Jedify raises $24M to help companies arm AI agents with context on their business

Jedify, a New York-based startup, has successfully raised $24 million in Series A funding. This investment aims to bolster its platform, which is designed to provide AI agents with critical business context. The funding round was led by Norwest, with significant participation from existing investors S Capital VC and Cerca Partners, new investor Oceans Ventures, and strategic investor Snowflake.

The core of Jedify's offering is its "context graph," an innovative solution that integrates with various enterprise knowledge sources via APIs. These sources can range from structured data in databases and data warehouses to unstructured information in reports, documentation, codebases, and even communication channels like Slack. By mapping the relationships between entities, data, permissions, and workflows, the platform enables AI agents to focus on information directly relevant to specific tasks, thereby enhancing their utility within complex business environments.

The company addresses a critical gap in enterprise AI: while vendors often present AI products as turnkey solutions, their effectiveness is greatly limited without specific business context. Jedify's technology ensures that AI agents understand company-specific terminology, operational assumptions, and access protocols, preventing issues such as an agent inadvertently exposing sensitive financial data. This capability is particularly crucial for mid-market and large enterprises with extensive and diverse data ecosystems.

Assaf Henkin, co-founder and CEO of Jedify, highlights how their multi-dimensional context graph differs from traditional semantic layers or knowledge graphs. Unlike older systems, Jedify's platform is model-agnostic and updates in real time, capturing dynamic relationships across various data points and organizational structures. This allows AI agents to achieve true autonomy and make informed decisions by accessing comprehensive, current information, significantly outperforming simpler semantic layers in capability.