Agentic AI often is the buzziest new frontier in enterprise tech, however for safety leaders, it’s rapidly changing into a nightmare. Austin-based Eve Safety has raised $3 million in seed funding to deal with the surge of “shadow AI brokers” making unsanctioned selections, shifting delicate knowledge, and creating new vulnerabilities inside companies.
The spherical was led by LiveOak Ventures, with participation from Tau Ventures. This funding will help additional product improvement and enlargement of Eve Safety’s go-to-market technique.
Introducing Eve
The corporate additionally introduced the trade’s first Agent-in-the-Loop (AITL) system, referred to as Eve. It’s designed to detect anomalies, cluster requests, assess dangers, and interrogate different brokers in order that CISOs and division heads don’t should.
Eve screens brokers and their interactions with each other, with people, and with vital enterprise methods, taking motion as wanted.
Eve operates constantly below comparable situations, permitting people to supervise selections with reliability. It follows clear, rule-based logic, offering the identical end result every time it receives the identical enter.
“Agentic AI is already contained in the enterprise, whether or not CISOs and different leaders approve it or not. Shadow AI brokers are making selections, shifting knowledge, and creating vulnerabilities that right this moment’s safety instruments merely can’t see. Proper now, most safety leaders are flying blind,” says Creighton Hicks, Accomplice at LiveOak Ventures. “Eve Safety permits safety groups to realize visibility and management over the know.
Filling the hole in AI safety
Based by CEO Nadav Cornberg, CRO Sharon Eilon and CTO Amit Eliav, Eve Safety is an agentic AI observability and coverage enforcement platform. It makes use of the trade’s first Agent-in-the-Loop (AITL) to make sure the safety of AI brokers interacting with an organization’s most vital enterprise methods.
EveGuard consists of ‘Intent- and Knowledge-Based mostly Coverage Enforcement’ to deal with points with English-language insurance policies that may result in misunderstandings as a consequence of translation and cultural variations.
Eve Safety’s platform can outline and implement insurance policies that think about the intent and knowledge in requests. These options will not be current in present safety merchandise, which is essential for managing identified and unknown agentic AI in a enterprise setting.
“CISOs right this moment are confronted with each nice alternative and nice duty,” mentioned Nadav Cornberg, co-founder and CEO, Eve Safety. “We assist them to innovate with AI with confidence and belief that their organisation’s crown jewels – the info and methods that matter most to their enterprise – are protected. Our implementation and mixture of observability, coverage and AITL permits CISOs and division heads to belief the exercise of their AI brokers. Our help for API, MCP and A2A protocols permits us to go deeper than rivals to watch, detect and react to safety considerations rapidly.”
Finish-to-end observability and management
Eve Safety’s platform combines observability, coverage enforcement, and AITL oversight into one system:
- Agent-in-the-Loop (AITL): Eve observes and manages agent exercise throughout financials, HR, and gross sales methods, industries the place compliance and auditability are vital.
- Deep protocol help: The platform helps Mannequin Context Protocol (MCP) and Agent-to-Agent (A2A) requirements for monitoring coordination chains and exterior instrument entry.
- Coverage enforcement: Goes past English-language-based guidelines to interpret intent and knowledge context, decreasing blind spots attributable to translation or cultural nuance.
- Frictionless deployment: Requires no extra infrastructure or plug-ins, delivering rapid worth.
