How I Think

Most AI transformations don’t fail because of technology.

They fail because strategy, operating models, governance, adoption, and measurement are treated as separate initiatives rather than a connected system.

My work focuses on helping organizations close those gaps and turn AI ambition into measurable business outcomes.

01

AI is not the transformation.

Technology alone does not change organizations.

Sustainable transformation requires changes in leadership, operating models, governance, incentives, and ways of working.

02

Adoption matters more than capability.

Most organizations already have access to powerful AI tools.

The challenge is creating sustained adoption, behavior change, and measurable business impact.

03

Business value is the final metric.

Strategies, pilots, governance frameworks, training programs, and transformation initiatives only matter if they create meaningful outcomes.

Value creation is the objective. Everything else is an enabler.

04

Transformation succeeds as a system.

Strategy, operating models, governance, adoption, and measurement should not be optimized independently.

Organizations create lasting value when all elements of the transformation system work together.

From Strategy to Business Value

AI Transformation Value Chain

The AI Transformation Value Chain is a high-level view of my AI Transformation Playbook.

It captures how I approach AI transformation as a connected system, linking strategy, operating models, governance, adoption, measurement, and execution to deliver measurable business outcomes.

From intent to outcomes

Six connected stages that turn AI ambition into measurable business value.

Each stage answers a distinct question. Weak links in the chain limit what the others can deliver.

  1. 01

    Strategy

    What are we trying to achieve?

    Define the business outcomes, strategic priorities, transformation ambition, and success criteria that AI should enable.

  2. 02

    Operating Model

    How do we organize for success?

    Establish ownership, decision rights, roles, capabilities, and execution structures required to turn strategy into coordinated action.

  3. 03

    Governance

    How do we make decisions and manage risk?

    Create the guardrails, accountability, policies, and decision mechanisms needed to scale AI responsibly without unnecessarily slowing execution.

  4. 04

    Adoption

    How do we change behaviors and ways of working?

    Build leadership alignment, capabilities, communication, communities, and change mechanisms that embed AI into everyday work.

  5. 05

    Measurement

    How do we know what is working?

    Define adoption indicators, outcome metrics, feedback loops, and value-tracking mechanisms that support evidence-based decisions.

  6. 06

    Business Value

    How do we create lasting outcomes?

    Translate execution and adoption into measurable productivity, innovation, operational, customer, and strategic outcomes.

Six connected stages · continuous flow

The companion kit.

The Value Chain is the framework. The AI coding adoption playbook is a companion kit for engineering teams, with guides, templates and a small adoption scorecard. It is open source.

The five moves of an AI coding rolloutFive steps in order. 1 Align: who owns the rollout, what problem are we solving, and how will we know if it worked. 2 Enable: how does each role get from access to a useful first week. 3 Embed: which useful behaviors should become repeatable team practices. 4 Measure: what should we measure, and where can the numbers mislead us. 5 Expand: should we expand, hold, or fix the rollout first.12345AlignWho owns the rollout, what problem are wesolving, and how will we know if it worked?EnableHow does each role get from accessto a useful first week?EmbedWhich useful behaviors should becomerepeatable team practices?MeasureWhat should we measure, and wherecan the numbers mislead us?ExpandShould we expand, hold, or fix therollout first?
The five moves of a rollout. Each one is a question the team has to answer before moving on.
Five measurement layers, kept apartFive layers, each answering a different question. Access: who holds a license. Activation: who has used it at all. Adoption: who keeps coming back. Depth: how do active users work with it. Outcomes: is anything about the engineering work improving. Between the layers: licenses do not prove use; one sign-in is not repeat use; use does not prove depth; depth does not prove productivity.01AccessWho holds a license?Licenses do not prove use.02ActivationWho has used it at all?One sign-in is not repeat use.03AdoptionWho keeps coming back?Use does not prove depth.04DepthHow do active users work with it?Depth does not prove productivity.05OutcomesIs the engineering work improving?
Five measurement layers, kept apart. Each answers a different question, and none proves the next.

View the adoption playbook repoRead the five guides