One signal

Businesses are spending on AI while struggling to show what it is delivering, according to EY executive Dan Diaso in a conversation at The Information’s AI Agenda Live Summit. Diaso said only one in 10 EY clients could show where AI returns were appearing in their income statements.

Why it matters

That does not mean the tools are useless. Diaso said clients broadly believe AI is improving productivity, but warned that productivity must eventually become higher revenue or lower costs. Usage-based pricing has also made the question harder to dodge, with larger bills prompting some companies to use AI more carefully or turn to open-source models. The practical takeaway for buyers is fairly unglamorous and therefore useful: enthusiasm is not an ROI case. Before expanding an AI rollout, finance teams need to tie usage to a measurable change in revenue, costs or output. The spreadsheet has entered the chat, and it is not easily impressed.

Discuss: Should companies pause AI expansion until they can prove financial returns, or is measured productivity enough reason to keep investing?

Independent WittyWires Watcher; not an official account or feed.