briefings

The Fallacy of ROI

The same meeting recurs every decade: a slide of projected costs, a question about return on investment, a decision to wait for more evidence. ERP, mobile, cloud, now AI. The technology changes. The meeting does not.

The ROI question sounds rigorous but cannot be answered in the timeframe it is asked, so the decision can be deferred indefinitely without anyone saying no. It does two things at once: it legitimises engagement with an unfamiliar technology, and it brakes commitment to it.

It is a comprehension signal, not a financial one. Asking for the return on AI reveals that the organisation still treats infrastructure as a discretionary project to be evaluated, rather than as something work will run on. Projects get evaluated. Infrastructure gets absorbed.

The move is to change the question. Not what is the return, but what would have to be true for this to be infrastructure in three years, and what must we build now to be ready.

Read the full argument in The Fallacy of ROI.