Kontext raises $4M to stop AI agents from going rogue at work

Kontext, the runtime security platform for AI agents, today announced $4 million in funding. The company is backed by 42CAP, a16z CSX, and HTGF, with 42CAP leading the financing. The funding will support the expansion of Kontext’s engineering team, continued development of its runtime enforcement platform, and help customers safely deploy AI agents with greater control and visibility.

The funding arrives at a critical moment for AI security. AI agents are moving beyond chat interfaces and software development, becoming personal assistants that directly integrate into the workplace, where they can write code, access files, and act with company credentials. A recent incident showed how quickly those risks can become real. In July, AI agents running in a cybersecurity evaluation circumvented their intended isolation, communicated through unauthorized channels, and compromised external infrastructure without direct human instruction. As businesses give agents greater access and autonomy, they need safeguards that control what those agents are allowed to do while they are running.

Kontext builds software that sits between AI agents and the tools and systems they act on. It evaluates agent activity in real time against security policies and risk signals, taking into account the agent’s identity, requested action, target resource, and assigned task. This gives organizations visibility into how agents behave, control over what they are authorized to do, and the ability to enforce clear boundaries.

An agent asked to fix a software bug, for example, may need permission to read a repository. That does not mean it should be allowed to send the code to an external service, modify unrelated infrastructure, or use the same access for a different task.

Teams can initially run Kontext in observe mode to understand agent behavior, identify risky activity, and see how policies would apply without interrupting work. When enforcement is enabled, Kontext can deny unauthorized actions before they execute and retain an auditable record of each decision.

“An AI agent can be properly authenticated, use an approved tool, and still take an action no one authorized,” said Jens Ernstberger, co-founder of Kontext. “As agents move from generating text to operating software, companies need a control point at the moment of action. Kontext connects identity with task context and policy to decide what an agent is allowed to do before it happens.”

Kontext was founded by Jens Ernstberger and Michel Osswald, who bring backgrounds in secure computing, applied cryptography, and AI systems. At Kontext, the team focuses on a specific failure mode in autonomous systems: agents can operate with valid credentials and approved tools while taking actions that exceed the authority of their assigned task. This work underpins Kontext’s task-aware approach to runtime enforcement.

“The identity and access tools built for the last twenty years assume a human is on the other end, clicking one thing at a time,” said Julian von Fischer, General Partner at 42CAP. “An AI agent authenticates once and then acts on its own, across a dozen systems, on a task nobody is watching step by step. That’s a structurally different problem, and the incidents we’ve already seen this year show it’s not theoretical. Jens and Michel bring years of secure computing and applied cryptography to it: most tools still stop at the credential, Kontext looks at the task the agent was actually given. We believe this infrastructure becomes essential the moment a company puts agents into production, which is why we’re leading this round.”

About Kontext
Kontext is a runtime security platform for AI agents. It gives teams visibility into how agents behave and control over what they are allowed to do. Kontext evaluates activity in real time using identity, task context, security policies, and cyber-risk signals. It can identify risky behavior, deny unauthorized actions before execution, and create an auditable record of what each agent attempted, what was allowed or denied, and why.

About 42CAP
42CAP invests in early-stage technology companies across Europe with global ambition. The team behind 42CAP are seasoned entrepreneurs turned investors. With Hybris and eCircle they previously built two of Europe’s largest B2B software companies culminating in $1.6b cash exits to SAP and Teradata respectively.
The 42CAP credo “Peers amongst Entrepreneurs” is reflected in our investment approach, with which we support B2B data- and technology-driven business models, product- focused founders and sustainable venture development.
More information at: https://42cap.com