Can we prove what AI is worth to the business?

• Who owns the benefitIT delivers, the business signs the number
• How many use cases to run at onceLearning across many versus going deep on three

• Productivity or growth casesEasier to measure versus bigger to win

• Portfolio or project fundingStrategic focus versus speed of approval

• When to re-fund, re-scope or retireA fixed review rhythm versus waiting for failure

• Pilots launched without a baseline
• Experimenting everywhere

• Counting licenses or active users as value

• Hours saved that never leave the budget
• Pilots that succeed in isolation and never connect

• Treating AI as a technology programme rather than an operating model change

The starting level. Nothing precedes it.

Nothing in another pillar. It builds on the level before it.

• A portfolio value model needs trusted baseline data.AI-Ready Data Estate at Establishing
• Someone has to own benefits across the portfolio.Organization, Skills & Operating Model at Establishing

• P&L-level value shows up once processes change.AI in the Process at Scaling

Only the next level in this pillar.

• Domains take ownership when business cases give them a reason.Organization, Skills & Operating Model at Scaling

• Redesign at scale is funded by proven value.AI in the Process at Compounding
• Incentives can reward value only once value is measured.Organization, Skills & Operating Model at Compounding

Top of this pillar. Nothing else depends on it.