Enterprise AI: Value, Control and Advantage
Tokyo | September 10, 2026
Outcomes for the day
- Enterprise lessons learned: repeatable patterns that increase AI innovation velocity inside large organizations - what is working, what is not, and what operating model changes enable safe scale.
- External signals: outside-in perspectives on what is emerging in real deployments, informing where to partner, what to build, and what to avoid.
The goal is not consensus; it is collective intelligence.
Opening and Framing
We will have a mix of existing and some new Executive Technology Board members at this meeting. We will start the day with a quick round of introductions and agree how we will work together for the meeting. Please plan on sharing one example of AI that is planned or is working in your company, and one assumption in your AI strategy that you are testing.
What we are seeing across the Board
Through the meeting topics we will also reflect on a summary of what the Board is seeing across the globe. This year’s meetings are in Toronto, Amsterdam, Sydney, Seattle, Tokyo, New York and London. The discussion covers what is working, what has proved harder than expected, and the lessons members have taken from it. This material is based on the collective knowledge of over 250 members globally.
Topic 1: Where AI is Working
We look beyond pilots and demonstrations to understand where AI is materially changing economics, capacity, customer experience, speed or quality.
The discussion will examine what distinguishes use cases that scale from those that remain stuck in experimentation: how the opportunity was selected, how the underlying process was redesigned, where business ownership sits, how value is measured, and what had to change beyond the technology itself.
We will also look at where expectations have proved wrong. Some use cases have been easier technically but difficult organizationally; others create impressive productivity gains without translating into enterprise value. The objective is to identify the repeatable conditions that turn AI capability into measurable business outcomes.
Questions we will explore include:
- Where is AI already producing measurable enterprise value?
- Which use cases have scaled beyond a team or function?
- Where has the constraint moved from technology to process, organization or leadership?
Topic 2: Data, Architecture and Control
As AI moves deeper into enterprise workflows, the architectural question is changing from which model to use to how intelligence is controlled across the enterprise.
We will look at how AI connects into existing applications and data, and where common enterprise layers such as gateways, orchestration, identity, observability and policy enforcement are emerging. We will examine how organizations preserve flexibility as model capabilities, costs and vendors change rapidly.
The discussion will also address the increasingly important question of control. Where can data be held? What information can leave the enterprise? Who can authorize an agent to take action? How do we know what an agent did, what information it used, and why a decision was made? As autonomous activity increases, auditability and the ability to interrupt or reverse actions become as important as model performance.
Cost is becoming an architectural consideration as well. We will discuss how enterprises are measuring inference costs, managing multiple models, and deciding when premium intelligence is worth paying for versus when smaller or specialized models are sufficient.
Questions we will explore include:
- How much model and platform optionality do we need?
- What must remain visible, auditable and interruptible as agents become more autonomous?
- How are data residency, sovereignty and intellectual property influencing architecture?
Topic 3: AI-enabled Transformation, Culture and Driving Change
The largest constraint on AI may increasingly be the enterprise itself.
We will look at how enterprises are approaching this transition: where ownership of AI transformation sits, how business and technology leaders work together, how employees are involved in redesigning work, and how management structures change when people increasingly work alongside agents.
The discussion will also examine where sustainable advantage comes from. Every enterprise can increasingly access similar models. Differentiation therefore shifts toward the things that are harder to copy: proprietary data, institutional knowledge, unique workflows, decision history, customer context and the ability to continuously learn from operations.
This changes both transformation and technology strategy. What should we build because it captures something unique about our enterprise? What should we buy because it will rapidly become commodity capability? And how do we capture the knowledge held by experienced employees before it disappears from the organization?
Questions we will explore include:
- Where does our most valuable institutional knowledge actually reside and how do we turn that knowledge into a reusable enterprise asset?
- What should we build ourselves because it creates differentiation, and what should we increasingly buy?
- What happens to roles, management structures and decision rights as agents take on more work?
We agree the main points from the day and what the Board takes forward. Members receive a written summary afterwards.
Networking Time and Group Dinner