← Back to all services Education

Governing AI for Boards

  • 3 days
  • For Boards
  • On site · On campus

A three-day in-company program for the members of one supervisory board, built around your own company. For insight, judgment and oversight of the risks and opportunities of AI. We work with your own revenue model, plans and AI systems, under strict confidentiality, at Columbia Business School in New York or at a location of your choice.

Request the brochure →Plan an introduction →

Is this for you?

Who it is for

The members of one supervisory board, or a comparable body such as a German Aufsichtsrat. Committee chairs play a central role.

Group size

All members of one board. Your HR director, your CIO or CISO and your management board join selected sessions.

Prior knowledge

No technical background needed. Nobody has to learn to build AI.

Core objectives

  1. 01See what AI changes. Understand how AI changes your revenue model, competition and work.
  2. 02Judge the plans. Assess management's AI plans on value, risk and substantiation.
  3. 03Ask for the right information. Set what the board needs from management on AI, and when.
  4. 04Test the controls. Check that the controls for safe AI use work, from data security to audit trail.
  5. 05Keep costs and dependence in check. Manage AI costs and dependence on a single model or cloud supplier.
  6. 06Leave with a plan. An AI risk appetite and an oversight agenda for the year ahead.
Why it matters

AI now touches the strategy, operations and risks of almost every large company.

The board has to oversee it, yet few supervisory directors can properly judge what management reports. Meanwhile, regulators are setting hard deadlines.

The program

Three days, one board

Three consecutive days around your own company, session by session. Every session ends with a result your board keeps.

Day1

Insight

What AI changes in your company

Directors get a workable picture of what AI can do, where value and competition are shifting in your sector, and what it asks of people and culture.

  1. 01

    The board's blind spot

    Why AI has become a board matter and how it changes the oversight role. The session maps where the board stands now and which oversight questions the program must answer for your company.

    Result The board's own list of oversight questions

  2. 02

    What AI can and cannot do

    What AI can and cannot do today, demonstrated live, and why its strengths and weaknesses matter to the board. Directors learn to weigh claims about AI, test them against practice and ask better follow-up questions.

    Result A clear guide to what AI can do today

  3. 03

    Revenue models under pressure

    How AI shifts value in your sector and what that means for your revenue model. Directors judge which parts are most vulnerable, where AI-native entrants can compete and where you keep a lasting advantage.

    Result An AI impact map for each revenue stream

  4. 04

    Attack your own company

    A competition exercise that looks at your company through the eyes of an AI-native challenger. Directors see where the business is vulnerable, what a credible response looks like and what inertia would cost.

    Result Three attack scenarios and the board's response

  5. 05

    People, culture and values

    How AI changes work, roles and leadership, and what it asks of your culture and values. Directors explore what builds trust and acceptance, and when the works council must consent. With your HR director.

    Result Five questions on people for management

  6. 06

    Reading management's AI plan

    How to test whether management's AI plans deliver real value. Directors learn to tell a credible business case from an overly optimistic one, how staged funding limits risk and which red flags matter.

    Result A one-page checklist for AI proposals

Optional dinner with the management board

An open conversation with management about the company's AI ambition, outside the formal agenda.

Day2

JudgmentThe core

Governance and control

The core of the program. Directors learn where oversight of AI sits within their board, which controls must be in place and how to test whether they work.

  1. 01

    Oversight in a two-tier board

    The supervisory board oversees and advises, but does not run the company. With AI, the work is about the right questions and the right information: what management reports, the board's own duty to ask, the role of the audit committee and which AI decisions need the board's approval.

    Result An information protocol: what the board receives on AI, and how often

  2. 02

    The governance architecture

    How responsibility for AI is divided between the board, its committees, management and internal audit. Directors learn which building blocks a company needs, how management puts an AI risk appetite to the board for approval, and how frameworks such as ISO/IEC 42001 help benchmark the approach.

    Result A draft AI risk appetite for your company

  3. 03

    Inside the control layer

    What happens between an employee's question to AI and the answer, and where your company keeps control: protecting data, limiting access, routing work to the right model, logging activity for audit and preventing misuse, including AI tools staff use without permission. The focus is on the questions the board asks management at each step.

    Result The questions for each control point

  4. 04

    The control layer in practice

    A live, vendor-neutral demonstration of the control layer, with your CIO or CISO. Directors see what good control looks like and what evidence they can ask management for.

    Result Eight controls the board must test

  5. 05

    Regulation and compliance

    Which rules apply to your company's use of AI, from the European AI Act to privacy, disclosure and sector rules. Directors learn what the board must know, what it can delegate and how to read compliance reports.

