What is the difference between OneTrust AI Governance and Arcjet?
OneTrust AI Governance inventories AI systems from cloud AI platforms such as Amazon Bedrock, Microsoft Foundry, Google Vertex and Databricks, runs risk assessments against regulations, and applies guardrails on the platforms it integrates with. Arcjet builds the AI inventory from managed devices, coding agents and the agents you build, and enforces decisions there.
The two products look at different places. OneTrust connects to the cloud platforms where teams build and host models and agents, and wraps them in the privacy, risk and compliance workflows that OneTrust customers already run. Arcjet observes the AI that people and agents use on laptops and in code: which AI providers each person reaches, what each coding agent does, and what the agents you build send and call. If your AI is mostly built on Bedrock, Foundry, Vertex or Databricks and your governance team already works in OneTrust, OneTrust covers that estate. If your risk sits in coding agents and employee use of AI providers, Arcjet covers ground that OneTrust's page doesn't describe. The two fit together, with Arcjet's observed inventory feeding the OneTrust register.
Arcjet publishes this comparison. OneTrust details come from OneTrust's public AI Governance product page and pricing page, reviewed on November 16, 2026. Products change, so check the linked sources before you decide.
How do OneTrust AI Governance and Arcjet compare at a glance?
OneTrust AI Governance governs AI built on cloud AI platforms, and Arcjet governs AI used on devices and in agents. The following table compares the two products across 10 areas.
| Area | Arcjet | OneTrust AI Governance |
|---|---|---|
| Category | Runtime-led AI governance and AI agent runtime security | Governance, risk and compliance (GRC)-led AI governance, part of OneTrust's privacy and risk platform |
| What it inventories | AI systems: applications and agents with the model providers and Model Context Protocol (MCP) servers they use, and AI providers reached from managed devices | AI systems, models, agents, datasets, vendors, projects and use cases |
| Discovery sources | Discovery on macOS and Windows devices, coding-agent hooks, software development kits (SDKs), the Claude Compliance API and OpenTelemetry | Amazon Bedrock, Microsoft Foundry, Google Vertex and Databricks Unity Catalog, plus Jira and Palo Alto Networks |
| Employee AI use | Each connection to an AI provider from a managed device, attributed to a person | The page mentions detecting shadow AI but doesn't describe browser, device or endpoint coverage |
| Coding agents | Hooks in Claude Code, GitHub Copilot, Cursor, OpenAI Codex and Muse Code decide each action | Not described on the product page |
| Risk assessment | Risk tier and EU AI Act role per AI system | Assessment templates, automated risk tiering, and reassessment when a model, agent or dataset changes |
| Runtime enforcement | Guard decisions on tool calls and prompts, hook decisions, and provider blocking on devices | Guardrail Enforcement, Guardian Agents and MCP policies, with coverage that varies by platform |
| Sensitive data | Personally identifiable information (PII), secrets and protected health information (PHI) detection in Guard and coding-agent policies | AI Guard Python SDK; PII findings from Bedrock and Foundry interaction logs |
| Evidence output | Evidence page listing every requirement as supported or not covered; JSON and CSV export; push to Vanta and Drata | Evidence from assessments, approvals, attestations, policies, telemetry and enforcement actions |
| Pricing basis | Platform plans, with Discovery through a demo | Based on admin users and AI inventory, quoted on request |
What is OneTrust AI Governance?
OneTrust AI Governance is OneTrust's product for inventorying AI, assessing its risk against regulations, and applying policy to it. It sits beside OneTrust's privacy, data and third-party risk products, so a team that already runs privacy impact assessments or vendor reviews in OneTrust can run AI governance in the same place.
The product page says OneTrust continuously discovers AI across changing environments and inventories AI systems, models, agents, datasets, vendors, projects and use cases. Each entry can record ownership, life-cycle status and dependencies. Amazon Bedrock, Microsoft Foundry, Google Vertex and Databricks Unity Catalog are supported out of the box, and OneTrust also integrates with Jira and Palo Alto Networks. The page doesn't describe how discovery works beyond those integrations.
