Mintive raises $7M to secure AI agents with identity and permissions

Recent funding activity highlights growing demand for governance tools in enterprise AI deployments. A startup in this space has secured additional capital from Cowboy Ventures and other backers, pushing its total funding to a figure reported by industry sources. The company’s platform aims to address a critical challenge: ensuring AI agents operate within defined boundaries while maintaining visibility into their actions.

The startup’s technology reportedly processes millions of agent interactions daily, serving a range of enterprise clients across sectors like fintech, education, and healthcare. Its solution focuses on creating distinct identities for agents, limiting their access to only what’s necessary for their tasks, and maintaining detailed records of their activities. This approach is designed to prevent agents from inheriting overly broad permissions—an issue that arises when they operate under employee credentials.

“Agents are becoming integral to company operations, and they require the same level of oversight as human employees,” a company spokesperson told a tech publication. The platform’s architecture separates agent identities from user credentials, allowing administrators to manage permissions independently. Agents authenticate through their own credentials, receive short-lived tokens, and can be revoked without disrupting other users or workflows. This structure also aims to solve continuity challenges, ensuring business processes don’t rely indefinitely on individual employees’ access.

The system combines tool access and model interaction layers, using a protocol to standardize connections between AI applications and external tools. Governance controls determine which tools an agent can access, with administrators able to disable specific functions or apply uniform policies across AI clients. Every interaction is logged, including details like the user, tool, destination, and guardrail outcomes, and these logs can be exported to security monitoring systems or streamed via industry-standard protocols.

Additional layers screen for risks like prompt injection or exposed data, and the company states its platform meets SOC 2 Type II compliance standards. However, the effectiveness of these controls depends on how organizations configure their policies and integrations.

Customer use cases illustrate the practical challenges of agent governance. One fintech client reportedly uses the platform to streamline workflows involving multiple internal tools, with the company claiming efficiency gains in certain scenarios. An educational institution employs it as a centralized gateway for tool access, while also offering agents in collaboration platforms like Slack and Teams. These agents combine tool access with features like long-term memory and scheduled tasks.

The founding team brings experience from major tech companies, including roles in AI research and infrastructure development. Investors include institutional firms and individual backers with backgrounds in enterprise software and AI.

This funding reflects broader trends in enterprise AI, where operational layers are emerging to define agent permissions, authority delegation, and audit trails. While the market for such solutions is still developing, the reported adoption suggests these concerns are already critical for organizations deploying agents at scale.

Earlier coverage of a similar startup focused on monitoring AI agents posing as users, highlighting the risks of unchecked agent activity. The latest funding suggests a shift toward proactive governance, embedding identity and permissions into agent workflows from the outset.

The key question remains whether enterprises will adopt centralized gateways like this one or continue managing agent permissions in a decentralized manner. While the startup’s customer base and usage metrics indicate traction, the broader market for AI agent governance is still taking shape. What’s clear is that traceability and control are no longer optional—they’re essential for companies running agents in production.

Sources: unite.ai