Jedify builds software that gives enterprise AI agents a live map of a company’s data, documents, permissions, and business language. That pitch just helped the New York startup land a $24 million Series A led by Norwest, at a moment when a lot of enterprise AI still looks smarter in demos than it does inside real companies. Founded by Assaf Henkin, Adi Elimelech, and Erik Shani, Jedify was founded in 2023 and launched its Semantic Fusion platform in 2024. The bet is pretty simple: AI agents don’t fail because the model is weak. They fail because they don’t understand the business they’re dropped into.
What is Jedify’s context graph for enterprise AI?
Jedify’s core product is a context layer that sits between enterprise systems and AI agents. A customer connects sources like databases, warehouses, SaaS apps, BI tools, docs, Slack, and other unstructured content. Jedify then builds a context graph, using its Semantic Fusion engine to connect entities and metrics. It also ties in permissions, workflows, and company-specific definitions into something an agent can reason over.
The workflow is more concrete than the usual “we organize your knowledge” pitch. In the docs, Ask Jedify breaks a user request into entities, metrics, time frames, filters, and groupings, maps those terms to the company’s semantic model, and then composes SQL to fetch the answer. That means a sales leader can ask a plain-English question about pipeline, win rate, or revenue and get a response grounded in internal definitions rather than whatever the model guesses “revenue” should mean.
Jedify also packages that layer into several product modules. There’s Ask Jedify for conversational analytics. There’s a Contextual MCP Server that plugs business context into outside agent frameworks, plus Smart Scenarios for recurring deliverables like QBR decks and reports, and an Insights Library for reusable prompts and shared analysis. The docs also show Slack and Looker integrations, which hints at where the company wants to live: inside the tools employees already use, not as another standalone analytics tab.
That setup helps explain the Kiteworks example in the source article. Kiteworks connected Snowflake, Tableau, Notion, and internal playbooks — including screenshots and documents — then used Jedify to build seller-facing and account-team workflows. Henkin’s description is basically a real-time prep layer for customer calls. The system assembles what the rep needs before a meeting, then surfaces specific details during the conversation itself.
Who founded Jedify and what’s its early traction?
Founding story
The founders have worked together around data and AI for more than 15 years. They previously built one of the earlier open-source intelligence platforms used by large organizations, then went on to run Singtel’s digital data business after that company was acquired. That history matters because Jedify isn’t coming from a pure research angle. It’s coming from people who’ve already lived through messy enterprise data sprawl.
Jedify’s formal story starts in 2023. The trio had watched dashboards, semantic layers, and catalogs pile up while enterprise data got more fragmented and more unstructured. Semantic Fusion — launched in early 2024 — is their answer.
Founder market fit
Henkin is the clearest public signal here. Before Jedify, he co-founded Kontera, a big-data consumer analytics company that Singtel acquired, and later helped build life-insurtech company Sproutt as co-founder, president, and COO. His background is heavy on data products, analytics, and operating inside large organizations after acquisition. That's the kind of experience you'd want if you’re selling deep infrastructure into enterprises.
Shani, Jedify’s CPO, brings more product and go-to-market muscle. He was previously co-founder and chief product officer at Fyllo, a compliance-focused martech company, and his professional profile centers on strategy, product management, sales, and GTM work. That’s useful because Jedify isn’t just building a graph database with nice branding. It’s trying to turn a hard technical layer into a product business.
Elimelech, the CTO, appears to be the technical architect behind the company’s semantic and agent stack. He’s the co-author attached to Jedify’s deeper product explainers on agentic analytics and the autonomous context graph, which lines up with his role as the person translating the thesis into product architecture.
Traction and operating signals
Jedify is still early, but it’s not pre-product. Henkin said the company has between 10 and 20 early customers, including The Weather Company, and is seeing demand from gaming, industrials, and consumer packaged goods. LinkedIn lists the company at 11-50 employees, which fits the picture of a startup that’s past the prototype phase but still small enough to be highly hands-on with deployments.
The product stack also looks live, not aspirational. Jedify has published docs for its conversational analytics product, Smart Scenarios, MCP integrations, Slack app, Looker extension, and a Deep Research Agent. That doesn’t prove scale. But it does show a working platform rather than a single flashy demo.
Fundraising details
The new round is a $24 million Series A led by Norwest. Returning investors S Capital VC and Cerca Partners joined again, with Oceans Ventures added as a new backer. Snowflake came in as a strategic investor and is integrating Jedify with Cortex AI, Semantic Views, and CoWork. The fresh capital will go to product development and hiring. It will also fund go-to-market. Total funding is now about $33 million.
