In partnership with

Evaluation and Real-Time Governance for AI Agents

Check your Description

Principled Intelligence builds evaluation and governance tools that keep AI agents within an organization's own policies.

Its Spectral product, in private beta, runs thousands of autonomous tests across threat categories such as prompt injection, hallucination and personal data leaks before agents reach users.

Orbitals, released as an open source Python guardrails library under Apache 2.0, makes sub-second allow or block decisions on every response and attaches a rationale to each one.

Every decision is logged as audit-ready evidence for engineering, legal and compliance teams, with multilingual support for mission-critical deployments.

The Rome-based company was founded by researchers Simone Conia and Edoardo Barba and raised a 1.85 million euro pre-seed round with participation from CDP Venture Capital.

Market Segments:

AI SecurityAI Governance

Categories:

AI GuardrailsAI Red TeamingModel TestingAI Governance