Half of big-company CEOs say AI drives their productivity. Only 16% can see what it returns.

EY-Parthenon published its latest CEO Outlook Survey on 1 October, and Postimees brought it to Estonian readers yesterday. EY surveyed 1,200 chief executives of large companies in 21 countries during August and September. Eleven of those countries are European, Finland, Sweden and Germany among them. Estonia is not in the sample, but the findings describe a problem we see in Estonian companies every week.

Half of the CEOs named AI as the single largest source of their productivity gains over the past 12 months. Business process redesign came second at 46%. So the technology works, at least in the eyes of the people who run these companies. The trouble starts when the gain has to show up in the accounts.

The gain is real. The proof is missing.

Only 16% of the CEOs say they measure AI returns with confidence and see AI costs and returns in real time. Another 24% can track some of the costs but cannot judge the overall return. Almost a quarter, 23%, say AI gains are often absorbed before they reach the bottom line. EY names legacy systems, organisational complexity and regulatory requirements as the factors still holding companies back.

In plain terms: an employee finishes a report in two hours instead of five, and the three saved hours disappear into more meetings, more approvals and more checking. Nobody wrote down what the old process cost, so nobody can show what the new one saves. EY’s own summary makes the same point. Productivity alone is not a strategy, and the work now is turning saved time into measurable value.

Skills, more than headcount

The workforce figures matter most for a company planning next year. Four in five CEOs, 80%, expect AI to change roles, skills and ways of working more than it changes the size of the workforce over the next three years. Nearly half, 47%, say their organisation is not building AI skills fast enough. And 72% expect skills shortages to become a bigger barrier to growth than access to capital within three years.

Then comes the gap. A third of the CEOs, 33%, already call reskilling a key driver of productivity. Only 15% say they reinvest productivity gains in developing their workforce. Most companies agree that skills decide the outcome, and most are not spending the gains on them.

What an Estonian company should do on Monday

• ✳Pick the two or three AI uses your teams rely on most and write down how long each task took before AI and how long it takes now. A rough number recorded today beats a perfect one you never collect
• ✳Decide in advance where the saved hours go: faster customer replies, a backlog cleared, a new service tested. Hours you do not assign will be absorbed
• ✳Put the AI licence, API and support costs in one line of the budget, so that cost and return sit on the same page
• ✳Look for the approval step or legacy system where the time savings stall, and fix that before buying another tool
• ✳Set aside a share of what AI saves for training. If 47% of large companies say their people are learning too slowly, a smaller company without a training budget is unlikely to be ahead

The survey covers corporations with revenues mostly above half a billion dollars, and they still cannot see what AI returns. An Estonian company with fifty staff has one advantage: it can measure three tasks by hand next week and know its answer before the big players do.