FREE · SAT, NOV 21The AI GRC Summit 2026: Governing AI in the Age of Agents. Live on Zoom.Save your seat →

Agentic AI governance training

We train the people who sign off on AI agents.

Become the person your company trusts to govern AI agents. Practice on a running agent in a six-week live cohort, then defend a Go or No-Go decision with evidence an auditor can follow.

  • 100+ five-star reviews
  • 24 CPE per participant
  • No coding needed
  • No production data needed
AGENT SIGN-OFF RECORD AGT-014

Claims-intake agent

Owner
Claims Operations
Risk tier
High
Boundaries
Reads claims, drafts letters. No payments.
Human approval
Every denial, every payout
Controls tested
14 of 14 held
Evidence mapped
ISO/IEC 42001 · NIST AI RMF · EU AI Act
DECISION Conditional Go Approve with a weekly trace review for the first 90 days.
Illustrative. You build one like this in the cohort.

Evidence mapped to the frameworks you report on

  • ISO/IEC 42001
  • NIST AI RMF
  • EU AI Act
  • AIUC-1
  • OWASP
  • MITRE ATLAS

Why this matters

“Can we trust this agent?” just landed on your desk.

Agents don’t just answer questions. They read data, call tools and take actions. Security reviews the access, privacy reviews the data, legal reviews the contract, GRC writes a policy. Everyone reviews a piece. Nobody owns the decision.

Sooner or later, someone asks three questions about every agent:

  1. 01What is it allowed to do without a person?Not what the policy says. What the permissions actually allow.
  2. 02Who approved it, and on what evidence?A name, a date, and a record someone else can check.
  3. 03If it did something wrong yesterday, could you prove what happened?That takes a trace, not a guess.

Most teams can answer one of the three. We train you to answer all three, for every agent.

The method

One shared way to review any agent.

You learn it on one agent, then run it on every agent after that. Same six steps, the same evidence, whoever is in the room.

  1. 01MAP IT

    Inventory

    Owner, purpose, users, data, tools, and how much the agent does on its own.

  2. 02SIZE IT

    Risk-tier

    Which agents need enhanced review, based on impact, autonomy, data and exposure.

  3. 03BOUND IT

    Set controls

    Boundaries and human checkpoints, each with an owner and the evidence it must leave.

  4. 04BREAK IT

    Test

    Prove the controls hold, then read the trace of what the agent actually did.

  5. 05SIGN IT

    Decide

    Go, Conditional Go or No-Go, with the residual risk stated plainly.

  6. 06WATCH IT

    Monitor

    Thresholds, alerts, change control, incident triggers, and when to retire it.

Ways to work with us

Start with yourself. Or train your whole team.

FOR TEAMS OF 10–20Most teams start here

The Team Program

A private six-week program for your GRC, security, audit, privacy and risk team, built around one of your own agents. It ends with your team presenting a Go, Conditional Go or No-Go readout to your executives.

  • Six weeks, live and remote, scheduled around your calendar
  • Capstone on a representative version of your agent: no production data, no system access
  • The full governance pack for that agent, mapped to ISO/IEC 42001, the NIST AI RMF and the EU AI Act
  • 24 CPE and a certificate of completion for every participant
Priced for your team.We scope it together on a 30-minute call: your agent, your people, your calendar. If it fits, you get a one-page proposal within 48 hours.
FOR INDIVIDUALS

The live cohort

Agentic AI Governance for GRC & Cybersecurity. Six weeks of hands-on labs with feedback, and a capstone you present and defend at Demo Day.

  • Hands-on labs in VerifyWise, Dify, Opik and MITRE ATLAS
  • Feedback on every submission
  • 24 CPE for ISACA, ISC2, CompTIA or IAPP
Next cohort starts January 12 · 30 seats
$2,500per seat · 5.0 on Maven
$500 off with code KNOW-YOUR-AGENTS
Save $500: join the January 12 cohort
MEMBERSHIP

The Agent Vault

Courses, templates, certification prep and monthly live sessions for people who govern AI agents, or want to.

  • Design your agentic AI governance program
  • AIGP certification prep (from January)
  • The Replay Vault: monthly workshops, AMAs and guest sessions
$49/monthor $499/year · founding rate
Founding rate for the first 100 members.
Join at the founding rate

What your team leaves with

A governance pack for your agent. And a team that can build the next one.

Your team keeps the pack, the templates and the method, and can run the same review on the next agent without us.

  1. BEFORE KICKOFF

    Scope. One call to choose the agent, confirm who attends and set the schedule.

  2. LIVE SESSIONS AND LABS

    Build. Inventory the agent, set its risk tier, boundaries and human checkpoints, using your own policies.

  3. LABS IN REAL TOOLS

    Test. Prove the controls hold and read the agent’s trace.

  4. FINAL SESSION

    Decide. Your team presents its readout to your executives, with the evidence index behind it.

GOVERNANCE PACK · AGT-0148 artifacts
  1. 01Inventory entryOwner, purpose, users, data access, tools and autonomy, in one record.
  2. 02Risk tierThe tier and the reasoning behind it, written so an auditor can follow it.
  3. 03BoundariesWhat the agent may read, send and do without a person.
  4. 04Human-oversight checkpointsWhere a person must approve, and the test that proves the gate works.
  5. 05Evaluation and monitoring planTest cases, thresholds, alerts and a review cycle.
  6. 06Incident runbookTriggers, owners and the first 24 hours when the agent misbehaves.
  7. 07Evidence indexEvery artifact mapped to ISO/IEC 42001, the NIST AI RMF and the EU AI Act.
  8. 08Leadership readoutGo, Conditional Go or No-Go, presented by your team to your executives.