    Result An overview of the regulation for your company

  6. 06

    Costs and independence

    How AI costs develop as use grows, and how dependent the company is on a single model or cloud supplier. Directors learn to judge spending, exit options and protective contract terms.

    Result Five questions on costs and supplier dependence

Day3

Oversight

The board in action

Directors practice the work itself: handling an incident, testing assurance, reading an AI dashboard and deciding on a real proposal from your company.

  1. 01

    Incident simulation

    A realistic simulation of an AI incident, built on your own systems and played in four rounds. Directors practice what they need to know when an AI system fails, when they need to know it and what they then expect from management.

    Result An escalation protocol between board and management

  2. 02

    Assurance you can rely on

    How directors can trust what management tells them about AI. The session shows who provides independent assurance, what a certification such as ISO/IEC 42001 does and does not prove, and how AI claims in the annual report stay accurate.

    Result A plan for who checks what, and how often

  3. 03

    The board's AI dashboard

    What the board should see on AI every quarter, and how to read it. Directors set a short list of indicators for value, risk, costs and compliance, so problems surface early and decisions rest on facts.

    Result An AI quarterly dashboard for your board

  4. 04

    Board meeting simulation

    A board meeting on a real AI proposal from your company. Directors practice questioning management critically, weigh value against risk and reach a decision, with conditions if needed. Afterwards they look back at the questions that mattered most.

    Result A decision memo for your next board meeting

  5. 05

    The board itself

    How the board builds its AI knowledge through its profile, self-evaluation and education, and how directors use AI in their own work without putting the confidentiality of board papers at risk.

    Result A development plan for the board itself

  6. 06

    Our oversight agenda

    Everything comes together in an oversight agenda for AI that the board sets for the next twelve months: its questions, information needs, committee roles and its own development.

    Result An oversight agenda for AI for the next twelve months

After the program

Follow-up session

A follow-up session with the board, alongside a regular meeting, on progress with the oversight agenda and the first decisions.

Eight controls the board must test

Directors do not build these controls themselves. They ask management to show that each control works.

  1. 01Data rulesWhich data may go to which tool
  2. 02Access controlWho may use which model and which data
  3. 03Approved suppliersA short list of tested models
  4. 04Model routingThe right model for each task
  5. 05Audit trailWho asked what, when, and with what result
  6. 06GuardrailsMisuse blocked and regularly tested
  7. 07Cost limitsBudgets and alerts per team and application
  8. 08Exit planSwitching without rebuilding everything

How the program works

  • Your company is the case

    We work with your own revenue model, plans and AI systems, under strict confidentiality. No standard cases.

  • Judgment over technology

    Directors learn which questions, evidence and red flags they need to hold management to account. Nobody has to learn to build AI.

  • Practice as a board

    An incident simulation and a meeting on a real proposal turn knowledge into decisions.

  • Independent and transparent

    Praexo also delivers AI solutions. We do not promote them in this program, and we disclose every interest in advance.

Before the program

Short preparation for each director, plus preparation with management, so the three days start from your own practice.

  1. 01

    Conversations

    Short conversations with the chair of the board, the CEO and the CFO about ambitions, concerns and the decisions ahead.

  2. 02

    Fact sheet

    An overview from management of the AI systems in use, plans, spending and incidents.

  3. 03

    Self-evaluation

    Each director answers twelve questions on how the board oversees AI today. The results are shared anonymously.

Where it takes place

Columbia Business School

On campus

At Columbia Business School in New York.

On site

At your company, or at another location of your choice.

Faculty and guests

Faculty
  • Marc De KuyperFounder of Praexo Group. Teaches AI at Columbia Business School.
  • Praexo specialistsAI governance, security and technology.
From your company
  • Your HR directorDay 1, people and culture
  • Your CIO or CISODay 2, live demonstration
  • Your management boardOptional dinner on day 1 and the simulation on day 3

Everything your board shares stays confidential and is used only for this program. Content and faculty may change.

Deliverables

  • AI impact map for each revenue stream
  • Checklist for AI proposals on one page
  • Information protocol: what the board receives on AI, and how often
  • Draft AI risk appetite for your company
  • Checklist of eight controls the board must test
  • Regulatory overview for your company
  • Escalation protocol between board and management
  • AI quarterly dashboard for your board
  • Oversight agenda for the next twelve months

What comes next

3 days

Governing AI for Boards

An AI risk appetite, nine working tools and an oversight agenda for the next twelve months.

Request the brochure →Plan an introduction → ← Back to all services