While the enthusiasm for AI is still over the top, business leaders are tamping down their expectations of it. CEOs, for one, are far less likely than a year ago to see AI as a growth engine. But yet, they continue to have high expectations for the year ahead.
That’s the word from a new study of CEO attitudes published by the IBM Institute for Business Value, which finds that while 49% of top executives expected advanced AI – then defined as generative AI – to be driving business growth within a two-year period. This year, that number – now applied to agentic AI – dropped to 10%.
Still, they see rapid progress on the horizon, with 74% expecting agentic AI to be delivering accelerated business growth by 2030. The IBM study is based on a survey of 2,000 CEOs from 33 geographies and 21 industries between February and April 2026, along with qualitative discussions.
The value of AI is shrouded in mystery. Only 24% of business leaders admit they know how that value will come about. But they need to get much more involved than they have been when it comes to getting everyone on board with AI and steering it in the right direction.
It comes down to vision, and this is something business leaders need to develop and propagate, the study’s authors urge. Importantly, such a vision needs to be built on customized data and models, not the large pre-trained models that will deliver sameness across companies. In the study, “CEOs who have defined a tailored AI vision are more optimistic about product and service innovation.” The future, they say, will be built on products and services not offered today, made possible through AI.
Such a successful AI vision comes from within, incorporating “smaller, tailored models or a combination of custom and foundation models” – not large pre-trained models, they explain. Business leaders acting on more customized AI models “expect 24% greater productivity gains, 55% higher operating profit margin improvements, and twice the reduction in process cycle times by 2030 compared to those relying predominantly on large pre-trained models.”
Successful business leaders are also recognizing that AI-based transformation means more than slapping AI on top of existing processes. “Today, AI augments people. By 2030, people will augment AI,” the study’s authors state. “The biggest shift won’t be structural – it will be cultural.”
This calls for human and machine collaboration on a scale not seen before. “Many organizations are still focused on using AI to support traditional ways of working, but our analysis shows that those who are proactively rethinking how people and technology collaborate are already seeing better results,” the study’s authors point out. This means more cross-functional teams exploring AI’s possibilities. Executives who promote this “are more than twice as likely to have delivered on their business objectives.”
When AI handles specialized tasks with speed and precision, “the human
advantage shifts to those who can see patterns across functions, ask questions AI doesn’t know to ask, and integrate machine intelligence with strategic context.”
The study’s authors provide the following pieces of advice:
- Create new roles at every level to take advantage of AI. Assign ownership and authority for AI leadership. Rethink who should have authority over what area of the business—then give those leaders, especially COOs and business line leaders, the power to change how work is done.”
- Appoint a chief AI officer – or similar role. Appoint a chief AI officer “with real authority to orchestrate transformation enterprise wide.”
- Redesign decisions. Look for “the handful of enterprise decisions that slow everything else down – AI deployment, pricing moves, capital shifts, partner selection,” they advise. Then, as suggested above, follow through with “a single owner, explicit authority, and clear escalation rules” for each deployment.
- Make leaders accountable for AI outcomes, and reward accordingly. AI success needs to be measured via margins and customer trends, and tied to leaders’ compensation. “Make sure people and technology decisions are made together, measured together, and delivered together.”
- Give AI leadership authority—within limits. This is likely to be in the domain of chief AI officers. “Clarify their mandate. Give them authority over AI priorities, standards, and funding gates—but not ownership of business results. Their job is to accelerate decisions, scale what works, and stop what doesn’t. Line leaders remain accountable for outcomes. That separation is what enables speed without chaos.”
- Reward AI success. Indeed. “Ask executives to reevaluate partners based on how easily third-party AI agents can discover, integrate with, and transact through your offerings. Identify shared growth plays, joint use cases, and reciprocal pipeline creation and prioritize the partners best equipped to scale with you.”
