Enterprise AI Governance Research

Foundational research, frameworks, and analysis for governing enterprise AI, autonomous agents, and control-plane adoption in regulated organisations.

Helixar Research

Research, frameworks, and analysis for enterprise AI governance, compliance, and agentic risk.

PinnedSpecificationv0.1IETF draft-01Apache 2.0

HDP v0.1 Specification

Human Delegation Provenance Protocol

The normative specification cited as [HDP-SPEC] by the IETF Internet-Draft draft-helixar-hdp-agentic-delegation-01. It defines the token structure, the Ed25519 root and hop signing procedures, the chain integrity rules, the seven-step offline verification pipeline, and the HTTP transport bindings.

Read the specification

Status: Draft, open for comment
Signing: Ed25519, RFC 8032
Canonicalization: RFC 8785
arXiv: 2604.04522

Flagship framework

Enterprise AI Governance Framework

A practical, jurisdiction aware framework that turns AI principles into decision rights, controls, and evidence a board and an auditor can test.

Flagship report

State of Enterprise AI Governance in ANZ 2026

The flagship Helixar research report on enterprise AI governance in Australia and New Zealand, covering regulation, standards, operational resilience, privacy, agentic AI, readiness, board assurance, and control-plane evidence.

Enterprise AI Governance Explained

The Helixar research primer for enterprise AI governance: accountability, risk, compliance, auditability, and control-plane thinking.

Incident analysis

Governing Agentic AI: The OpenAI and Hugging Face Incident

What the OpenAI and Hugging Face incident reveals about bounding agent authority and producing accountable evidence in enterprise deployments.

Enterprise AI Governance Best Practices

Practical, grounded practices for governing enterprise AI well, framed as patterns to adapt rather than a checklist to copy.

Programme design

AI Governance Programme Design

Designing an AI governance programme that actually operates, with real authority, adequate resourcing, and a rhythm that outlasts the launch.

Enterprise AI Governance Roadmap

A phased roadmap for standing up AI governance that lasts, sequencing visibility before control and control before assurance.

Control objectives

AI Governance Control Objectives

The control objectives that make AI governance testable, bridging policy and the specific controls that meet them.

Reporting framework

AI Governance Reporting Framework

Reporting that lets the board govern AI, concise at the top and deep underneath, so any figure can be traced to the record behind it.

Evidence framework

AI Governance Evidence Framework

The evidence that turns governance from a narrative into a fact, retained across the lifecycle and strong enough to survive challenge.

Audit framework

AI Governance Audit Framework

Making AI governance auditable, and auditing it well, so an independent function can test whether controls are designed and operate.

Assurance framework

Enterprise AI Assurance Framework

Turning governance activity into credible assurance, so trust rests on independent challenge and evidence rather than management assertion.

Vendor governance

AI Vendor Governance

A lifecycle view of governing AI vendors and model providers, complementing the risk lens with the management of the relationship itself.

Operational risk

AI Operational Risk Management

AI as an operational risk, not only a model risk, brought into the resilience discipline that already governs critical processes.

AI Model Risk Governance

Model risk management, updated for AI and generative models, so the risk that a model is wrong, misused, or misunderstood is governed across its life.

Third-party risk

Third-Party AI Risk Management

Governing the AI risk you buy, not only the AI you build, across the whole life of the third-party relationship.

Enterprise AI Risk Register Framework

An AI risk register that makes AI risk visible, owned, and trackable, and connects it to the enterprise risk management the board already trusts.

Metrics and KPIs

Enterprise AI Governance Metrics and KPIs

Measuring AI governance so a board can see whether it works, using a small set of indicators that are traceable to underlying records.

Oversight models

AI Governance Oversight Models

Proportionate oversight that scales with AI risk, designed for the speed of agents rather than the pace of a review.

Decision accountability

Enterprise AI Decision Accountability

Keeping a named human answerable for AI influenced decisions, with oversight that is real and evidence that can reconstruct what happened.

Accountability model

AI Governance Accountability Model

The accountability chain that gives AI governance authority, and the decision rights that make it operate.

Policy management

AI Governance Policy Management

The policy lifecycle that keeps AI rules current, enforceable, and evidenced, and the taxonomy of policies an enterprise needs.

Reference model

Enterprise AI Governance Reference Model

The architecture diagram for enterprise AI governance: the layers that must connect for governance to hold together, and how any control framework hosts inside them.

Capability assessment

Enterprise AI Governance Capability Assessment

Turning the capability model and assessment methodology into a scored current versus target view and a plan the enterprise can act on.

