The strategy question
An AI strategy is not a tool list. It is a position.
Most AI strategies are a pile of pilots and a tooling budget. A real one names the one or two things that change at the core of your business, and builds the capability to act on them first.
Build an AI strategy in four moves. First, name what fundamentally changes about your business as AI enters your industry, the billable hour, the data moat, the commercial layer. Second, decide the position that change creates and who you are willing to disappoint to hold it. Third, build the capability, starting with the leadership team, because you can only set strategy for what you can imagine. Fourth, sequence it: what to build, what to prove, what to stop. Tools come last, not first.
What changes
Start with what changes, not what to automate
An AI strategy is a decision about competitive position, not a plan for which tools to buy. Every industry has one or two things AI structurally changes: pricing logic, the moat, the leverage model, the commercial layer. An AI strategy names your version of that shift and commits to a position around it before competitors do. A list of use cases is not a strategy. A named position, and the capability to reach it, is.
What it includes
What a real AI strategy includes
A real AI strategy includes four things: a diagnosis of what AI changes at the core of your business, a stated position you will build toward, a plan to build the capability (which starts with your leadership team, not your IT department), and a sequence, what to build now, what to prove next, what to stop doing. It does not need a long list of tools. Tools are downstream of the position, and they change every quarter.
The choice
The choice that decides everything
Underneath every AI strategy is one fork: automation or augmentation. Automation cuts cost and shows clean ROI this quarter, then plateaus, and everyone reaches the same parity. Augmentation multiplies what your people can do, compounds, and builds an advantage that takes years to copy. Most companies pick automation because the ROI is visible now. The company that picks augmentation looks worse in Q1 and wins by year three.
The J-curve
Expect it to look like failure first
Real transformation follows a J-curve: performance dips while the organization redesigns itself around the new capability, then compounds. Clean ROI in the first quarter usually means you are automating around the edges, not transforming. Plan for the dip, measure the right things during it, learning speed and new capability rather than next quarter's margin, and hold the course. The companies that quit in the trough are the ones that waste the investment.
Who owns it
Who should own it
The CEO and the leadership team own AI strategy. It cannot be delegated to IT or an innovation team, because those teams propose what they can see from their seat, tools and efficiency, while the questions that decide position can only be answered by the people who own the P&L. The CEO does not need to become technical. The leadership team needs enough hands-on experience to imagine what is possible, because an organization cannot set an ambition higher than its leaders can picture.
That is why GRAIL's programme starts with the management team building real AI fluency in its own work. The full arc is on the services page, and the leadership case is on AI for leadership teams.
A strategy you cannot state in one sentence is a strategy you do not have.
Name the position. Build the capability. Sequence the moves. Everything else is a tooling budget.
Questions leaders ask
How do you build an AI strategy?
Build an AI strategy in four moves. First, name what fundamentally changes about your business as AI enters your industry, the billable hour, the data moat, the commercial layer. Second, decide the position that change creates and who you are willing to disappoint to hold it. Third, build the capability, starting with the leadership team, because you can only set strategy for what you can imagine. Fourth, sequence it: what to build, what to prove, what to stop. Tools come last, not first.
What should an AI strategy include?
A real AI strategy includes four things: a diagnosis of what AI changes at the core of your business, a stated position you will build toward, a plan to build the capability (which starts with your leadership team, not your IT department), and a sequence, what to build now, what to prove next, what to stop doing. It does not need a long list of tools. Tools are downstream of the position, and they change every quarter.
Who should own AI strategy?
The CEO and the leadership team own AI strategy. It cannot be delegated to IT or an innovation team, because those teams propose what they can see from their seat, tools and efficiency, while the questions that decide position can only be answered by the people who own the P&L. The CEO does not need to become technical. The leadership team needs enough hands-on experience to imagine what is possible, because an organization cannot set an ambition higher than its leaders can picture.
How long does an AI strategy take to pay off?
Real transformation follows a J-curve: performance dips while the organization redesigns itself around the new capability, then compounds. Clean ROI in the first quarter usually means you are automating around the edges, not transforming. Plan for the dip, measure the right things during it, learning speed and new capability rather than next quarter's margin, and hold the course. The companies that quit in the trough are the ones that waste the investment.
Get to a real position in weeks, not quarters.
A 30-minute conversation, your real situation on the table. No pitch.