AI Agent Security: Cyata Raises $8.5M to Tame Autonomous Enterprise AI
August 3, 2025, 9:31 pm
Tel Aviv's Cyata emerges with $8.5M seed funding, led by TLV Partners, to secure enterprise AI agents. As AI integration accelerates, autonomous agents pose new risks beyond traditional cybersecurity. Cyata offers a novel control plane for "agentic identities," featuring automated discovery, forensic observability, and granular access control. Their platform uniquely uses AI to interrogate agents, ensuring compliance and mitigating risks. This addresses critical gaps for businesses deploying AI copilots and autonomous systems, providing essential governance in a rapidly evolving digital landscape. It safeguards sensitive data and ensures accountability for AI actions. Cyata aims to future-proof AI security for all enterprises.
A new digital frontier has arrived. Enterprises rapidly adopt artificial intelligence. Generative AI tools, copilots, and autonomous agents are transforming workflows. They bring unprecedented efficiency. Yet, this transformation introduces profound security challenges. Traditional cybersecurity models struggle to keep pace. A new paradigm is emerging for AI agent security.
Cyata, a Tel Aviv-based cybersecurity startup, steps into this critical void. The company recently launched from stealth. It secured $8.5 million in seed funding. TLV Partners led the investment round. Notable angel investors also participated. These include Ron Serber and Yossi Carmil, former CEOs of Cellebrite. Cyata’s CEO, Shahar Tal, previously served as Cellebrite’s VP of business development. This experienced leadership team signals a serious approach to a complex problem.
The problem is urgent. A vast majority of IT and data executives, an estimated 96%, plan to increase their use of AI agents this year. These agents operate autonomously. They execute tasks with minimal human oversight. They can spin up in milliseconds. They fork into sub-agents. They make privileged calls. Then, they vanish. Existing identity and access management (IAM) or privileged access management (PAM) systems cannot cope. Legacy tooling is simply outmatched. AI agents are faster. They are more privileged. They can be more error-prone. This creates significant vulnerabilities for sensitive corporate data and systems. Businesses face a new wave of cybersecurity risk.
Cyata delivers a dedicated solution. It introduces a control plane for "agentic identities." These are the autonomous digital workers within an enterprise. The platform features three integrated capabilities. Each addresses a crucial aspect of AI security.
First, Cyata offers automated discovery. It continuously scans desktop and SaaS environments. It finds all AI identities. It assesses their risk profiles. Each agent maps to a human owner. This provides vital visibility. Many organizations underestimate their AI agent usage. Unexpected agents, like those embedded in popular tools, often operate with elevated permissions. They impersonate users. They access sensitive data without oversight. Cyata uncovers these "shadow AI" instances.
Second, the platform provides forensic observability. It creates detailed audit trails of agentic AI activity. This includes capturing intent. Agents are prompted to justify their reasoning in real-time. This is a crucial innovation. Cyata interrogates agents in natural language. An agent attempting a task must explain its actions. Rules-based logic and AI models then evaluate this justification. This is AI evaluating AI. It scores context and intent. This added layer of interpretability strengthens risk assessment. It detects anomalous behaviors. It flags risky access patterns. It ensures accountability for every AI action.
Third, Cyata enforces granular access control. Permissions apply just-in-time. Human-in-the-loop approvals remain for sensitive operations. This maintains essential human oversight. The system recognizes agents upon authentication. It traces their activity. It enforces the principle of least privilege in real time. This ensures AI agents only access what they absolutely need, when they need it. It mitigates the risk of unauthorized data access or system compromise.
Deployment is rapid. Cyata highlights a 48-hour integration process. It quickly scans cloud environments, copilots, and other tools. It surfaces agentic identities and their associated risks. This offers near-immediate value to enterprise security teams. The system integrates with common platforms. Microsoft Copilot, Salesforce AgentForce, and various identity providers are supported. This ease of integration accelerates adoption. It bridges the gap between traditional identity systems and the burgeoning AI workforce.
The implications extend beyond technical teams. Initially, developers were seen as the primary audience for AI agents. Cyata quickly observed broader adoption. Non-developers, including sales, finance, and support teams, are deploying agents rapidly. This widespread use makes centralized governance essential. Cyata’s platform operates in multiple modes. It ranges from passive monitoring to active enforcement. This flexibility allows security teams to adopt it without disrupting existing workflows. It can flag risky activity. It can suggest mitigations. It can enforce human approvals for high-privilege actions.
Cyata’s pricing follows a SaaS model. It is based on the number of managed agentic identities. This straightforward approach aids enterprise budgeting. The company’s vision is clear. It aims not just to patch current security gaps. It seeks to prepare enterprises for a fundamental shift in work execution. With its team of cybersecurity veterans, Cyata is poised to lead this emerging category. The company plans to unveil new research at Black Hat. It also builds a partnership program. This will deepen integrations with identity vendors and enterprise platforms.