Risk assessment uses built-in templates for the EU AI Act, the National Institute of Standards and Technology (NIST) AI Risk Management Framework (AI RMF) and ISO/IEC 42001. OneTrust tiers risk automatically by use case, system, component, deployment context or data sensitivity, and reassesses when a model, agent, dataset or usage pattern changes materially. AI Policy Manager defines acceptable use, scope and required controls, and OneTrust tracks policy versions, exceptions and remediation.
For enforcement, the page describes Guardrail Enforcement, which takes technical actions where AI runs: block, allow, redact, escalate or route an action by policy. OneTrust supplies seeded policies and generative AI guardrails aligned to NIST AI RMF and the EU AI Act, with starting points for Databricks and Amazon Bedrock. Agents are registered with a purpose, permissions and allowed actions, and the page says OneTrust enforces MCP policies with audit logs and revalidates agents when their tools or data change. Guardian Agents provide runtime oversight of agent activity and apply controls at the point of action.
The AI Guard SDK is a Python SDK that finds sensitive data in prompts and responses and blocks or redacts it before it reaches the model or a downstream system. OneTrust also analyzes Bedrock and Foundry interaction logs for PII, and monitors quality, safety, performance and guardrail signals for models and agents.
The page's FAQ says platform coverage varies by integration. Amazon Bedrock provides usage, token, log analysis and PII findings. Microsoft Foundry provides evaluation visibility and guardrail enforcement. Check which capabilities apply to the platforms you run.
OneTrust's pricing page says AI Governance is priced on admin users and AI inventory, and quotes on request.
What is Arcjet?
Arcjet is an AI agent runtime security platform that records governance on top of its runtime layer. It builds the AI inventory from what it observes, records who owns each system, enforces provider decisions and agent policies, and maps the records to framework evidence.
An AI system is an application or agent together with the model providers and tools it uses. Arcjet suggests AI systems by grouping an actor with the providers and MCP servers it uses, and a person confirms each grouping. The observations come from the Discovery endpoint agent on managed macOS and Windows devices, which records each connection to an AI provider and attributes it to a person; from coding agent hooks in Claude Code, GitHub Copilot, Cursor, OpenAI Codex and Muse Code; and from the agents you build through Arcjet's SDKs, the Claude Compliance API and OpenTelemetry.
Each AI system carries an owner, who must be a member of the team, a purpose, a review every 180 days, a risk tier and its EU AI Act role. An owner who leaves the team reads as no owner. Only team administrators edit these fields, and each change goes to the WorkOS audit log. Retiring an actor revokes its keys and keeps the record.
Provider decisions approve or prohibit an AI provider, or one client of it, such as the ChatGPT app as opposed to the OpenAI API. The Discovery agent enforces the decision on managed devices, and an observation of a prohibited provider at or after the decision time is a violation attributed to a person. Purpose drift compares the data classes and tools an owner declared with sensitive-information findings and captured tool calls, and notifies the owner. A weekly digest sends each owner their overdue reviews, violations and drift.
Arcjet Guard decides tool calls and prompts in the agents you build, with dry run, and a rule must pass a test before it goes live. Publishing a policy can require a second approver, and Hold decisions go to the system's owner. Sensitive-information detection covers PII, secrets and PHI.
The evidence page maps these records to NIST AI RMF 1.0, ISO/IEC 42001 and the EU AI Act, plus frameworks you define, and marks every requirement as supported or not covered. You export it as JSON or CSV or push it to Vanta or Drata. On the Enterprise plan, a log drain also exports decisions to your SIEM or log store.
How do OneTrust AI Governance and Arcjet differ?
OneTrust AI Governance and Arcjet differ in where each looks for AI, where each enforces, and how much of the compliance program each runs. The following sections cover each difference, then the areas where OneTrust goes further.
