Every large enterprise today operates under an explicit artificial intelligence mandate, yet very few organizations truly understand what happens the moment an autonomous agent is granted direct access to enterprise systems and internal data. Questions regarding what a system will be able to reach, how actions can be tracked, and what happens to sensitive records remain unresolved across the corporate landscape. According to industry insiders and venture capital backers, these critical uncertainties—rather than model quality, infrastructure costs, or a scarcity of specialized talent—explain why so many enterprise AI programs remain permanently stuck in the pilot phase. A specialized security startup known as Cymphony was built specifically to solve these foundational dilemmas.
Agent data access is the constraint
Modern employees increasingly work side-by-side with autonomous AI agents to execute daily operational workflows. However, the typical software agent bears very little resemblance to a human employee. An agent can be provisioned within minutes rather than being recruited and onboarded over the course of several weeks. It operates continuously without fatigue, and it can traverse a vast landscape of systems, databases, and functional capabilities that far exceeds the reach of any single individual. Every one of these unique properties is precisely what makes deploying an AI agent valuable to a business—and every one of them is simultaneously what prompts corporate security teams to deny deployment requests.
Governance and security concerns have already transformed into major barriers to broader AI adoption. Market research from Gartner indicates that approximately 40% of enterprises are expected to formally demote or decommission autonomous AI agents over governance anxieties by next year. Industry analysts emphasize that this trend does not reflect a failure of underlying artificial intelligence technology to perform its intended tasks. Instead, it highlights a scenario where enterprises are pulling back fully functional agents because they lack the necessary visibility and control to determine what those agents are accessing. The fundamental challenge facing modern corporations is not utility, but accountability.
For decades, enterprise security teams have constructed their defensive strategies around two distinct and separate concerns: identifying where sensitive data lives, and verifying who possesses access to it. The first concern gave rise to specialized software tools designed to monitor files, email inboxes, and corporate databases for accidental or malicious leaks. The second concern built the massive enterprise identity and access management industry. However, both of these traditional paradigms fundamentally assume that the entity being monitored is a human employee.
Artificial intelligence has effectively broken that core assumption. A traditional identity management tool can confirm that an agent exists within the network, while a data security tool can flag a specific file as sensitive. Neither tool, however, can determine whether a given agent—acting dynamically on behalf of a specific human employee—can reach that sensitive file at any given moment. In the era of autonomous systems, traditional identity security and data security are no longer separable disciplines.
Cymphony is the solution
To address this growing architectural gap, Cymphony continuously maps both human employees and autonomous agents across the entire enterprise ecosystem. The platform constructs a dynamic graph that outlines precisely who and what can reach specific systems and sensitive information. Solving this challenge represents a formidable engineering hurdle. In a typical organization, the same person routinely holds a dozen different user accounts scattered across disparate software systems that were never originally designed to communicate with one another, while autonomous agents inherit access permissions from both the humans and the background applications that spawned them. Stitching all of these complex relationships back together into a single, accurate operational picture requires far more than a basic dashboard that merely aggregates security alerts.
Industry observers note that the approach taken by Cymphony co-founders Shy Dekel, Idan Berkovits, and Edi Gotlieb centers on building for the underlying workflow rather than stopping at the alert. Instead of merely identifying a potential security risk and leaving a human security analyst to figure out the appropriate response, Cymphony goes a step further by recommending and automating remediation strategies while actively coordinating with the personnel who need to take action.
Early enterprise customers, including major global organizations such as Syngenta, KKR, Cass Information Systems, and Athennian, are already deploying the platform and observing tangible operational results. Sequoia Capital, which also utilizes the security platform internally, announced that it led Cymphony’s initial seed funding round and is now returning to co-lead the startup’s Series A financing.
Why we partnered
The founding team behind Cymphony was discovered through Sequoia’s extensive technology network in Israel, where they were introduced independently by multiple sources who shared a consistent observation: these three founders had spent their professional careers tackling difficult, unglamorous security problems that most entrepreneurs actively avoid, and they were constructing an entirely unique approach to solving them. All three co-founders emerged from Israel’s elite Talpiot military program and bring deep domain expertise to the table. Shy Dekel previously led the cyber department for Unit 8200, Idan Berkovits directed a specialized research group within the Office of the Prime Minister and earned an Israel Defense Prize, and Edi Gotlieb spent years building advanced hardware systems at Apple and the Israeli Ministry of Defense.
Venture capital investors point out that every major enterprise software category has been built on a generational transition where the underlying unit of corporate work changes, rendering older security paradigms obsolete. The rise of cloud computing fundamentally altered where software applications were executed, and the security industry was forced to rebuild itself around that new reality. Mobile computing subsequently redefined what constituted a corporate endpoint, forcing security frameworks to adapt once again. During each of these structural shifts, the companies that ultimately provided comprehensive visibility and operational control over the new environment eventually became essential infrastructure that everyone took for granted.
Artificial intelligence represents a parallel shift, though it stands out as the first technological transition where the new unit of work possesses the capacity to act entirely on its own accord. The company that successfully empowers an enterprise to say yes to autonomous agents—confidently, at scale, and with a reliable audit trail of their actions—will function as much more than a simple feature nestled inside the broader AI technology stack. It will serve as a fundamental precondition for enterprise AI adoption as a whole. Backers believe Cymphony is actively constructing that vital infrastructural layer.
Comments (0)