Jedify融资2400万美元，为AI代理构建企业实时上下文图谱

做的是什么生意

Jedify, founded in 2023 by CEO Assaf Henkin, CTO Adi Elimelech and CPO Erik Shani, connects an enterprise's knowledge sources through APIs and assembles a context graph of the business: definitions, entity relationships, permissions, workflows, operational assumptions and company-specific terminology. Sources include databases, data warehouses and lakes, SaaS apps, BI tools, reports, documentation, code bases, Slack channels and meeting recordings. AI agents use the graph to narrow their attention to relevant data instead of searching everything. The platform inherits access rules from identity, file, SaaS and database systems, including row-, column- and table-level permissions, updates in real time as data flows, is model-agnostic, and includes observability and governance tooling. The company is New York-based with 35 employees and targets mid-market and large enterprises with mature data stacks.

启动资金：Series A of $24M (2026-06-10) led by Norwest, with S Capital VC and Cerca Partners returning, Oceans Ventures new, and Snowflake as strategic investor; brings total funding to about $33M after an $8.5M seed in September 2023

起因

Jedify was founded in 2023 by Assaf Henkin, Adi Elimelech and Erik Shani, who watched enterprise AI deployments stall beyond prototypes because models lacked reliable business context at runtime: how the company defines revenue, who may see which file, which workflows and assumptions apply. The founders' answer was to treat context as infrastructure, not a model trick, and they raised an $8.5 million seed in September 2023 to build it.

经过

By 2026 Jedify had a platform connecting structured sources such as warehouses, CRMs, financial systems and BI tools with unstructured ones such as documents, Slack and meeting recordings, and it had 10-20 customers including The Weather Company and Kiteworks. On 10 June 2026 TechCrunch exclusively reported a $24 million Series A led by Norwest; Norwest partner Assaf Harel joined the board. Snowflake invested strategically and is integrating Jedify with Cortex AI, Semantic Views and CoWork, an endorsement that cuts both ways, since Snowflake and other data platforms are also trying to build context capabilities themselves.

结果

As of early September 2026 Jedify is live with early customers and using the Series A for product development, hiring and go-to-market. The bet is that enterprises will standardize on a neutral context layer between their data systems and whichever models they use, rather than letting a single AI or data vendor own that position. The open question is whether context graphs become a durable infrastructure category or a feature that the large platforms absorb, and whether 10-20 lighthouse customers can convert into the mid-market and enterprise pipeline Jedify is targeting.

可借鉴之处

当模型可互换时，就拥有让它们在客户业务中发挥作用的那一层，并保持中立，以免被单一AI供应商夺走。