You give an AI agent access to your company’s Slack workspace. The prompt is clear, the guardrails are set. But the agent, acting on its own, starts scraping channel histories it wasn’t meant to see, or pinging service accounts with overly broad permissions. The risk isn't in the model's reasoning, but in the identity you just handed it,an identity that now lives, untracked, among thousands of other human and machine credentials. This is the moment Cygrid is built for, a moment of silent, automated overreach that traditional security tools were not designed to see.
Founded in Berlin in 2026, Cygrid sells what it calls an "Identity Intelligence Graph" [Cygrid]. It is a platform that maps every actor in a company's digital environment,human employees, service accounts, API keys, and now, AI agents,into a single, scored model of risk. Its promise is not just visibility, but automated enforcement: the ability to revoke access, rewrite a command, or block an action before it executes [Cygrid]. In October 2026, the company emerged from stealth with a €2.66 million pre-seed round led by Atlantic, with participation from D11Z.Ventures and several angel investors [EU-Startups, October 2026].
The Graph as a New Security Surface
Cygrid’s core bet is that identity is the unifying layer for modern enterprise security, and that this layer has been fractured by the arrival of autonomous AI. Traditional identity governance tools were built for a world of human users and scheduled service accounts. They audit, they report, they flag anomalies after the fact. Cygrid’s founders argue that in a world where AI agents can act at machine speed, post-incident reporting is too late. The product, therefore, is engineered for pre-emptive control.
The platform connects to major SaaS environments like Google Workspace, Microsoft 365, GitHub, Slack, and HubSpot, discovering every identity and its permissions [Cygrid]. It then applies policy in real-time. If an AI agent attempts an action outside its scope, or a service account token is used in an unusual context, Cygrid claims it can intervene programmatically. The goal is to turn sprawling identity data into a quantified, enforceable security posture [Crunchbase].
A Team Built for Intelligence Operations
The founding team’s background reads less like a typical SaaS startup and more like a security intelligence unit. Co-founder and CEO Dmitry Kharchenko brings a mix of military intelligence experience and a track record as a multi-time startup CTO, with prior roles at companies later acquired by SCHUFA, Cover Genius, and Proofpoint [PERPLEXITY SONAR PRO BRIEF]. Co-founder and Chief Security Officer Dekel Paz is described as a cybersecurity and intelligence veteran [PERPLEXITY SONAR PRO BRIEF]. This pairing suggests a foundational focus on threat modeling and operational security, rather than just feature development.
Their early commercial beachhead appears to be design partnerships with companies like Barfer’s and Horizn Studios, where the platform is reportedly already running in production [EU-Startups, October 2026]. The €2.66 million in capital is earmarked for expanding the team, with seven open roles advertised across engineering, product, and go-to-market functions [Cygrid Careers].
| Role | Focus Area | Location |
|---|---|---|
| Founding Backend Engineer (Go) | Engineering | Not Specified |
| Founding DevOps Engineer / SRE | Engineering | Not Specified |
| Founding Product Manager | Product | Not Specified |
| Account Executive, Full-Cycle | Go-to-Market | Berlin |
| GTM Engineer | Go-to-Market | Remote EU or Berlin |
| Founding Associate to CEO | Operations/Growth | Berlin |
| Lead Generation Specialist | Go-to-Market | Not Specified |
The Risks of Building a New Category
Cygrid’s ambition is to define and own a new category,identity enforcement for AI-driven enterprises. This is a potent but perilous position. The primary challenge is not technological skepticism, but market education and integration depth.
- Category creation. The company must convince security teams that AI agent governance is a discrete, urgent problem requiring a dedicated platform, not just an extension of existing cloud security or identity governance tools. This requires narrative discipline and early, vocal customer advocates.
- Integration burden. Real-time enforcement is only as good as the platform’s reach. While it lists major SaaS systems, enterprise environments are mosaics of legacy and custom tools. Each new integration is a development and maintenance cost, and gaps in coverage could become attack vectors.
- The competitive horizon. While no direct competitors are named in available sources, the space is adjacent to well-funded giants in cloud security, identity governance, and AI safety. Cygrid’s window is to establish deep technical differentiation and customer loyalty before larger players decide to build or buy in.
The company’s answer to these risks appears to be a focus on deep, actionable intelligence,the "graph",rather than mere monitoring, and a team built to operationalize that intelligence. The pre-seed funding provides runway to prove that this wedge is sharp enough to carve out a market.
What to Watch in Berlin
The next twelve months for Cygrid will be about translating its intelligence-grade platform into commercial proof. The key milestones are clear: converting design partners into referenceable case studies, landing its first few full-cycle enterprise deals, and demonstrating that its enforcement capabilities tangibly reduce risk in environments where AI agents are actively deployed. Success will be measured not just in revenue, but in whether security leaders begin to frame the problem of AI agent access in Cygrid’s terms.
The platform, in the end, is answering a cultural question that has quietly arrived with the proliferation of AI assistants. We have spent years debating what these models should think, but we have only just begun to manage what we have allowed them to do. Cygrid’s graph is an attempt to draw the map for that new territory of action, a ledger of all the things,human, machine, and now artificial,that have been granted the power to act in our name.
Sources
- [Cygrid] Identity Enforcement for AI-Driven Enterprises | Cygrid | https://www.cygrid.io/
- [EU-Startups, October 2026] Berlin-based Cygrid raises €2.66 million pre-seed to enforce identity security for AI agents | https://www.startup.eu/investments/cygrid-2-66-m-pre-seed-10-2026
- [PERPLEXITY SONAR PRO BRIEF] Cygrid Company Brief
- [Crunchbase] Cygrid - Crunchbase Company Profile & Funding | https://www.crunchbase.com/organization/cygrid
- [Cygrid Careers] General Application (Talent Pool) - Careers at Cygrid GmbH | https://careers.cygrid.io/o/general-application-talent-pool