London-based Pallma AI has announced the completion of a $1.6 million pre-seed funding round, led by Marathon Venture Capital. The goal is to develop a security platform specifically designed to protect enterprise systems based on artificial intelligence agents.

The security challenge in AI agents

The adoption of AI in the enterprise is growing rapidly, with large language models (LLMs) and autonomous AI agents moving from pilot programs to core operational roles. This transition increases reliance on systems capable of operating with limited human supervision. However, this evolution introduces new security and risk management challenges.

Traditional security tools are not designed to address the probabilistic and adaptive behavior of autonomous systems, which can be vulnerable to prompt manipulation, unintended data exposure, or excessive autonomy. When these systems process sensitive data or are authorized to take actions, errors or misuse can have a significant impact on security and operations.

Pallma AI's solution

Pallma AI aims to solve these problems by providing a centralized layer of security, visibility, and control for AI agents. The platform integrates with existing enterprise technology stacks, collects application and system data, and uses AI models to monitor, detect, and mitigate risks in real time. This includes protection against prompt injection, unintended data exposure, and unsafe agent behavior.

The platform combines real-time threat detection with actionable guidance and automated response capabilities, going beyond a purely passive security approach. The company plans to use the new funding to expand its team, accelerate product development, and support the growing demand for its AI-powered enterprise security platform.