Business cases are decision instruments. Their purpose is not to demonstrate that a technology could create value. Their purpose is to show how the organization will turn a change in work into an economic outcome.
That means every significant benefit assumption needs an operating mechanism behind it.
Four categories of AI benefit
- Cashable: spend can actually be removed.
- Avoidable: future spend or hiring can be prevented.
- Redeployable: capacity is moved to other valuable work.
- Strategic: better decisions, growth, quality, risk or customer outcomes create value that may not be direct cost reduction.
The categories can coexist, but they should not be added together as though they were identical.
The most common FTE fallacy
If 1,000 employees each save 10 minutes per day, the arithmetic produces a large FTE-equivalent number. But if those minutes are spread across roles and shifts, no headcount may actually be removable. The capacity may still be valuable—but it is not automatically cashable savings.
Why benefit ownership must exist before approval
Technology teams can deliver the capability. Finance can validate the arithmetic. Only the operating business can usually make the decisions that convert capacity into value: staffing, workload, service levels, vendor spend, pricing, growth or process redesign.
The 12-point AI Savings Reality Test
- Is each benefit labeled cashable, avoidable, redeployable or strategic?
- Is the baseline observable and agreed?
- Is the productivity assumption evidenced?
- Is adoption ramp included?
- Are AI run costs included?
- Are support and governance costs included?
- Is the capacity concentrated enough to act on?
- What exact operating action realizes the benefit?
- Who has authority to make that action happen?
- When should the benefit appear?
- How will realization be measured after go-live?
- What happens if benefits do not materialize before the next investment stage?
Research and further reading
- Gartner — Three Pillars for Deriving Value from AI — 9 Mar 2026. Gartner reports only 44% of organizations had adopted financial guardrails or AI FinOps practices, despite rapid expansion of AI deployment.
- Gartner — CFO AI investment survey — 20 Jul 2026. In a survey of 204 finance leaders, 45% of finance AI investments leaned toward productivity while 20% leaned toward decision quality; Gartner warns that productivity-heavy portfolios may fall short of board expectations for enterprise value.
- BCG — Making AI Productivity Pay Off — 5 May 2026. BCG argues that productivity gains do not automatically become lower cost or better performance; capacity must be deliberately redirected and work redesigned.
- Gartner — AI ROI Requires a Focus on Value, Not Feasibility — 22 Apr 2026. Public abstract: deployment processes can optimize technical feasibility while neglecting business value; value alignment must be explicit before scaling.