By Neng Nana 
For the past two years, the corporate world has been swept up in the generative AI gold rush. Enterprises have rushed to adopt large language models (LLMs), expecting an immediate transformation in productivity. However, the reality has been starkly different: AI agents, out of the box, are often "hallucination-prone" and functionally illiterate when it comes to the specific, nuanced operational requirements of a business. They struggle to distinguish between a company’s internal revenue definitions or navigate the labyrinthine permissions of a global corporation.
New York-based startup Jedify is positioning itself as the bridge across this "last mile" gap. By building what it calls a "context graph," the company aims to provide AI agents with the foundational knowledge they need to actually function within a business. Today, Jedify announced it has successfully closed a $24 million Series A funding round led by Norwest, with significant strategic involvement from data giant Snowflake.
The Core Challenge: Why "Turnkey" AI Isn’t Working
AI vendors frequently market their products as seamless, plug-and-play solutions. Yet, in practice, the deployment of these tools often requires a small army of engineers to manually integrate systems and fine-tune models to understand company-specific workflows.
The problem lies in context. An LLM may be a genius at general knowledge, but it knows nothing about the specific, often messy, data silos of an enterprise. It lacks awareness of the complex relationships between a firm’s databases, its SaaS application architecture, its historical documentation, and its internal communication channels like Slack or Microsoft Teams. Without this context, AI agents are essentially guessing, leading to fragmented results and significant security risks.
The Jedify Solution: The Context Graph
Jedify’s platform acts as an intelligent overlay. By connecting to an organization’s existing APIs, it aggregates data from structured sources (data warehouses, data lakes, and BI tools) and unstructured sources (reports, codebases, meeting transcripts, and chat logs).
Unlike a static metadata catalog, Jedify’s "context graph" is dynamic and multi-dimensional. It maps the relationships between entities—people, projects, permissions, and domain-specific terminology—allowing an AI agent to narrow its focus. Instead of searching a company’s entire digital footprint, an agent guided by a context graph can pinpoint exactly the information relevant to a specific task, significantly reducing noise and improving accuracy.
Chronology of Growth and Strategic Alignment
The journey for Jedify has been one of rapid iteration and high-level validation. Since its inception, the company has moved from conceptualizing the "context graph" to proving its utility with high-stakes, data-heavy clients.
- Initial Development: The founding team identified that the primary friction point for enterprise AI was the "cold start" problem—the time it takes for a model to learn the idiosyncrasies of a new environment.
- The Pilot Phase: Jedify began onboarding early adopters, focusing on companies with complex data stacks, such as The Weather Company and various players in the gaming and industrial sectors.
- Strategic Partnership: The most significant milestone to date is the partnership with Snowflake. Snowflake is not merely an investor; it is actively integrating Jedify’s technology into its core AI products, including Cortex AI, Semantic Views, and CoWork. This move signals that the industry’s biggest data infrastructure providers recognize the need for a specialized layer to handle the semantic and contextual "glue" that keeps agents aligned.
- Series A Funding: With the recent $24 million injection, Jedify has now raised a total of $33 million. This latest round was led by Norwest, with participation from returning investors S Capital VC and Cerca Partners, and the addition of Oceans Ventures.
Supporting Data: Why Context is the New Moat
The business logic behind Jedify’s rise is rooted in the shifting economics of AI. As companies face ballooning costs associated with high token usage, they are increasingly scrutinizing their AI spend.
Efficiency vs. Token Usage
Training an AI model to build a proprietary, in-house context layer is a cost-prohibitive endeavor for all but the largest tech giants. By offloading this task to Jedify, enterprises can achieve higher efficiency with fewer API calls. By feeding the model exactly what it needs to know—and nothing more—Jedify helps companies manage the "token bill" that has recently become a major concern for CFOs globally.
Security and Governance
One of the most critical aspects of the enterprise AI equation is data sovereignty. Jedify addresses the fear of unauthorized data leakage by inheriting permissions directly from a company’s existing identity and access management (IAM) systems. Whether it is row-level security in a database or file-level restrictions in a cloud drive, Jedify ensures that the AI respects the existing "guardrails." This level of governance is a prerequisite for any enterprise that handles sensitive financial or personal data.
Official Perspectives: The CEO’s Vision
Assaf Henkin, co-founder and CEO of Jedify, argues that his company’s approach is fundamentally different from existing semantic layers or traditional knowledge graphs.
"When you want to enable an agentic solution to really be autonomous—to drive decisions across CRM data, Zendesk tickets, and real-time telemetry—that’s when a context graph is much better," Henkin noted in an interview.
He highlighted the work done with Kiteworks, a compliance-focused enterprise. By using Jedify to connect Snowflake, Tableau, Notion, and internal playbooks, Kiteworks was able to build a dual-purpose application: a dashboard and a real-time, conversational assistant for their sales teams.
"When they go into a customer conversation, Jedify builds for them, on the fly, everything they need to know," Henkin explained. "During the conversation, they can get very specific details surfaced proactively." This is the "agentic" future that Jedify aims to facilitate—one where AI moves from being a static search tool to an active participant in business workflows.
Implications: The Future of the AI Stack
The rise of companies like Jedify suggests a maturing AI market. The initial phase of "AI excitement" is being replaced by an "AI infrastructure" phase, where the winners are those who can solve the mundane, difficult problems of data fragmentation and system interoperability.
A New Moat for Enterprises
Henkin makes a compelling point regarding the "commoditization" of AI models. As OpenAI, Google, and Anthropic continue to leapfrog each other with more powerful models, the intelligence of the model itself may become a commodity. However, the context—the specific, proprietary relationships between a company’s data and its people—remains unique.
By building a high-quality context graph, enterprises are essentially building a defensible moat. This context makes their AI agents more capable than competitors’, and it is portable, meaning the company is not locked into a single model provider.
Challenging the Cloud Giants
Jedify’s existence poses an interesting question for companies like Snowflake, Databricks, or AWS. While these providers argue that companies should "bring all their data" to their platforms, the reality is that most large organizations are inherently multi-cloud and multi-database. Jedify’s platform, by remaining model-agnostic and source-agnostic, captures the institutional knowledge that sits in the "gaps" between these massive data environments.
The Road Ahead
With $33 million in total funding, Jedify is now in a position to scale its go-to-market efforts. The company plans to focus heavily on hiring and product development, aiming to make the "context graph" a standard component of the modern data stack.
As the industry shifts from "chatting with an AI" to "deploying autonomous agents," the success of those agents will depend entirely on how well they understand the world they operate in. Jedify is betting that the company that owns the map of that world will be the one that defines the future of enterprise automation.
Dwi Wanna 
Laily UPN 
Suro Senen 