How Cara is pioneering domain-specific AI for business insurance brokerages with AWS
Insurance is an $8 trillion global industry that is struggling with labor flows and a growing talent shortage. Cara brings an AI solution to AWS that automates back office processes for insurance brokerages.
Insurance agents often spend hours on repetitive tasks. This includes completing applications, analyzing policy coverage, rekeying data across systems, and transferring information between customers and carriers. As the industry faces an ongoing talent shortage, brokerages need to increase revenue without increasing headcount.
In this post, we explore how Cara, built in partnership with AWS, addresses these challenges. We walk through the technical design decisions and AWS services that support the solution. We also share the measurable results that Cara has brought to business brokerages.
Challenge: Why generic AI is failing in insurance
Insurance brokerages operate in a highly regulated environment. Everything you do needs to be accurate, legible, and consistent. The data involved includes sensitive personally identifiable information (PII), financial records, and underwriting information.
Generic AI tools are not designed for this complexity. Effective insurance AI must understand domain-specific data models and salesperson workflows. It must also handle carrier-specific requirements and regulatory restrictions while meeting business security standards.
Cara's founding team has seen these gaps firsthand. Vic Yeh, Nikhil Kansal, and Jon Patel previously founded a digital brokerage. They raised it and sold it to The McGowan Companies, one of the largest private insurance companies in the US.
Meanwhile, the team is building an in-house AI copy powered by large-scale linguistic models (LLMs). Flight attendant reduced turnaround times, improved data accuracy, and streamlined agent workflow. Encouraged by strong adoption, they expanded the concept into a stand-alone product: Cara.
Architectural overview
Cara is built on AWS services chosen for reliability, scalability, and security. Figure 1 shows the high-level components of Cara's productivity distribution.
Cara Architecture on AWS
Count and orchestration
Cara uses Amazon Elastic Kubernetes Service (Amazon EKS) for container orchestration across multiple availability zones. EKS manages Cara's microservices, including import pipelines, workflow engines, and the indexing layer.
This architecture supports scalable scaling to handle demand during high refresh and service times. It supports thousands of simultaneous users and workflows per brokerage. It is the responsibility of each organization to work in individual areas to separate tenants.
AI and meaning
Cara's AI capabilities are powered by LLMs hosted on Amazon Bedrock. Amazon Bedrock provides access to underlying models through a fully managed API. This allows Cara to use inference without managing the GPU infrastructure. Cara uses Amazon Bedrock for several basic capabilities:
- Covering and citing intelligence – compares carrier quotes, summarizes coverage differences, and highlights exclusions or gaps.
- Application and default form – cross-fills ACORD and supplemental forms using source documents, pre-submissions, and agency guidelines.
- Proposal creation and renewal – generates branded, customer-friendly proposals and updates spreadsheets.
- Information-driven workflow – references to specific agency references, carrier wishes, and historical placement to guide decisions.
Data security and privacy
Data protection is a basic requirement for insurance companies. Cara architectures use account-specific deployments in AWS. Each brokerage's data and workflows are separated between dedicated, secure workstations. This project supports compliance with industry regulations and provides auditability at the organizational level.
Integration
Cara also includes leading agency management systems (AMS) and customer relationship management (CRM) tools. Synchronizes accounts, policies, and documents to reduce duplicate data entry. AI-driven workflows work directly within consumer technology stacks. This design helps minimize changes to the systems their agents are already using.
Distribution and performance characteristics
One of Cara's design goals is quick time to value. Business brokerages can log in within hours and implement customized workflows within days. Cara deployment in EKS uses parameterized templates for each new tenant. It provides distributed namespaces, storage, and endpoints without manual setup.
In production, Cara's infrastructure on AWS provides:
- High availability – multi-AZ deployment in EKS with automatic failover.
- Elastic Measurement – Kubernetes Horizontal Pod Autoscaler adjusts capacity according to real-time demand. This supports thousands of simultaneous users during peak periods.
- Business security – data segmentation per tenant, encryption at rest and in transit, and integration with AWS Identity and Access Management (AWS IAM).
Measurable results
Cara's AI-driven workflows have delivered measurable results for business insurance brokerages:
| Metric | The result |
| Time saved per user | ~ 10 hours per week on workflow automation and contextual information retrieval |
| Ride speed | Business brokerages boarded within hours; Custom workflows last for days |
| The same dose | Thousands of concurrent users and workflows per brokerage |
| Adoption | Used by hundreds of leading insurance companies and brokerages |
These results come from the automation of some organization and the retrieval of context information. They rely on Cara's unique AI background and the robust, secure infrastructure provided by AWS.
Looking ahead
The insurance industry is still in the early stages of AI adoption. As business demand grows, Cara continues to expand its AI-driven mobility across sales, service, and operations.
“We are excited to push the boundaries of domain-specific AI in real-world insurance use cases with AWS,” said Vic Yeh, CEO of Cara. “Our goal is to help insurance professionals get back to the core of our industry: relationships.”
The conclusion
In this post, we show how Cara built a domain-specific AI solution for insurance sales using Amazon EKS and Amazon Bedrock. The architecture delivers standalone, scalable workspaces. It supports thousands of users simultaneously while meeting the security and compliance requirements of the insurance industry.
To learn more about building AI-enabled applications on AWS, visit the AWS Architecture Center. To get started with Amazon Bedrock, see Getting Started with Amazon Bedrock. For Amazon EKS, see Getting Started with Amazon EKS.
About the writers
