Operant AI, a Silicon Valley-based security startup, recently started operations in India. Now, the company has unveiled its first product for Indian businesses. It’s called AI Gatekeeper, which aims to secure AI systems, including autonomous agents and multi-agent workflows, in complex environments like Kubernetes, hybrid cloud, and edge deployments.
The adoption of AI agents is rising among Indian enterprises. A recent report by Deloitte revealed that over 80% of Indian enterprises are exploring autonomous agents, with 50% working on setups requiring minimal human oversight. However, as AI adoption increases, the risks associated with rogue AI behavior, data leakage, and model poisoning are also rising.
If we go by the company’s shared press release, the AI Gatekeeper will try to address those issues by offering real-time runtime protection. It includes features like agent trust scoring, access controls, and threat blocking for protocols like Model Context Protocols (MCPs) and Non-Human Identities (NHIs).
Third-Party Dependence Adds Security Complexity
According to Operant, Indian companies often rely on third-party vendors for AI deployment. This creates data governance and security challenges.
“The AI that we are now securing is a completely new beast compared to even two years ago,” said Vrajesh Bhavsar, Operant AI’s CEO and co-founder. He added that today RAG applications to AI Agents to AI Inference systems operate at a completely new scale, because of which AI can’t be secured in isolation. AI Gatekeeper can bring Operant’s unique defensive capabilities to everywhere customers are deploying AI, alongside critical new capabilities for protecting sensitive data and the rest of the application environment from the new attack surface that is being fueled by rapid Agentic AI adoption.”
The AI that we are now securing is a completely new beast compared to even two years ago. Today RAG applications to AI Agents to AI Inference systems operate at a completely new scale, because of which AI can’t be secured in isolation. AI Gatekeeper can bring Operant’s unique defensive capabilities to everywhere customers are deploying AI, alongside critical new capabilities for protecting sensitive data and the rest of the application environment from the new attack surface that is being fueled by rapid Agentic AI adoption.
Vrajesh Bhavsar, Operant AI’s CEO and co-founder
The company aims to provide end-to-end visibility and threat defense for all AI workloads, including platforms like Amazon Bedrock, Azure, Databricks, and Snowflake. The goal is to protect where the data resides, not just where it’s processed.
What AI Gatekeeper Offers
AI Gatekeeper comes with several new capabilities:
- Runtime protection across cloud and hybrid environments beyond Kubernetes.
- Live inventories of all AI agents, tools, and models used across an organization.
- Threat modeling and analytics, including OWASP mappings for risks like prompt injection and API key leakage.
- Defense for agent supply chains, unauthenticated agents, and least-privilege execution.
- Support for Model Context Protocols and Non-Human Identities, expanding access and identity protections.
We are seeing three trends happening right now: First, incredibly fast deployment of AI models and AI Agents for novel use cases; second, adoption of new platforms beyond the traditional cloud providers; and lastly, the requirements and responsibilities for security, infrastructure, data infosec and AI converging. Operant has built a solution that helps teams protect their most business-critical transformations, while AI Gatekeeper makes it possible for AI-native teams to innovate securely at a completely new pace.
Raj Yavatkar, CTO of Juniper Networks
Securing AI Agents is a critical priority for AI-native companies because you can’t hand off that level of autonomy at scale to these systems without appropriate controls in place.
Martin Choluj, CISO of Clickhouse
