Our Take
A headline question is not reporting—until FT shows us the math, we don't know whether this is an investigation with findings or a provocative framing with no answer.
Why it matters
CFOs and boards are committing billions to AI infrastructure and projects while actual ROI remains opaque. This matters now because we're past the pilot phase and into real budget allocation.
Do this week
Finance team: Document your AI project's cost per unit of output (inference cost, processing time, or error rate reduction) before your next steering committee review, so you can answer the question yourself.
FT asks the hard ROI question
The Financial Times is asking directly: how much value is AI actually creating? The question implies a gap between AI spending and measurable returns. Without access to the full article (it is paywalled), the scope of FT's reporting remains unclear. The headline suggests an investigation into whether AI investments deliver on their promised business value, or whether the market is pricing in returns that don't yet exist.
Enterprise buyers need proof, not promises
This question arrives at a critical moment. Enterprises have moved from proof-of-concept budgets to capital allocation. IT leaders and CFOs are defending billion-dollar spending decisions without standard frameworks for measuring AI ROI. Unlike traditional software or infrastructure, AI value is often claimed in abstract terms (efficiency, speed, quality) rather than concrete metrics. A major publication asking publicly how much value AI is creating signals that the market itself is starting to demand specifics.
Measure your own value
If you own an AI initiative, stop waiting for the industry to define ROI. Map your project to one concrete outcome: reduced processing time per transaction, cost per classification, or error rate reduction. Give it a number and a baseline. When leadership asks the question FT is asking, you will have an answer that is not a forecast.