Darwinbox launches Claude connector for HR tasks

Integrating HR context with AI

Darwinbox has announced a new connector that brings its human capital management (HCM) platform into the Claude AI assistant. This integration allows employees, managers, and HR teams to perform specific work tasks directly within a conversation. The goal is to combine Claude's intelligence with the specific data regarding an organization's people and policies.

The platform provides Claude with secure, permission-aware access to more than 280 tools. These tools cover the entire employee lifecycle, from the initial hiring process to daily administrative tasks. By providing this context, the system aims to help AI provide more accurate answers based on how a specific company actually operates.

According to Gulf Today, the launch builds on Darwinbox Cortex, which is an AI-native HCM platform. This new connector is now available in the Claude Marketplace for organizations to use. It serves as a bridge between the AI's reasoning capabilities and the actual data stored in a company's HR systems.

Automating the employee lifecycle

The connector enables a wide range of everyday actions that usually require navigating separate software modules. Employees can use Claude to apply for leave, regularize their attendance, request shift changes, or file expense reports. These actions happen in the same conversation where the user might first ask a question about company policy.

For recruiting teams, the integration provides a dedicated workspace for finding and managing talent. Recruiters can use the tool to shortlist candidates, schedule interviews, and record feedback. It also supports high-volume tasks such as sending pre-offers in bulk or initiating the onboarding process for new hires.

The system includes specialist agents designed for specific roles, such as an Employee Support Agent and a Recruiter Agent. These agents feature an interactive user interface powered by A2UI, which allows for visual interaction within the Claude environment. This setup helps move HR work from manual data entry to guided, conversational workflows.

Data security and permissions

A major component of the Darwinbox connector is its adherence to existing organizational security protocols. Every action performed through Claude runs as the signed-in user, meaning the AI respects the specific permissions assigned to that individual. Users can only see and perform actions that their specific role allows within the company structure.

The system is designed to ensure that no changes are made to the HR records without explicit user oversight. Claude is programmed to ask the user for confirmation before any update is finalized in the Darwinbox system. This creates a governed environment where the AI acts as an intermediary rather than an autonomous actor with unrestricted access.

This approach addresses the common concern of how to provide AI with sensitive organizational data safely. By using the permissions and approvals already established in the HR platform, the company ensures that the AI stays within the boundaries of corporate policy. It allows for the use of rich context without compromising the privacy of the employees.

The Context Graph and MCP

To provide deeper insights, the connector utilizes the Darwinbox Context Graph. This is described as a live map that shows how people, roles, policies, and workflows connect across a company. When a user asks a complex question, Claude can reference this graph to provide answers grounded in the organization's specific structure.

The technology is built on the Model Context Protocol (MCP), a standard for connecting AI models to data sources. Darwinbox became the first major HCM platform to launch its own MCP server in May 2025. This allows the platform to connect to various workplace apps, including Slack and Microsoft Copilot, in addition to Claude.

By using these protocols, the company aims to make AI interoperable across different enterprise tools. This means the HR data can be accessed by the AI wherever the work is already happening. It helps create a unified experience where the AI understands the organizational