Cloud AI platforms compared with devices and agents
OneTrust discovers AI through the platforms where it is built and hosted. A connection to Bedrock or Databricks Unity Catalog lists the models, agents and datasets registered there, which is an accurate source for AI that your engineering and data teams deploy on those platforms.
Arcjet discovers AI where it is used. A developer who installs Cursor, a sales team that uses the ChatGPT app, or an agent you built that calls the Anthropic API all show up in Arcjet from the connections, hooks and SDK records they produce. None of those runs on a cloud AI platform, so a platform integration doesn't see them. The reverse also holds: a model registered in Databricks that no Arcjet-protected application calls doesn't appear in Arcjet. For the general tradeoff, see observed vs survey-based AI inventory.
Where enforcement happens
OneTrust enforces through the platforms it integrates with and through its own SDK and agents. The FAQ says Microsoft Foundry provides guardrail enforcement, the AI Guard SDK blocks or redacts sensitive data in Python applications that call it, and Guardian Agents apply controls at the point of action. The page doesn't describe which agent frameworks Guardian Agents attach to or how an MCP policy reaches an MCP client.
Arcjet enforces in three places it controls: the hook of each supported coding agent, the Guard call in your application, and the endpoint agent on managed devices. A provider decision made by an administrator in governance is the same value the endpoint agent reads, so prohibiting a provider blocks it on devices and counts later attempts as violations.
Assessment workflows compared with computed evidence
OneTrust builds evidence from assessments, approvals, attestations and policies, as well as telemetry and enforcement actions. That covers organizational requirements, such as an approved policy or a completed impact assessment, that runtime records can't.
Arcjet computes evidence from records: the share of AI systems with an owner, the share reviewed within 180 days, the share of observed providers with a decision, and violations since each decision. Each value drills down to the records behind it, and the page lists the requirements it doesn't cover as well as the ones it does. Arcjet doesn't hold policy documents or run impact assessments, so the ISO/IEC 42001 management system and the EU AI Act's fundamental rights impact assessment in Article 27 stay in your GRC tool.
Where OneTrust goes further
OneTrust AI Governance does several things that Arcjet doesn't. It inventories models, datasets and agents directly from Bedrock, Foundry, Vertex and Databricks. It runs assessment templates, tiers risk automatically and triggers reassessment on material change. It tracks policy versions, exceptions and remediation, and it shares a platform with OneTrust's privacy and third-party risk products, so AI governance can reuse the vendor records and data maps those teams already keep. It also monitors model quality and safety signals on supported platforms, which Arcjet doesn't measure.
When should you choose OneTrust AI Governance?
Choose OneTrust AI Governance when your organization builds AI on cloud AI platforms and runs its privacy and risk programs in OneTrust. OneTrust fits when any of the following apply:
- Your models and agents are built on Amazon Bedrock, Microsoft Foundry, Google Vertex or Databricks.
- Your privacy, data and vendor risk teams already work in OneTrust.
- You need assessment templates, automated risk tiering and reassessment on change.
- You want policy versions, exceptions and remediation tracked with the AI inventory.
- You want guardrails and PII findings on the platforms OneTrust supports.
When should you choose Arcjet over OneTrust AI Governance?
Choose Arcjet when your AI risk is in what people and agents do on devices and in code, and you need decisions enforced there. Arcjet fits when any of the following apply:
- You need to know which people reach which AI providers from managed macOS and Windows devices.
- You want provider decisions, including client-level ones such as the ChatGPT app versus the OpenAI API, enforced on devices.
- Your developers use Claude Code, GitHub Copilot, Cursor, OpenAI Codex or Muse Code, and you want each action decided before it runs.
- You build agents and want tool calls and prompts decided at runtime, with dry run and tested rules.
- You want evidence computed from records and exported as JSON or CSV, or pushed to Vanta or Drata.
Can you use OneTrust AI Governance and Arcjet together?
