Shippable Strategy
Strategy that ships, with a number you can defend.
Most AI strategy work produces a roadmap and stops. A roadmap is an output of two of the five layers at a point in time. Without the mechanisms that produce and revise it, it is a document. This engagement installs the AI operating system at the scope you own and pushes the first change through it.
In practice
We start with one recurring piece of work you own: the month-end close commentary, the weekly labour forecast, the player reinvestment review. We settle who decides about it, who pays for it, what data it runs on, who builds it and how, and what number it moves. Then we run the first change through all five with your team and leave those arrangements in place, owned by named people. The roadmap for what comes next is part of the package. It is not the package.
Diagnose, decompose, deliver.
Diagnose
Which of the five layers exist in your scope today, how heavy they are, and where the gap is. Most organizations in the middle have some layers and not others: an enablement team with no one deciding above it, executives building outside it with nothing to run on, or a controller's team doing the work by hand because no one has decided it is allowed to change.
Decompose
The workflows, taken apart into people, platforms and processes. This is where the number gets built: what the work costs today, what changes, and who owns the result.
Deliver
The mechanisms installed and the first change pushed through them, so the system is proven on something real before the engagement ends.
One system at any scale.
A workflow, a unit, an enterprise. Each is an operating system with the same five layers. What lets them stay one system is a shared layer of data and context; deciding and funding across units is its own work, and the engagement treats it as such.
Enterprise
Decide
- AI: AI drafts the portfolio view: what is running, what is proposed, what overlaps
- Person: Executives set priorities and risk appetite
- Gate: Risk and regulatory review before anything touches players or money
Fund
- AI: AI drafts the funding case from each unit's number
- Person: Capital committee funds, in tranches, against the number
Units inside it
- Decide
- Executive committee sets priorities and risk appetite.
- Fund
- Capital committee, in tranches, against each unit's number.
- Data and context
- Definitions, policies, authority limits, shared across units.
- Deliver
- Enablement team and vendors, under the decisions above.
- Account
- Return by unit, reviewed quarterly.
Unit
Workflow
Data and context
Inherited by every unit definitions, policies, authority limits, reported results.
Bottom-up or top-down.
Bottom-up is one piece of work with all five layers, kept light. Top-down is a mandate with all five layers at enterprise weight. Neither is wrong. Start with one workflow and you will reconcile its definitions and controls with the next one later. Build the shared layer first and you pay for it before you have a number to show. Most organizations in the middle started bottom-up without deciding when that reconciling would happen. Starting small is fine. Owning the five layers from the start is what keeps it from becoming another pilot.
What you receive.
For each of the five layers at your scope: who owns it, how it runs, what it produces, how often, and what number it moves, in place and running by the end of the engagement. The roadmap is in there as a component. So is the delivery method your teams will use, chosen for the teams you have, whether that is your existing enablement group, a vendor, or the analyst who owns the workflow. We do not sell a platform.
A 30-minute conversation.
Bring the mandate, the context, or the problem.