Singulr AI is an AI governance and security company. It gives enterprises one place to see the AI systems and agents in use across the organisation, decide what each is allowed to do and enforce those decisions while the AI is running, with evidence that the rules were followed.
The platform covers three ways AI enters a business: apps and agents built in house, public AI tools that staff sign up for, and AI features embedded in SaaS applications. It is built for security, risk and compliance teams in financial services, insurance, healthcare, software and hi-tech manufacturing.
At A Glance
| Vendor | Singulr AI |
|---|---|
| Category | AI governance and compliance |
| Identities covered | AI agents, MCP servers, AI services and the users who access them |
| Agentic AI | Agent Pulse agent discovery, context graph, permission boundaries and endpoint agent governance |
| Customers | KVC Health Systems, Delta Dental Plans Association |
| Certifications | ISO/IEC 27001:2022, SOC 2 Type 2 |
| Funding | Seed round led by Nexus Venture Partners and Dell Technologies Capital |
| Market classification | AI Governance & Compliance; also Agentic AI Identity |
What Singulr Does
Most organisations now have AI policies, but few can show those policies are applied to the tools and agents people actually use. Singulr’s mission is to “make AI governance enforceable”. Teams set intent, ownership and risk thresholds for each AI system or agent, and a natural-language policy engine turns that intent into boundaries that are checked at runtime. Governance, runtime enforcement, security and independent assurance are joined in a closed loop, so findings from monitoring feed straight back into policy.
Products
Singulr Runtime Governance
Runtime Governance is where risk and compliance teams set AI policy. Powered by the Singulr Pulse risk engine, it runs red teaming against OWASP LLM Top 10, NIST and MITRE scenarios so weaknesses show up before attackers find them.
Singulr Runtime Security
Runtime Security watches live AI interactions for attacks. It detects prompt injection, jailbreaks and unauthorised data exfiltration in live AI traffic, and keeps forensic traceability of each interaction for investigation.
Singulr Runtime Control
Runtime Control is the enforcement layer for organisations that need to decide which AI tools may be used and how. It can:
- Block unapproved AI tools at service level
- Tell personal accounts apart from enterprise accounts and treat them differently
- Enforce agent permission boundaries at the moment an agent acts
- Detect and redact PII and PHI, with custom classifiers for other sensitive data
- Apply domain-level controls and block file uploads
- Map applications and AI services into a topology view
Singulr Agent Pulse
Agent Pulse is aimed at teams that cannot say how many AI agents are running, on which platforms or with what access. It discovers agents wherever they run, along with the MCP servers and tools they depend on, keeps a running risk score for each and records what every agent did.
How It Works
The platform works as a sequence of discovery, assessment, policy and enforcement.
| Component | What it does |
|---|---|
| Discovery and context graph | Singulr finds AI services and agents and builds a context graph of their tool connections, data access, MCP servers and permissions. |
| Policy engine | Natural-language governance intent becomes enforceable policy based on agent type, data sensitivity, tool access and scope. |
| Enforcement points | Policy is applied inline to AI services and agent execution and, through Agent Pulse, on endpoints. Each decision can permit, restrict, warn, redact or raise a security ticket, and the Singulr Pulse risk engine feeds its scores into those decisions. |
| Connected tools | Signals and actions flow to Splunk, CrowdStrike, Zscaler, Microsoft Purview, Code42 and ITSM tooling, with native SIEM, EDR, SSE and SASE integration. Okta and Microsoft Entra supply identity context. |
Identities And Systems Covered
| AI agents | Agents on any platform, with their tools, data access and permissions. |
|---|---|
| Endpoint AI agents | Claude Code, Cowork, Cursor, ChatGPT Desktop and MCPs. |
| MCP servers | Discovered and mapped alongside the agents that call them. |
| Users and accounts | Personal and enterprise accounts handled separately. |
| AI services | Public AI tools and AI embedded in SaaS applications. |
Agentic AI Capabilities
Agent Risk Scoring And Drift
Each agent is scored continuously on its model access, MCP configurations and tool connections. Agent Pulse also tracks drift in configuration and runtime behaviour, so a change to an agent’s setup raises its risk.
Permission Boundaries
Excessive agent permissions are flagged before they can be exploited, and boundaries are checked at execution. Enforcement steps up when an interaction moves into higher-risk territory.
Endpoint Agents
Runtime governance reaches coding and desktop agents on employee machines, so tools such as Claude Code and Cursor fall under the same policies as server-side agents.
Company Background
| Headquarters | Palo Alto, California, United States. |
|---|---|
| Leadership | Shiv Agarwal (founder and CEO) and Abhijit Sharma (co-founder and CTO). |
| Funding | A reported $10 million seed round led by Nexus Venture Partners and Dell Technologies Capital in February 2025. KAAJ Ventures and individual investors including Dheeraj Pandey also invested. |
| Recognition | Included in the Gartner Market Guide for Guardian Agents. |
Note: This overview is compiled from Singulr AI’s publicly available information as of 3 November 2026. It is provided for information only. NHI Mgmt Group does not endorse any vendor, and nothing on this page represents the opinions of NHI Mgmt Group.