Artificial intelligence vendors often market their enterprise tools as ready-to-deploy solutions, but the reality is that AI agents rarely function effectively from day one. Without investing time to train a model on your specific business operations, it won’t grasp how your company defines key terms like "revenue" or understand which employees have access to particular files. This challenge helps explain why many AI firms now send engineers to assist with integrating their products into client systems. Jedify, a startup based in New York, is directly addressing this issue. The company’s platform connects to enterprise knowledge sources through APIs, building a "context graph" of the business that AI agents can leverage to perform better. These sources include databases, data warehouses and lakes, SaaS applications, BI tools, and unstructured data like reports, documentation, code repositories, Slack channels, and meeting recordings.

The funding round included participation from existing investors S Capital VC and Cerca Partners, as well as new backer Oceans Ventures. Snowflake, the data giant, also joined as a strategic investor and is incorporating Jedify’s technology into its AI offerings, such as Cortex AI, Semantic Views, and CoWork. Jedify’s core argument is that for AI agents to be truly useful within enterprises, they need access to relationships between entities, data, permissions, domain knowledge, workflows, operational assumptions, and company-specific terminology. This context allows an AI agent to focus on information relevant to a particular task, rather than searching across everything the company possesses.

Co-founder and CEO Assaf Henkin (pictured above, far right) highlighted Kiteworks, a compliance firm, as a case study. Kiteworks connected Snowflake, Tableau, Notion, and internal playbooks—including documents and screenshots—to Jedify, then built agentic tools for various customer workflows. "They wanted to arm their sellers and account teams with a sophisticated app—think of it as both a dashboard and a real-time conversational application. When they enter a customer conversation, Jedify builds everything they need to know on the fly. During the conversation, they can get very specific details surfaced proactively in real time," Henkin said.

Henkin argues that Jedify’s context graph differs from existing semantic layers, metadata catalogs, and knowledge graphs because it is multi-dimensional, capturing relationships across entities, data, people, permissions, and customers. It is also model-agnostic and updates in real time as information flows into and out of connected systems. "When you want an agentic solution to be truly autonomous, driving decisions across CRM data, Zendesk tickets, or real-time telemetry data, a context graph offers far better capabilities than a semantic layer," he explained.

Permissions present a clear challenge—an intern should not access the CFO’s revenue projections, for example. Henkin said his platform addresses this by inheriting permissions from identity systems, file systems, SaaS tools, and databases, including row-, column-, and table-level access rules. Customers can then create additional groups to define what agents or workflows can access. The platform also includes observability and governance tools to ensure AI agents behave as intended.

Jedify currently targets mid-market and large enterprise customers with mature data stacks and multiple databases or data warehouses. Henkin noted the company has between 10 and 20 early customers, including The Weather Company, and is seeing interest from data-heavy sectors like gaming, industrials, and consumer packaged goods. Snowflake’s investment and partnership are significant because large data platforms are also building similar capabilities to help customers use AI more effectively with their data. However, Henkin argues that Jedify complements these efforts, as much of a company’s data—and most of its institutional knowledge—is not typically stored with a single cloud provider.

"[Large data companies] will say, 'Oh yeah, just bring everything.' But in reality, companies have multiple databases, warehouses, and data solutions. The big thing is that not all of your data is in those environments, and most of your knowledge is not there, so it’s a disadvantage for them," Henkin said. He also noted that for companies attempting this independently, training an AI model to build a comparable context layer can be prohibitively expensive, especially as organizations scrutinize and limit their AI token usage. Rapid advances in AI model development support Jedify’s broader bet: as models become more capable and interchangeable, proprietary context that enhances their performance within businesses will become a valuable and durable advantage. The startup will use the fresh funding for product development, hiring, and go-to-market efforts, bringing its total funding to approximately $33 million.