Flagship annual studies and benchmark reports
Original Clariantix research on the state of enterprise AI governance, risk exposure, and the emerging AI trust economy.
An external reviewer is not automatically an independent evaluator. What access and freedom did the reviewer actually have?
An evaluator can be outside the company yet constrained by limited access, selected evidence, contractual controls or publication restrictions. Independence must be demonstrated, not inferred from the label.
Your Policy Says the Agent Cannot Do It. What Actually Stops the Agent?
Policy defines what an AI agent is allowed to do. Runtime controls determine what the agent is actually capable of doing when instructions fail, context changes or safeguards are bypassed.
Your AI Provider May Not Call It an Incident. Should You?
A provider's explanation of an AI event is an important input, but it should not determine the customer's governance response. Organizations need an independent way to classify control failures, exposure and required action.
Your Company Authorized the AI Agent. The Other Party Did Not.
Internal approval does not grant an AI agent permission to act on another organization's platform. Enterprises need controls for identity, platform terms, credentials, data capture and counterparty consent.
Lack of Evidence Is Not Evidence of AI Safety
Risk reviews should not convert missing evidence into confidence. When impact could be severe or irreversible, uncertainty should increase scrutiny, containment and the burden of proof for deployment.
State of Enterprise AI Governance 2026
Artificial intelligence has rapidly evolved from an experimental technology into core enterprise infrastructure. Organizations across every sector are deploying generative AI, machine learning, intelligent automation, and predictive analytics to improve productivity and decision-making. Yet while AI adoption has accelerated dramatically, governance maturity has not kept pace. Many organizations now operate dozens of AI systems with limited visibility, inconsistent ownership, and fragmented oversight — producing an emerging trust gap. This report examines the current state of enterprise AI governance and identifies the foundational capabilities organizations should establish to build trustworthy, transparent, and accountable AI programs.
Enterprise AI Risk Benchmark Study
Enterprise AI introduces a new category of organizational risk. Unlike traditional software, AI systems can evolve, generate unexpected outputs, and influence business decisions in ways that are difficult to predict. Understanding these risks is the first step toward managing them. This benchmark study examines the six major AI risk domains, the persistent threat of shadow AI, third-party accountability gaps, and the practical priorities for risk reduction.
The AI Trust Economy
Trust has always been one of the most valuable business assets. In the age of artificial intelligence, trust becomes even more important. Customers, regulators, employees, and investors increasingly want to know: can this organization be trusted to use AI responsibly? This report explores how AI changes the nature of trust, the five pillars that sustain it, and why trust is becoming the currency of the next digital economy.
The EU AI Act is the world's first comprehensive horizontal AI regulation. It categorizes AI systems by risk (prohibited, high-risk, limited-risk, minimal-risk) and imposes obligations on providers, deployers, importers, and distributors operating in or selling into the EU.
ISO/IEC 42001 is the first international management-system standard for artificial intelligence. It defines requirements for establishing, implementing, maintaining, and continually improving an AI management system (AIMS), with certifiable controls across governance, risk, lifecycle, and operations.
The NIST AI Risk Management Framework is a voluntary, widely adopted framework for managing risks across the AI lifecycle. It is organized around four functions — Govern, Map, Measure, Manage — and is paired with the Generative AI Profile for foundation-model risks.