AI agents are quickly moving from tools that retrieve information to actors that take actions: making purchases, opening accounts, moving money, and interacting directly with businesses on behalf of people and companies.
That shift creates a new set of problems. The infrastructure built to understand and manage human customers wasn’t designed for autonomous software acting on their behalf. Businesses need to know not just that an agent can be trusted, but whether that agent should be allowed to take a particular action, under a particular set of policies, at a particular moment.
That’s the problem Beltic is building to solve.
Beltic is building an agent verification stack that helps enterprises safely accept and interact with agentic traffic. Their vision starts with “Know Your Agent,” or KYA: connecting agents back to verified people or businesses, translating existing enterprise policies into rules agents can operate within, and continuously evaluating activity as agents interact with a business.
We invested because we believe this will become a critical piece of infrastructure as the internet shifts from human-driven to increasingly agent-driven commerce.
A New Identity Layer for the Agent Economy
The existing internet has well-established systems for determining whether to trust a person or business. KYC, KYB, fraud tooling, authentication, and risk engines collectively answer some version of: Should we allow this user to do this?
Agents complicate that question.
An enterprise may trust the person behind an agent, but it still needs to know what the agent is authorized to do, whether its behavior matches that authorization, and how much risk to accept in any individual interaction.
Beltic is building specifically for that future. Importantly, the company isn’t asking enterprises to rip out their existing risk infrastructure. Instead, Beltic can make those systems agent-ready by connecting existing identity and policy frameworks to a new class of machine customers.
Built for a Fragmented Ecosystem
The agentic ecosystem is developing quickly, and there is unlikely to be a single dominant protocol through which every agent interacts with every business.
We believe that makes Beltic’s protocol-agnostic approach particularly compelling. By sitting across emerging agent protocols and transaction surfaces, Beltic has the opportunity to become the place enterprises understand who an agent represents, what it is permitted to do, and whether its ongoing behavior remains trustworthy.
Over time, that position can become increasingly valuable as more commerce moves from humans clicking buttons to software acting autonomously on their behalf.
The Right Team for the Problem
Beltic is led by Isha Bhatnagar and Mike Allan, two founders who have spent their careers dealing with the exact identity, compliance, payments, and risk infrastructure now being rethought for AI agents.
Isha has spent her career building regulatory and identity products at Coinbase. Mike previously built a banking infrastructure company in Brazil through acquisition and went on to help launch a bank inside one of the country’s largest financial institutions.
That combination of entrepreneurial experience and deep domain expertise stood out to us. This is a market where understanding enterprise risk and compliance infrastructure matters just as much as understanding AI.
The agent economy is still early. Exactly how agents transact, which protocols win, and which use cases scale first will evolve significantly from here.
But we have conviction in the underlying shift: agents are becoming actors for transaction, engagement, communications, and much more, and enterprises will need infrastructure purpose-built to safely do business with them.
We’re excited to partner with Mike, Isha, and the Beltic team as they build that infrastructure.
© 2026 Restive®, Inc.