As AI agents become ubiquitous, the need for robust AI security tools becomes paramount. Enterprises must understand who – or what – is acting on their behalf. Cyata’s solution offers a vital control plane. It delivers necessary visibility, accountability, and governance. It secures the future of enterprise AI. This makes it a crucial investment for any organization navigating the complexities of autonomous digital workers. Cyata addresses enterprise AI governance directly. It protects sensitive data from advanced AI threats. It is essential for AI risk management.
A new digital frontier has arrived. Enterprises rapidly adopt artificial intelligence. Generative AI tools, copilots, and autonomous agents are transforming workflows. They bring unprecedented efficiency. Yet, this transformation introduces profound security challenges. Traditional cybersecurity models struggle to keep pace. A new paradigm is emerging for AI agent security.
Cyata, a Tel Aviv-based cybersecurity startup, steps into this critical void. The company recently launched from stealth. It secured $8.5 million in seed funding. TLV Partners led the investment round. Notable angel investors also participated. These include Ron Serber and Yossi Carmil, former CEOs of Cellebrite. Cyata’s CEO, Shahar Tal, previously served as Cellebrite’s VP of business development. This experienced leadership team signals a serious approach to a complex problem.
The problem is urgent. A vast majority of IT and data executives, an estimated 96%, plan to increase their use of AI agents this year. These agents operate autonomously. They execute tasks with minimal human oversight. They can spin up in milliseconds. They fork into sub-agents. They make privileged calls. Then, they vanish. Existing identity and access management (IAM) or privileged access management (PAM) systems cannot cope. Legacy tooling is simply outmatched. AI agents are faster. They are more privileged. They can be more error-prone. This creates significant vulnerabilities for sensitive corporate data and systems. Businesses face a new wave of cybersecurity risk.
Cyata delivers a dedicated solution. It introduces a control plane for "agentic identities." These are the autonomous digital workers within an enterprise. The platform features three integrated capabilities. Each addresses a crucial aspect of AI security.
First, Cyata offers automated discovery. It continuously scans desktop and SaaS environments. It finds all AI identities. It assesses their risk profiles. Each agent maps to a human owner. This provides vital visibility. Many organizations underestimate their AI agent usage. Unexpected agents, like those embedded in popular tools, often operate with elevated permissions. They impersonate users. They access sensitive data without oversight. Cyata uncovers these "shadow AI" instances.
Second, the platform provides forensic observability. It creates detailed audit trails of agentic AI activity. This includes capturing intent. Agents are prompted to justify their reasoning in real-time. This is a crucial innovation. Cyata interrogates agents in natural language. An agent attempting a task must explain its actions. Rules-based logic and AI models then evaluate this justification. This is AI evaluating AI. It scores context and intent. This added layer of interpretability strengthens risk assessment. It detects anomalous behaviors. It flags risky access patterns. It ensures accountability for every AI action.
Third, Cyata enforces granular access control. Permissions apply just-in-time. Human-in-the-loop approvals remain for sensitive operations. This maintains essential human oversight. The system recognizes agents upon authentication. It traces their activity. It enforces the principle of least privilege in real time. This ensures AI agents only access what they absolutely need, when they need it. It mitigates the risk of unauthorized data access or system compromise.
Deployment is rapid. Cyata highlights a 48-hour integration process. It quickly scans cloud environments, copilots, and other tools. It surfaces agentic identities and their associated risks. This offers near-immediate value to enterprise security teams. The system integrates with common platforms. Microsoft Copilot, Salesforce AgentForce, and various identity providers are supported. This ease of integration accelerates adoption. It bridges the gap between traditional identity systems and the burgeoning AI workforce.
The implications extend beyond technical teams. Initially, developers were seen as the primary audience for AI agents. Cyata quickly observed broader adoption. Non-developers, including sales, finance, and support teams, are deploying agents rapidly. This widespread use makes centralized governance essential. Cyata’s platform operates in multiple modes. It ranges from passive monitoring to active enforcement. This flexibility allows security teams to adopt it without disrupting existing workflows. It can flag risky activity. It can suggest mitigations. It can enforce human approvals for high-privilege actions.
Cyata’s pricing follows a SaaS model. It is based on the number of managed agentic identities. This straightforward approach aids enterprise budgeting. The company’s vision is clear. It aims not just to patch current security gaps. It seeks to prepare enterprises for a fundamental shift in work execution. With its team of cybersecurity veterans, Cyata is poised to lead this emerging category. The company plans to unveil new research at Black Hat. It also builds a partnership program. This will deepen integrations with identity vendors and enterprise platforms.
As AI agents become ubiquitous, the need for robust AI security tools becomes paramount. Enterprises must understand who – or what – is acting on their behalf. Cyata’s solution offers a vital control plane. It delivers necessary visibility, accountability, and governance. It secures the future of enterprise AI. This makes it a crucial investment for any organization navigating the complexities of autonomous digital workers. Cyata addresses enterprise AI governance directly. It protects sensitive data from advanced AI threats. It is essential for AI risk management.