Competition and market positioning
This is where Jedify gets interesting — and where the pitch gets tougher.
Glean now pushes its own Enterprise Graph as the layer that captures relationships across people, projects, teams, and processes for enterprise AI. Moveworks pairs agentic search with a conversational assistant that reaches across connected business systems. Contextual AI, meanwhile, sells an enterprise platform for building specialized RAG agents over structured and unstructured data. Jedify is entering a crowded zone where everyone agrees context matters. They just disagree on where that context layer should live.
Jedify’s differentiation is narrower and more technical. Henkin argues it goes beyond a semantic layer or metadata catalog because it combines structured data and unstructured knowledge. It also pulls business rules and permissions into one graph that updates continuously. Unlike the “just bring everything into our cloud” logic from big platform vendors, Jedify is pitching itself as multi-system and model-agnostic. That’s a good argument if buyers don’t want one warehouse or one AI vendor to own the full control plane. It’s also a risky one, because Snowflake is both a partner and proof that the larger platforms are marching toward the same goal.
Why does this context graph for enterprise AI round matter?
The obvious answer is money. The more useful answer is timing.
A $24 million Series A tells you Jedify’s backers think the company has a shot at becoming infrastructure, not just another workflow app bolted onto an LLM. That matters because infrastructure companies can survive model churn. If Anthropic, OpenAI, Google, or open models keep leapfrogging each other, the value may shift toward the layer that knows how a business actually works.
Snowflake’s involvement sharpens that point. Strategic checks can be cosmetic. This one looks more substantive because it comes with product integration into Cortex AI, Semantic Views, and CoWork. If that integration turns into real distribution, Jedify gets an easier path into data teams that already trust Snowflake. If it doesn’t, Jedify still has the harder but more durable pitch: it helps enterprises stitch together the stuff that never lived in one platform to begin with.
There’s a customer angle here too. Enterprises are starting to get impatient with generic copilots that answer nicely but can’t reliably act. Jedify is selling the less glamorous layer underneath — permissions, entity relationships, metric definitions, governance, observability. That’s the kind of plumbing buyers complain about at first and then refuse to live without once it works.
How big is the context graph for enterprise AI market?
The demand side is real. Gartner forecast worldwide AI spending will hit $2.59 trillion in 2026, up 47% year over year, even as enterprises stay selective and favor practical deployments over moonshot transformation projects. That’s a useful backdrop for Jedify because it suggests buyers still want AI spend tied to measurable workflow gains, not abstract model bragging rights.
The more direct market is smaller but rising fast. Grand View Research estimates the enterprise knowledge graph market was about $2.89 billion in 2025 and could reach $13.37 billion by 2033, a 21.3% CAGR. Put differently: the industry is finally spending real money on the missing layer between raw enterprise data and AI systems that are supposed to do something useful with it.
Final take on Jedify’s enterprise AI context graph
Jedify isn’t promising a magic agent. It’s selling the context graph for enterprise AI that those agents need before they stop embarrassing themselves in production.
That’s a smarter pitch than most. The next question is whether Jedify can turn early design-heavy deployments into a repeatable product and keep its edge while bigger platforms build their own context layers.
Read how Exponent Energy raised ₹200 crore to expand its rapid-charging platform for commercial EVs, using a full-stack system that combines batteries, chargers, and energy management to cut charging times and keep fleet vehicles moving.
FAQ
- What funding did Jedify raise? Jedify raised a $24 million Series A. Norwest led the round, with S Capital VC and Cerca Partners returning, Oceans Ventures joining, and Snowflake investing strategically as part of a broader product integration plan.
- How does Jedify’s product actually work? It connects enterprise systems through APIs and builds a semantic context graph. That layer helps AI agents reason over business-specific data and definitions. In practice, the platform can turn plain-English questions into structured queries. It can also generate recurring reports and feed outside agents through its Contextual MCP Server.
- Who are Jedify’s founders? Jedify was founded by Assaf Henkin, Adi Elimelech, and Erik Shani. Henkin previously co-founded Kontera and later Sproutt, while Shani previously co-founded Fyllo. The team has spent more than 15 years building around data and AI.
- Is Jedify an enterprise search company or an AI infrastructure company? It looks closer to AI infrastructure than classic enterprise search. The company is building a context layer for agents — one that spans structured data, unstructured knowledge, permissions, and semantic definitions — which puts it nearer to the enterprise knowledge graph and agent-enablement category than to a simple search box.