Yes. OneTrust covers the AI your teams build on cloud AI platforms and the assessment and policy workflow around it. Arcjet covers the AI that people and coding agents use on devices, and the runtime decisions on the agents you build. Together, the register holds both halves.
Arcjet's evidence page exports as JSON or CSV. You can bring the observed AI systems, owners, provider decisions and violations into the OneTrust inventory, and run OneTrust assessments on the systems Arcjet found. Neither product's public material documents a direct integration with the other, so plan for an import your team maintains. Choose one tool as the record of owner and purpose for each system, so the two copies don't disagree.
For the layering in general, see AI governance vs AI runtime security.
What are the alternatives to OneTrust AI Governance?
The alternatives to OneTrust AI Governance differ in whether they start from a register, from models or from runtime records. The following products cover overlapping ground:
- Arcjet builds an observed AI inventory from devices, coding agents and the agents you build, enforces decisions at runtime, and maps the records to framework evidence.
- Credo AI runs an AI registry, a risk assessment library and policy packs for the EU AI Act, NIST AI RMF, ISO/IEC 42001 and SOC 2. For more information, see Arcjet vs Credo AI.
- IBM watsonx.governance tracks AI use cases and models through their life cycle with factsheets, evaluations and IBM OpenPages. For more information, see Arcjet vs IBM watsonx.governance.
- Jamf AI Governance inventories AI apps on managed Macs and deploys each AI tool's enterprise settings. For more information, see Arcjet vs Jamf AI Governance.
For the categories side by side, see AI governance platforms compared.
Frequently asked questions
How is OneTrust AI Governance different from Arcjet?
OneTrust AI Governance discovers AI through cloud AI platforms such as Amazon Bedrock, Microsoft Foundry, Google Vertex and Databricks Unity Catalog, and runs assessments, policies and guardrails around it. Arcjet builds the AI inventory from Discovery on managed macOS and Windows devices, coding-agent hooks and the agents you build, records an owner, purpose and 180-day review per AI system, enforces provider decisions on devices and agent policies at runtime, and maps the records to NIST AI RMF 1.0, ISO/IEC 42001 and the EU AI Act on its evidence page. OneTrust looks where AI is built and hosted; Arcjet looks where AI is used.
Is Arcjet an alternative to OneTrust AI Governance?
For AI used on managed macOS and Windows devices and by coding agents such as Claude Code, GitHub Copilot, Cursor, OpenAI Codex and Muse Code, yes. For AI built on Bedrock, Foundry, Vertex or Databricks, and for assessment templates, risk tiering and policy exception tracking inside OneTrust's privacy and risk platform, OneTrust covers more. Many teams use both.
Does OneTrust AI Governance cover employee use of ChatGPT and coding agents?
OneTrust's AI Governance page, reviewed on November 16, 2026, mentions detecting shadow AI but doesn't describe browser, device, endpoint or coding-agent coverage. Arcjet records each connection to an AI provider from managed macOS and Windows devices, attributes it to a person, enforces provider decisions on the device, and decides each coding-agent action through hooks.
Where does OneTrust AI Governance enforce guardrails?
OneTrust describes Guardrail Enforcement that can block, allow, redact, escalate or route actions by policy, a Python AI Guard SDK that blocks or redacts sensitive data in prompts and responses, Guardian Agents for runtime oversight, and MCP policies with audit logs. Its FAQ says coverage varies by integration: Microsoft Foundry provides guardrail enforcement, and Amazon Bedrock provides usage, log analysis and PII findings.
How is OneTrust AI Governance priced?
OneTrust's pricing page says AI Governance is priced on admin users and AI inventory, and quotes on request. It doesn't list prices. For Arcjet Discovery pricing and availability, book a demo.
AI runtime security in your code
Protect your AI agent workflows with Arcjet
Arcjet runs inside your application, where it can use runtime context to enforce agent actions and budgets, detect prompt injection, and protect sensitive information before a workflow acts.