rules, workflows, governance, and human expertise across the company's insurance decisioning solutions. Earnix said AIOS is designed to orchestrate these capabilities within insurers' existing technology environments, rather than requiring them to replace the systems they already use. This allows agents to draw on the business context needed for specific insurance decisions while remaining subject to defined controls. The approach reflects the increasing importance of context as insurers move towards agentic AI, with agents needing access to relevant data, models, and business rules if they are to move beyond simply generating recommendations.
Be'eri Mart, chief product and technology officer at Earnix, said context was key to unlocking the value of agentic AI.
"Agentic AI becomes much more powerful when it can work with the data, models, and business context relevant to the task. The opportunity is not simply to automate a task, but to keep the information current as risk, customer behavior, and market conditions change. That is how insurers become more agile without losing control," said Mart.
The expansion also puts greater emphasis on governance as AI moves closer to decisions affecting pricing, underwriting, customer outcomes, and profitability. Earnix said Agent Hub incorporates permissions, traceability, and human oversight so that insurers can retain control as agents take on more work within operational processes.
Meredith Barnes-Cook, senior principal at Datos Insights, said agentic AI was entering a more consequential phase.
"Agentic AI is entering a more consequential phase for insurance. The question is no longer whether insurers can build agents, but whether they can deploy them into the decisions that affect growth, profitability, risk, and customer outcomes—within the guardrails that regulation and governance demand," said Barnes-Cook. "That raises the bar considerably. Insurers will need to think as carefully about authority, accountability, and governance as they do about the intelligence itself. The companies that master that balance—where capability and control move together—will turn agentic AI into real business value."
