OneTrust University · Release brief
Sample module · not an official OneTrust course
New this release TrustWeek 2026 · Sept 30

Governing AI agents at runtime

What changed in the September 30 platform release, why it matters to your governance program, and how to get your team ready.

Lessons5
TimeAbout 12 minutes
ForPrivacy, risk and AI governance admins
Lesson {{ lessonNum }} of 5 · {{ lessonTitle }}
Lesson 1 · What changed

Four new pieces in the OneTrust Platform

At TrustWeek on September 30, 2026, OneTrust introduced an evolution of its platform built to govern AI at enterprise scale while keeping people in control. It centers on CORIE, a shared intelligence layer, with the AI Control Plane, the Governance Command Center and a headless MCP Gateway built around it.

{{ tab.kicker }}
{{ tab.name }}

{{ tab.body }}

{{ p }}

The shift in one picture

Before
Governance as an assessment

Reviews happen before or after an AI system is deployed. Once an agent is running, its individual actions aren't checked as they happen.

Now
Governance while agents work

Policies are enforced at runtime. Each agent action proceeds, is blocked, or waits for a person, and the decision is recorded as evidence.

i
Check availability before you plan
Most announced capabilities are headed to private preview first, with Fall or Winter timelines. Confirm with your account team which ones your organization can access.
Lesson 2 · Why it matters

Agent use is outpacing governance

Organizations that encourage AI agent use 87%
Have clear governance, oversight and controls 47%
Source: OneTrust 2026 AI-Ready Governance Report

That 40-point gap is the problem this release targets. When agents act faster than reviews can keep up, governance has to move into the workflow itself. Here's what that means for the teams that use OneTrust.

Select each team to expand.
{{ a.body }}
Lesson 3 · How to apply it

Practice: you set the outcome

Your company runs an AI support agent. Each card below is an action it's trying to take. Decide what the AI Control Plane should do with it, based on the sample policy.

Sample policy Refunds auto-approve up to $500 Personal data only goes to approved vendors
Agent action {{ s.n }} support-agent · {{ s.tool }}
{{ s.action }}
{{ s.verdict }} {{ s.fb }}
● evidence recordeddecision={{ s.correctKey }}policy={{ s.policy }}

Get your team ready for preview

{{ step.n }}
{{ step.t }}
{{ step.d }}
Lesson 4 · Knowledge check

Three quick questions

Question {{ qNum }} of 3
{{ q.q }}
{{ qVerdict }} {{ q.fb }}
Your result
{{ score }} of 3
{{ scoreMsg }}
Lesson 5 · Recap

What to take back to your team

What changed
CORIE, the AI Control Plane, the Governance Command Center and the MCP Gateway bring governance into AI workflows as they run.
Why it matters
Most organizations encourage agent use, but fewer than half have clear controls. Runtime enforcement closes that gap without slowing work down.
How to apply it
List your agents, name the decisions they make, set approval thresholds and assign evidence owners before preview access arrives.
{{ finishTitle }}
{{ finishSub }}
Sample module based on public information: OneTrust TrustWeek 2026 announcements (Sept 29–30, 2026) and press coverage. Not an official OneTrust course.
{{ continueHint }}