AI Agent Security Startup Beltic Raises $7.3M Seed Funding

Silicon Valley-based startup Beltic has successfully closed a $7.3 million seed funding round to build a global infrastructure for verifying the identities of AI agents. The funding was led by prominent venture capital firm Norwest Venture Partners, with participation from Restive Ventures and the Oxford Seed Fund. This capital infusion arrives less than a year after Beltic secured $1.5 million in pre-seed funding, underscoring strong investor confidence in its mission to secure the burgeoning agent economy.

Beltic is tackling a critical challenge: existing fraud prevention and authentication systems are designed for human behavior, not for the autonomous actions of AI agents. As artificial intelligence agents are increasingly tasked with executing transactions, purchases, and payments on behalf of individuals and businesses, a new paradigm for security is essential. The company's platform aims to provide a robust solution by analyzing AI-driven transactions across three key layers: identifying the entity behind the agent, confirming the agent's authorization for the specific action, and ensuring the executed task aligns with the user's original intent.

The first layer of verification can range from simple checks, such as verifying a registered email address with a merchant, to more stringent measures like biometric authentication or document submission for high-value transactions. The subsequent layers delve into the agent's operational protocols and the fidelity of its actions compared to the initial human directive. This is crucial, as AI agents can sometimes misinterpret instructions or "hallucinate" during complex processes, potentially leading to unauthorized or unintended outcomes. Beltic's system flags or blocks such discrepancies, safeguarding against errors and malicious activity.

Founded by Brazilian entrepreneur Mike Allan, who previously founded the acquired startup Atar, and Isha Bhatnagar, a former Coinbase executive with extensive experience in customer verification, Beltic is strategically positioned to address the complexities of the agent economy. Allan notes that while the infrastructure for AI agents is rapidly developing, a lack of standardization across payment networks and technology providers creates significant integration hurdles for businesses. Companies like Visa, Mastercard, Amex, and Google are each developing proprietary protocols, making it challenging for retailers to seamlessly integrate AI agent transactions.

Beltic's multi-protocol approach offers a unified solution, allowing businesses to interact with various AI agent protocols without needing to build custom integrations for each. This capability is particularly vital as the adoption of AI agents for tasks beyond simple e-commerce browsing, such as complex financial services, insurance procurement, and automated payment processes, is expected to grow significantly. While not anticipating a complete shift to AI-driven transactions in the immediate future, Beltic anticipates a substantial portion of automated, high-risk decisions and payments will be managed by AI agents, necessitating robust security frameworks.

The market for AI-driven automation and agent-based transactions is projected for substantial growth. As more sophisticated AI applications move from concept to execution, the demand for specialized security and identity verification solutions like Beltic's will intensify. The company's focus on bridging the gap between human intent and AI execution, coupled with its multi-protocol compatibility, positions it as a key player in enabling the secure and scalable deployment of AI agents across various industries.