Proof

Practitioners who’ve done the reps.

Consultants, analysts and security leads who learned the method on a running agent, in their own words.

five-star reviews across our programs
100+
average rating on Maven
5.0
CPE for every participant
24
graduates whose capstone defenses you can watch
6
I walked away feeling genuinely prepared to introduce AI responsibly in my day-to-day.
Rumteen T.Senior Consultant, Deloitte
★★★★★
“It gave me the exact practical experience I need for my consulting work in this space.”
Oyinjadesola E.AI Consultant, Nexs Risk & AI Advisory
★★★★★
“This course was an excellent hands-on learning experience that helped me understand how AI governance works in real business environments.”
Peter D.Security Analyst
★★★★★
“The topics and content were very well laid out with very good explanations, step by step guidance and lots of reference material to study.”
Altaf (AL)Security Project Manager, Altech Consultancy

Why now

Training is part of the evidence.

The frameworks you report against already ask for trained, competent people. Every participant’s certificate and CPE record becomes part of that evidence.

EU AI ACT · ARTICLE 4

AI literacy is already a legal duty

Providers and deployers must take measures to ensure a sufficient level of AI literacy among the staff who operate and use AI systems. It has applied since February 2, 2025.

ISO/IEC 42001 · CLAUSE 7.2

Competence needs proof

Determine the competence AI work requires, make sure people have it through education, training or experience, and keep documented evidence that they do.

NIST AI RMF · GOVERN 2.2

Training is part of governing

Personnel and partners receive AI risk management training so they can carry out their duties consistent with your policies and procedures.

François B. Arthanas, founder of CyberProsAI

Who teaches it

François B. Arthanas

I Train GRC & Cyber Pros to GOVERN AI Agents | Founder, CyberProsAI | Agentic AI Governance | CISSP · CISA · AIGP | Ph.D. Candidate, Cyber Defense

François works across cybersecurity, audit, governance and technology risk, and is a Ph.D. candidate in Cyber Defense at Dakota State University. He runs the same method in his public cohort, where practitioners build and defend real agent decisions in hands-on labs.

Every Team Program session is taught live by François.

Start free

Not ready for a call? Start here.

FREE EVENT · SAT, NOV 21 · 10 AM–4 PM ET · LIVE ON ZOOM

The AI GRC Summit 2026: Governing AI in the Age of Agents

Real implementations, not theory. Practitioners from Trainline, Reco and Cottage Health show what they built, the tradeoffs they faced and what they’d do differently. Certificate of attendance included.

FREE KIT · 3 TEMPLATES + A WORKED EXAMPLE

The Agent Pre-Deployment Review Kit

The three templates every agent review ends up needing: the Agent Register Entry, the Control Test & Trace Review and the Decision Record. Each comes blank and filled in with a worked example, so you see what good looks like.

FREE COMMUNITY · ON SKOOL

Agentic AI Governance community

1,300+ GRC, security and audit professionals working through agent governance together. Ask questions, share what you’re reviewing, and see the method before you commit.

Free newsletter, every two weeks

The Agent Review Board

One real AI agent, reviewed every two weeks, with the verdict at the end. For GRC, audit and security professionals who govern AI agents.

  • 01The agent, and what it’s allowed to do
  • 02The controls and the evidence to ask for
  • 03The verdict: Go, Conditional Go or No-Go

One email every two weeks. Unsubscribe in one click.

FAQ

Questions buyers ask.

Something else? Write to francois@cyberprosai.com. François reads every message and replies personally.

Do we have to share production data or give you system access?

No. The capstone uses a representative version of your agent: its purpose, data types, tools and approval flow. Anything sensitive follows a process we agree on before kickoff.

Who should attend?

10 to 20 people across GRC, security, audit, privacy, legal, compliance and the teams that build or own agents. Technical and non-technical people learn it together. No coding needed.

Does it count for continuing education?

Yes. Each participant earns 24 CPE (ISACA, ISC2, CompTIA or IAPP) and a certificate of completion. Those records also serve as training evidence for EU AI Act Article 4 and ISO/IEC 42001.

How does payment work?

One proposal, one invoice, paid directly: 50% at signing, 50% at kickoff. We’ll work through your vendor onboarding.

What if we have fewer than 10 people?

Enroll them in the public cohort at the per-seat price. The method and labs are the same, and the next cohort starts January 12.

Do I need a technical background?

You do not need to be a software engineer. You do need to be willing to understand how an agent uses data, models, tools, identities, permissions, integrations, memory and monitoring.

What happens after the six weeks?

Your team keeps the governance pack, templates and method, and can run the same review on the next agent without us.

Your next agent approval should come with evidence.

Book a 30-minute Team Fit Call with François, not a sales rep. You’ll leave knowing whether this fits, even if it doesn’t.

  1. 1Which agents you run or plan, and who signs off on them today
  2. 2Whether the Team Program fits, or a smaller option does
  3. 3If it fits: a one-page proposal within 48 hours
Training a team of 10+?Book a Team Fit Call
© 2026 CyberProsAI. All rights reserved.Privacy policy