Matthew McBride

The Value Creation Framework

One frame for every program

AI doesn't create value. Better business capabilities do. This is the operating model I run every program on — the same three motions, whatever the industry.

The usual delivery model splits ownership three ways: consultants define the opportunity and step back, integrators build against problems they didn't help define, and change teams arrive after the solution is designed. Value is measured — if at all — after everything ships. The framework fixes that with one operating model that owns the whole lifecycle.

ENTERPRISE STRATEGYENTERPRISE ASSURANCE & ENABLEMENT LAYERGOVERNANCE & ASSURANCE · ORCHESTRATION & ENABLEMENTSHAPEAre we solving the right problem?SERVICE & EXPERIENCE DESIGN · ANALYSISBUILD & RUNCan we build and run the right solution?FORWARD-DEPLOYED ENGINEERINGREALIZEAre we delivering measurable value?VALUE & ANALYTICSREALIZE FEEDS THE NEXT SHAPE — A CLOSED LOOPBUSINESS PROCESS LAYERL1–L5 PROCESS TAXONOMY · DECISION RIGHTS · HUMAN–AI WORKFLOW DESIGNAI ENABLEMENT FOUNDATIONDATA · INTEGRATION · AI & AGENT PLATFORM · REUSABLE COMPONENTSENTERPRISE CORE SYSTEMS

Three motions, three questions

  1. 01

    Shape

    What should we build, and is it worth it?

    Assess, prioritize, quantify, design, roadmap. Capabilities are scored, not use cases; the value baseline is agreed before anything is built.

    Assess → Prioritize → Quantify → Design → Roadmap

  2. 02

    Build & Run

    Can we deliver it and keep it running?

    The team that builds a capability runs it. Shared foundations — data, integration, governance, the agent platform — are built once and reused by everything after.

  3. 03

    Realize

    Did it return what we said it would?

    Adoption is engineered into every motion, and benefits are claimed only on adoption evidence. What's learned goes straight back into Shape.

The continuous lifecycle

Shape. Build. Run. Realize.
Then back into Shape.

Inside a program the weighting shifts — shape-heavy early, then build-and-run, then sustained realization. Across a portfolio all three run continuously. It is not a waterfall.

SHAPEFRAME & DESIGNBUILDENGINEERRUNOPERATEREALIZEPROVE & SCALEONE LIFECYCLEA FLYWHEEL, NOT A LINE

Principles that govern it

Score capabilities, not use cases.

A capability that serves thirty activities is worth an order of magnitude more than its own business case suggests. The first build is chosen for how far it travels, not for what it returns on its own.

Build and run are one accountability.

No throw-it-over-the-wall integration. Engineers sit with the business, iterate in short cycles, and own the capability from first build through live operation.

Measure value from the outset.

The baseline is agreed with Finance before build, not argued after. A value realization lead carries that number from the business case to the end.

Adoption gates the benefit claim.

A capability nobody uses returns nothing. Change is engineered into every motion, and benefits are only claimed on adoption evidence — so change work sits on the critical path.

Build the foundation once.

Data, integration, governance, and the agent platform are shared assets. Every build after the first starts from a proven capability, an evaluation harness, and a control set — not from zero.

Transformation is a capability, not a project.

Shape, Build & Run, and Realize run continuously across the portfolio. Nothing is ever “finished.” That is what makes an enterprise autonomous, not merely automated.

We practice what we advise

We don't just tell enterprises how to work with AI. We run the framework on it ourselves — AI augmenting people to move faster, think further, and decide better, with a person making every call.

  • Shape

    Discovery synthesized in days, not weeks

    Interviews, workshop output, and existing documentation are synthesized with AI into themes, capability gaps, and a first-pass capability map — which the team then challenges, corrects, and owns.

    More time with stakeholders, less time transcribing them. Shape lands in weeks instead of months.

  • Shape

    Value cases built from drivers, not adjectives

    Every benefit is calculated from named drivers, marked measured or estimated, and benchmarked against past work. AI drafts the case; every number traces back to the record, not to a model.

    Investment decisions a CFO can audit — and a benchmark library that makes each next case faster and better evidenced.

  • Shape

    From first meeting to a priced plan

    Notes from a first conversation become a recommended engagement, a roadmap, a proposal, and an investment memo — drafted for the team to judge, never sent unread.

    Senior people spend their time on judgment and the client, not on assembling documents.

  • Build & Run

    Systems built so AI can't get them wrong

    Design systems, taxonomies, and governance are engineered as structure agents can read — closed registries, typed vocabularies, and checks with no model in the approval path.

    AI accelerates delivery without lowering the bar: what isn't in the system can't ship.

  • Build & Run

    Agents with earned, bounded authority

    AI agents run in shadow beside the team and take over a task only when their output matches expert work. Every consequential action is checked against what was intended before it runs.

    Automation that a regulated enterprise can actually put into production.

  • Realize

    Scorekeeping that never stops

    Benefit tracking, adoption signals, and risks are kept current from the program record, with the reason stated for every red or amber — and fed straight back into the next Shape.

    Value is claimed on evidence, and the same team can stay with more programs at once.