Assessment methodology

Enterprise AI Governance Assessment Methodology

How to run a defensible, evidence led assessment of enterprise AI governance, end to end, so results can be trusted and retested.

Capability model

Enterprise AI Governance Capability Model

The EAGCM: the governance capabilities that make up a working enterprise AI governance function, and how to judge whether each one exists.

Operational risk

APRA CPS 230 and AI Governance

A practical research guide to interpreting AI governance through the lens of APRA CPS 230 operational risk management.

Information security

APRA CPS 234 and Autonomous AI

A practical research guide to information security governance for autonomous AI in APRA-regulated environments.

Readiness model

Enterprise AI Readiness Assessment

A practical research guide for assessing whether an enterprise is ready to adopt, govern, monitor, and evidence AI systems and agents.

Management system

ISO/IEC 42001 for Enterprise AI Governance

A practical research guide to applying ISO/IEC 42001 as an enterprise AI governance management system.

Framework mapping

NIST AI RMF for Enterprise AI Governance

A practical research guide to using the NIST AI RMF as an enterprise AI governance system, from Govern, Map, Measure, and Manage to operational evidence.

NZ Privacy Act and Enterprise AI Agents

A practical research guide to privacy-by-design controls for enterprise AI agents in New Zealand.

Critical infrastructure

AI Governance for Critical Infrastructure

AI governance for energy, water, transport, telecoms, financial market infrastructure, and essential services where resilience matters.

AI Governance for Government Agencies

How public sector teams can govern AI with transparency, record keeping, human oversight, privacy, procurement discipline, and public trust.

AI Governance for Healthcare

Governance guidance for hospitals, health insurers, digital health providers, administrators, and health technology teams using AI.

AI Governance for Insurance

How insurers can extend risk discipline into AI-enabled underwriting, claims, fraud, pricing, service, document processing, and modelling.

AI Governance Maturity Model

A five-level Helixar maturity model for boards, CISOs, risk teams, and compliance leaders assessing enterprise AI governance readiness.

Operating model

AI Governance Operating Model

How enterprises should organise AI governance across the board, executive teams, risk, security, privacy, legal, audit, engineering, and business owners.

Enterprise AI Risk Management

A risk-management view of AI governance: model risk, privacy risk, security risk, operational risk, third-party risk, autonomy risk, and accountability risk.

1 / 7

Infrastructure for the agentic layer, published openly

Helixar Labs ships open protocols and open-source tools under Apache 2.0 / MIT, freely usable in production without licensing or vendor lock-in.

HDP

Human Delegation Provenance protocol, additive delegated-authorization and provenance across agent hops. In the official Google Gemma cookbook.

HDP-P

HDP for Physical AI, embodied agents, robots, vehicles. Public Hugging Face demo.

ReleaseGuard

Artifact policy engine: secrets scanning, SBOM, Sigstore signing, SLSA attestation.

Sentinel

26-rule MCP security scanner with SARIF output for GitHub Code Scanning.

Unpinched

Point-in-time scanner for PinchTab and agentic browser-bridge artifacts.

MCP Security Checklist

Community-maintained 7-domain hardening framework for production MCP.

Published research and standards work

Hosted on arXiv, SSRN, and the IETF datatracker. Click through for the full text and supplementary material.

How we publish, and the bar we hold

Editorial guidelines: Every draft begins as a structured threat record and is reviewed by a member of the Helixar Research Team before merge, sources verified, entities and CVEs checked against authoritative records, original analysis required.
AI-assist disclosure: The drafting pipeline uses large language models; every published article is the product of an automated draft and a human review. The byline is Helixar Research Team. We do not invent named author personas.
Sources & allowlist: A minimum of three references resolving to original disclosing sources, NIST NVD, CISA, MITRE, vendor advisories, the IETF datatracker. We do not cite aggregators as primary sources.
Quality bar: A pre-merge lint requires 800+ words for threat pieces, three external references, every CVE validated against NIST NVD JSON, and original analysis above 80% of the body.
Corrections policy: We acknowledge reports within two business days, publish a correction at the top of the affected article, update its dateModified, and never silently edit to remove errors.
Editorial independence: We write about competitors, and about open standards we are commercially involved with (HDP, HDP-P, ReleaseGuard), and say so where it matters. No paid placements; vendors do not pre-review coverage.

Published research and threat intelligence on this site are informational and do not constitute professional security, legal, or compliance advice. References to third-party platforms and standards are for technical context only; Helixar is not affiliated with or endorsed by them unless explicitly stated.

Spotted an error? Write to [email protected] , we correct in the open.