Insights

Straight talk on enterprise AI strategy.

No hype, no vendor pitch decks — the same diagnostic thinking behind our advisory work, written out.

Featured research · Executive briefing

Enterprise AI: what happens after the demo?

Six operating questions that connect AI ambition to cost, data, control, reliability, and accountable decisions.

Explore the briefing ↗
Operating model

Who owns the AI decision?

A practical division of authority across the business, the AI function, security, finance, and executive leadership.

September 2026
Operating Doctrine

The twelve first principles of enterprise AI strategy

An interactive field guide to twelve durable rules for enterprise AI decisions. Explore a leadership scenario and build your own discussion agenda.

September 2026
Data Exposure

Shadow AI is your biggest exposure

The logo may be familiar. The account, data path, and permission boundary still need to be understood.

August 2026
Economics

Why AI pilots stall before production

A working demo is the beginning. Explore the economics, ownership, and operating evidence that earn the next investment.

August 2026
Vendor Strategy

You don't control the control plane

Know which decisions belong to your organization, which belong to your providers, and what it would take to change course.

August 2026
Operating System

The 90-day platform stand-up

A practical first-quarter plan for installing decision rights, a usable approved path, and evidence that the operating model works.

August 2026
Vendor Reality Check

Five things the AI vendor pitch leaves out

Five questions that turn an attractive AI proposal into an operating decision you can defend.

August 2026
Governance & Law

The AI Act obligations already binding you

The EU AI Act has several clocks. Start with the system, your role, and the obligation—not a headline about delay.

July 2026
Security

Your AI's supply chain is now an attack surface

Model artifacts, application dependencies, and tool servers need distinct trust decisions before they enter a production workflow.

July 2026
Security

Excessive agency — the AI risk that doesn't need an attacker

A correct login and an authorized tool call can still produce the wrong action. Explore how permissions change an agent’s reach.

July 2026
Reliability

Model deprecation is a reliability risk, not an IT ticket

A model change is a change to the service your business relies on. Plan the evaluation, cutover, and fallback before the deadline.

July 2026