Jedify raises $24m Series A to build context graphs for AI agents
Jedify, a New York-based startup building context-graph infrastructure for enterprise AI agents, has closed a $24 million Series A led by Norwest, with a strategic investment from Snowflake Ventures and participation from S Capital VC, Cerca Partners and Oceans Ventures. Combined with an $8.5 million seed round in September 2023, total funding now stands at just over $33 million. Norwest partner Assaf Harel will join the company's board.
The proceeds will fund accelerated product development, go-to-market expansion and headcount growth. Jedify's core proposition is that the runtime context problem — not model capability — is the primary reason enterprise AI agents stall between prototype and production.
The context problem
Co-founder and chief executive Assaf Henkin describes the challenge bluntly: "Enterprise data is fragmented across systems, definitions, permissions, and workflows. Jedify turns that fragmented knowledge into a live context graph that agents can use to produce accurate, cost-efficient, business-ready answers."
The company argues that large language models alone cannot determine which definition of revenue applies to a query, which customer record is canonical, or which operational assumptions are in scope — unless that context is surfaced at inference time. Without it, agents either hallucinate from insufficient context or waste tokens by processing irrelevant data. Jedify's platform autonomously constructs a customer-specific context graph using its patent-pending Semantic Fusion technology, connecting structured sources — data warehouses, CRMs, financial systems, BI tools — with unstructured knowledge including documents, Slack conversations and meeting recordings. The graph is described as continuously updated, capturing metric definitions, entity relationships, lineage, permissions and business rules.
Matthew Drooker, CTO of The Weather Company, is named as an early customer: "Jedify's context graphs give our agents and analysts the business context they need to operate at Weather Company scale."
Market context and competitive positioning
The enterprise AI infrastructure market is fragmenting rapidly as buyers discover that foundation-model access alone does not deliver production-grade agentic systems. A growing category of vendors — spanning semantic layers, knowledge graphs, data-observability platforms and retrieval-augmented generation frameworks — is competing to own the layer between enterprise data and agent reasoning.
Jedify's release pointedly frames the major model providers — naming OpenAI, Anthropic and Google — as structurally conflicted when they offer professional services to build context layers, given they also charge per token. The argument for an independent, model-agnostic layer resonates with enterprise procurement teams increasingly wary of single-vendor lock-in, a concern also embedded in EU AI Act and DORA governance obligations that push financial-services firms and critical-infrastructure operators toward interoperable, auditable AI stacks.
The Snowflake Ventures participation is strategically significant. Jedify's integration with Snowflake Cortex AI — specifically Semantic Views and Cortex Analyst — positions the startup as complementary infrastructure within the Snowflake Data Cloud, which already has deep penetration in enterprise data teams. This partnership accelerates distribution without requiring Jedify to compete head-on with Snowflake's native semantic capabilities.
Standards and governance read-across
As EU AI Act obligations for high-risk systems phase in from 2026, enterprises deploying AI agents in operational workflows will face increasing pressure to demonstrate data lineage, auditability and controlled access to the context their agents consume. A purpose-built, permissioned context graph — rather than ad hoc retrieval from uncontrolled data lakes — maps more naturally to those transparency requirements. UK and US enterprises pursuing SOC 2 or ISO 27001 certification for AI-assisted workflows will similarly benefit from a structured, auditable context layer.
Jedify will next need to demonstrate compounding value in practice: publishing benchmark data on hallucination-reduction rates, token-cost savings, and time-to-production metrics will be the milestones investors and enterprise buyers are watching for.