Building an agentic AI strategy that pays off - without risking business failure

Agentic AI promises transformational gains—Accenture calls it a new type of capital, KPMG estimates $3 trillion in annual productivity. But a new ZDNET feature by David Gewirtz urges caution. Gartner found that over 40% of agentic AI projects will be canceled by 2027 due to escalating costs, unclear business value, or inadequate risk controls. Many vendors engage in 'agent washing,' rebranding traditional chatbots and RPA as agentic. In reality, fewer than 13% of claimed agentic vendors ship true autonomous systems.

Costs are another hidden trap. Unlike simple Q&A queries, agentic loops can consume thousands of tokens per task, leading to runaway API bills from providers like OpenAI, Google, or Anthropic. Gewirtz advises executives to prioritize measurable outcomes over hype—start with small proof-of-concept projects that can pay for themselves before scaling. The smartest strategy balances ambition with rigorous risk controls to avoid joining the 40% failure statistic.