Architects of intent: How CIOs can convert AI into enterprise value

PwC argues that most organizations have accumulated AI tools, agents, and proofs of concept without converting that activity into measurable EBITDA impact—framing this as an AI value problem, not an adoption problem. The central prescription is for CIOs to reposition themselves as 'architects of intent': redesigning workflows from scratch around AI as the default task owner rather than inserting AI into existing processes. Human roles are not eliminated but refocused toward defining outcomes, resolving exceptions, and maintaining accountability, with the article arguing that the human is not merely 'in the loop' but the reason the loop exists. Governance is treated as a performance discipline: CIOs must continuously monitor model performance, autonomy levels, compute costs, and reliability, selecting models by business requirement rather than capability novelty. AI investment should be managed on a value-acceleration path tied directly to cost-per-transaction, cycle times, and revenue capacity—not judged solely on whether IT spending declines year over year. CIOs who cannot connect technical performance to economic outcomes risk accumulating more pilots and costs with no clearer path to enterprise value.