By Lina Irawan 
The promise of enterprise artificial intelligence is often sold as a "turnkey" revolution—a plug-and-play solution that magically synthesizes corporate wisdom into actionable intelligence. However, the reality for most businesses is far more granular. Off-the-shelf AI agents frequently falter when tasked with high-stakes corporate operations because they lack the "tribal knowledge" required to function accurately. They struggle to define company-specific metrics like net revenue, they fail to navigate complex internal permission structures, and they remain blind to the nuanced workflows that define daily operations.
New York-based startup Jedify is attempting to bridge this chasm. By building what it calls a "context graph," the company aims to provide AI agents with the relational map necessary to act with precision. This week, the company announced a $24 million Series A funding round, signaling significant market confidence in the need for structured, intelligent data orchestration layer between raw data and generative AI models.
The Core Challenge: Why AI Fails in the Enterprise
For years, the industry narrative suggested that simply pointing a Large Language Model (LLM) at a database would yield a business-ready assistant. Experience has proven otherwise. Without deep, context-aware training, AI models lack the "business literacy" to distinguish between sensitive internal data and public-facing information, nor can they intuitively understand the hierarchical relationships between a company’s departments, CRM entries, and project documentation.
This is why major AI vendors are currently forced to deploy armies of engineers to manually integrate products into customer environments. It is a costly, time-consuming process that prevents AI from scaling across the enterprise. Jedify’s solution is to automate this integration by aggregating knowledge from across the entire tech stack—including databases, SaaS applications, BI tools, and even unstructured repositories like Slack threads and meeting transcripts—to form a comprehensive, real-time context graph.
Chronology: Building the Bridge to Context
Jedify’s trajectory reflects the rapid maturation of the enterprise AI sector:
- The Conceptual Phase: Co-founder and CEO Assaf Henkin identified that while models were becoming increasingly powerful, the "data plumbing" connecting those models to business reality was fractured.
- Initial Traction: The company began working with mid-market and enterprise clients, including The Weather Company, to test the efficacy of their context-mapping platform.
- The Strategic Pivot: Recognizing that data exists in silos, the team focused on creating an API-first approach that is model-agnostic, ensuring the context graph remains useful regardless of which LLM a company chooses to adopt.
- Series A Milestone: With $24 million in new funding led by Norwest Venture Partners, and participation from Snowflake, S Capital VC, Cerca Partners, and Oceans Ventures, Jedify is now positioned to expand its engineering team and accelerate its go-to-market strategy.
Supporting Data and Technical Architecture
The "context graph" is not merely a keyword-search engine or a traditional metadata catalog. According to Henkin, it is a multi-dimensional architecture that tracks the relationships between entities, people, data permissions, and operational workflows.
How the Platform Works:
- Ingestion: Jedify connects via APIs to an enterprise’s entire ecosystem, from Snowflake data warehouses to Zendesk tickets and internal Notion documentation.
- Contextual Mapping: The platform creates a live map of how data points interact. For example, it understands not just that a "customer" exists, but which account manager is responsible for that customer, what their recent Slack sentiment has been, and which documents they have access to.
- Permission Inheritance: A critical feature of the Jedify platform is its ability to respect existing security governance. It ingests row-level, column-level, and table-level access rules from the underlying identity and storage systems. This prevents the "hallucination" of access, ensuring an intern never sees sensitive financial projections.
- Observability: The platform includes governance tools that allow IT leaders to monitor agent behavior, ensuring that autonomy does not lead to operational risk.
Official Perspectives and Real-World Application
The partnership with Snowflake is particularly telling. As large data platforms scramble to provide their own AI capabilities—such as Cortex AI and Semantic Views—they have identified Jedify as a strategic partner.
Assaf Henkin argues that while data giants like Snowflake are powerful, they cannot solve the problem of fragmented knowledge. "They will tell you, ‘Oh yeah, just bring everything,’" Henkin noted. "But in reality, companies have multiple databases, warehouses, and solutions. The big thing is that not all of your data is in those environments, and most of your knowledge is not there. It’s a disadvantage that they have."
A prime example of Jedify’s impact is Kiteworks, a compliance-focused company. Kiteworks utilized Jedify to unify their disparate data sources—Tableau, Notion, and internal playbooks—into a singular, conversational interface.
"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 turns an AI from a static search engine into a real-time, side-by-side business partner.
Implications for the Future of Enterprise AI
The success of this funding round highlights several critical shifts in the AI market:
1. The Death of the "Generalist" AI
As companies realize that generic AI is insufficient for complex enterprise needs, the demand for "specialized context" will skyrocket. The value of an AI implementation will no longer be measured by the model itself—which is becoming a commodity—but by the proprietary context layer that allows that model to navigate the enterprise.
2. The Economics of Token Usage
Many firms are currently struggling with the "token bill." By using a context graph, companies can be much more precise in the information they feed into an LLM. Instead of searching a massive, unstructured database—which consumes expensive tokens and increases the risk of inaccuracy—the context graph narrows the focus to only the information relevant to the task at hand. This efficiency is rapidly becoming a competitive advantage.
3. The "Moat" of Context
Henkin’s broader thesis is that as AI models grow more interchangeable, the "moat"—the durable competitive advantage—will shift to the context layer. A company that possesses a well-maintained, real-time map of its own business data, permissions, and institutional knowledge will be able to pivot to new, better AI models without losing the intelligence it has built over years.
4. Governance as a Catalyst for Adoption
Perhaps the biggest hurdle for enterprise AI is fear—fear of data leaks, hallucinations, and security breaches. By baking governance and permission inheritance into the foundation of the context graph, Jedify is addressing the primary objection held by CISOs and CTOs. When security is "baked in" rather than "bolted on," organizations are significantly more likely to green-light autonomous agent deployments.
Conclusion: The Road Ahead
With $33 million in total funding to date, Jedify is entering a growth phase that will test whether its context-graph approach can become the standard operating layer for enterprise AI. As the industry moves away from the "hype" phase of generative AI and toward the "utility" phase, the ability to make AI agents truly "business literate" will be the defining factor in which companies successfully derive ROI from their technology investments.
The company is currently targeting data-heavy sectors—including gaming, industrial manufacturing, and consumer packaged goods—where the volume of data is high and the cost of human error is significant. If Jedify can continue to prove that its graph is the missing link for autonomous enterprise agents, it may well become the nervous system upon which the next generation of business intelligence is built.
Evan Lee Salim 
Nana Wu 