For more than two decades, Data Loss Prevention (DLP) has sat in a strange place in the security stack. In theory, it protects the crown jewels: product roadmaps, source code, customer lists, financial documents, and sensitive internal conversations. In practice, it has become synonymous with noise, friction, and false promises.
Most security leaders I speak with are caught between two bad options. Either they run legacy DLP “for compliance” and drown in noisy alerts, constant tuning, and unhappy end users. Or they quietly avoid DLP altogether because the operational cost is guaranteed and the value is uncertain, while knowing full well that sensitive data is walking out the door every day.
The stakes are high. According to Verizon’s 2025 Data Breach Investigations Report, the human element is involved in roughly 60% of data breaches - simple mistakes, manipulated employees, or insider misuse. The risk is not theoretical; it is built into how modern work happens.
Jazz was created for this exact moment.
Jazz is turning DLP from a rigid rule engine into an intelligence system, one that actually understands how data is used, why a given action is happening, and whether it represents real risk. Today, Jazz is emerging from stealth with $61M in Seed and Series A funding, co-led (for the first time) by Glilot Capital Partners and Team8, with participation from Ten Eleven Ventures (1011), Merlin Ventures, Encoded Ventures, and MassMutual Ventures. Jazz is already in production with more than a dozen customers, including AlphaSense, SimilarWeb, and CAVA.
We believe Jazz is the beginning of a new chapter for data protection: one where DLP finally understands intent and context, instead of forcing security teams to guess them in advance.
Why legacy DLP failed
DLP was originally designed around a simple idea: define what is sensitive, write rules to catch it leaving, and block or alert.
For a different era, that made sense. But modern work looks nothing like the environments DLP was built for.
Data lives everywhere: SaaS apps, shared drives, cloud storage, collaboration tools, GenAI interfaces, unmanaged devices. Employees share, paste, copy, sync, and export as part of their daily work. The difference between “normal” and “dangerous” behavior is often a matter of intent and subtle context, not pattern matches.
Rule-based DLP simply can’t keep up. To make it work, teams are forced to predict every pattern of misuse in advance, encode it in rigid rules, and update those rules every time the business, tools, or workflows change.
The outcome is familiar to anyone who has lived with DLP in a large organization. Alerts multiply until no one believes them. Policies are tightened until they break legitimate work. Exceptions are added until policies are meaningless. Eventually, DLP either becomes a noisy background system that everyone ignores, or a stalled initiative that never leaves the pilot phase.
Meanwhile, the core problem remains: sensitive data leaves the organization through everyday actions - an attachment sent to the wrong place, a copy-paste into a GenAI tool, a file synced to an unmanaged app.
Legacy DLP doesn’t struggle because security teams don’t care enough. It struggles because it doesn’t understand what is actually happening. It sees patterns, not behavior.
Jazz’s thesis: DLP that deeply understands your business
Jazz starts from a different premise: to protect data in modern enterprises, you have to understand how people work, not just what the data looks like.
Instead of asking security teams to predict and encode every risky scenario, Jazz uses an autonomous Agentic Investigator that learns the organization’s business processes and real usage patterns. The system analyzes the full context of every relevant event: who the user is, what data is involved, which system is in play, and what business process this action appears to belong to.
From there, Jazz focuses on the key question: what was the intent?
Was someone following an approved workflow, using the right channels and destinations for that type of data? Or were they doing something that is potentially risky - whether negligent or malicious: sending a sensitive file outside approved domains, pasting confidential text into a GenAI system, or exfiltrating data to a personal app?
By modeling intent and context, Jazz automatically distinguishes between legitimate workflows and actual risk. It doesn’t just trigger a rule; it produces a pre-investigated explanation in plain English: what happened, why it matters, and whether it represents real exposure.
The impact of this approach is already clear. In one 1,000-employee customer deployment, Jazz reduced daily DLP noise from thousands of low-confidence detections to an average of two pre-investigated incidents per day. Instead of trying to triage a sea of alerts, security teams can focus on the handful of moments that truly matter.
This is the core of Jazz’s value: turning DLP from a blunt, rule-driven system into an intelligent investigator that understands how data is actually used inside the business.
Why now: GenAI, collaboration, and the new data perimeter
The timing for Jazz is not incidental.
On one side, enterprises are rushing to adopt GenAI and modern collaboration tools. Employees routinely paste text into AI assistants, export data into external tools, and connect new services that were never part of the original security design. The traditional “data perimeter” has dissolved into a distributed, constantly changing landscape.
On the other side, legacy data protection has become a bottleneck. The choice for many security leaders has been stark: accept the risk, or accept the operational pain
In this environment, a rule-only approach to DLP simply doesn’t scale. You cannot maintain enough static patterns to cover all the tools, workflows, and edge cases where sensitive data might appear. You cannot freeze collaboration without damaging the business.
What you can do is build a system that observes how data is used, learns the patterns that represent normal work, and highlights only the deviations that genuinely matter. That is exactly what Jazz is doing.
This is also why Jazz resonates so strongly with customers like AlphaSense, SimilarWeb, and CAVA. They are not looking for yet another DLP product. They are looking for a way to adopt AI and modern tools without losing control of their data.
The founders
Our conviction in Jazz is ultimately rooted in the team.
Jazz was founded by Ido Livneh (CEO), Jake Tuertskey (Chief AI), Noam Issachar (CBO), and Yonatan Zohar (CTO), veterans of Unit 81, one of Israel’s elite technology units, and alumni of companies including Axonius and Laminar.
They bring a rare combination of deep security engineering, large-scale data experience, and real-world exposure to how DLP and data protection actually fail in modern enterprises. They have seen, up close, how much effort goes into tuning and re-tuning legacy systems, how often real incidents are buried in noise, and how hard it is to get business stakeholders to trust DLP programs that constantly interrupt legitimate work.
From our first conversations, it was clear they were not interested in adding another rule engine or UI layer on top of the same model. Their focus was on rebuilding DLP from first principles: an AI-native, context-driven platform that combines a forensic endpoint agent for total visibility with an Agentic Investigator that can explain what happened and why in clear, human terms.
That is the mindset we look for at Team8: founders who have lived the problem at scale, who are willing to challenge a 20-year-old category, and who can execute quickly enough to prove that the market is ready. The fact that Jazz has already earned more than a dozen paying customers before launch is a strong signal that they are solving a problem the industry has been waiting to address.
Looking ahead
Jazz’s $61M in funding will be used to amplify its Agentic Investigator, expand enterprise deployments, and grow its engineering, research, and go-to-market teams. But the more important story is what Jazz represents.
For over twenty years, DLP has forced security teams into an unfair tradeoff: either accept the risk of data walking out the door, or accept the operational pain of systems that don’t understand how the business works.
Jazz is showing that this tradeoff is no longer necessary. By giving DLP an actual understanding of intent and context, it turns data protection from a blunt instrument into a precise, intelligence-driven capability and, in the process, turns security from a blocker into an enabler for how modern organizations want to work.
That is why we chose to co-lead Jazz’s funding round with Glilot, and why we’re excited to support the team as they remaster DLP for the AI era.
Learn more about Jazz here - https://jazz.security/
Co-Founder & Managing Partner
Liran Grinberg is the Co-founder and Managing Partner of Team8, where he invests in Cyber and Software Infra companies.
