Bring this conversation to your leadership community

The AI Decision Boundary

What machines may decide, and what humans must own

Every organization already has a decision boundary for AI. Very few have chosen it. It is being set quietly, one configuration at a time, by whoever set up the tool. This session shows leadership teams how to draw the boundary deliberately: which decisions AI may recommend, draft, or act on within limits, which stay with people, who remains accountable when the system is wrong, and what evidence a decision must carry.

For Boards, executive teams, risk and governance groups

From Pilot Gridlock to an AI Operating System

Most enterprises are not short of AI pilots. They are short of a way to decide which ones deserve to live. This session lays out the operating model underneath the pilots: how value is defined before spend, how context is owned, how authority is assigned, and how a portfolio of experiments becomes a small number of workflows that actually changed.

For CIO, CTO, COO, transformation and portfolio leaders

Context Before Agents

Why enterprise AI fails without organizational context

Enterprise AI rarely fails because the model is weak. It fails because the model answers without the definitions, constraints, and prior decisions that an experienced employee carries. This session covers what organizational context actually is, how to bound it so it stays useful, who has to own it, and why agent programmes built on top of missing context produce confident work that nobody can sign.

For Data, platform, and AI leaders, and the executives funding them

From AI Spending to AI Value

Kill, scale, or redesign

Adoption is not value, and usage dashboards have quietly become a way to avoid the harder number. This session gives finance and operating leaders a defensible way to review an AI portfolio: the measure each initiative must move, the proof window, the evidence review, and the discipline to stop things that earned the right to be stopped.

For CFO, finance business partners, operating partners, portfolio owners

Human Gates for High-Stakes AI Decisions

A human in the loop is not a control if the human has no time, no context, and no authority to say no. This session covers how to design gates that hold: where they belong in the workflow, what evidence they require, how escalation works, what an override costs, and how to tell a real gate from a signature that was always going to be given.

For CISO, risk, compliance, internal audit, regulated operations