Businesses are seeing stronger returns from artificial intelligence investments, but their ability to deal with the risks is falling behind, according to a new global report by technology services provider Kyndryl.

The company’s third annual Readiness Report found that 64 per cent of business leaders reported positive returns from their AI investments, an increase of 10 percentage points compared with 2025.

However, the average level of preparedness for external risks fell by seven percentage points to just 24 per cent, with the sharpest declines recorded in cybersecurity and companies’ ability to respond to changes in competition.

The findings are based on responses from 3,700 senior executives across 21 countries and point to a growing gap between the commercial benefits of AI and the ability of businesses to manage its wider implications.

The challenge is particularly relevant for Cyprus, where sectors such as financial services, energy and professional services increasingly rely on technology to support their operations.

Marios Kapiris, general manager of Kyndryl Cyprus, said the findings emphasised a significant challenge for the market, as organisations were increasingly recognising the potential of artificial intelligence and investing in its adoption, although their technological and organisational readiness was not necessarily progressing at the same pace.

“For an economy such as Cyprus, with strong business activity in sectors such as financial services, energy and professional services, the challenge is not limited to the adoption of artificial intelligence, but concerns its safe and effective integration into everyday operations,” he said.

Kapiris explained that this required modern technological infrastructure, strong cybersecurity and operational resilience mechanisms, as well as the continuous development of workforce skills.

He added that “organisations that succeed in combining these elements will be better placed to turn their investments in artificial intelligence into meaningful and sustainable business results.”

The report nevertheless shows that businesses are making progress in moving AI beyond the experimental stage and into everyday operations.

The proportion of organisations still experimenting with the technology fell by 13 percentage points to 49 per cent, while those reporting more pilot projects than they could expand on a larger scale declined by 11 percentage points to 61 per cent.

This transition is also prompting companies to reconsider how they use AI.

Nearly half of respondents, 48 per cent, revised their AI strategies during the year, introducing changes affecting technology, employees and governance arrangements.

At the same time, 24 per cent reported that an AI investment had failed, while 21 per cent had cancelled a major AI initiative.

The changes are proving particularly difficult when it comes to employees and the organisation of work.

Only 27 per cent of businesses have formally revised roles and responsibilities to reflect AI adoption, while just 29 per cent have fully implemented change-management procedures.

The findings show that although businesses are increasingly putting AI into operation, changes to their workforce structures and internal processes are progressing more slowly.

Cybersecurity remains the biggest concern.

For the third consecutive year, security topped the list of external risks facing businesses and was the main contributor to the decline in overall preparedness.

Just 26 per cent of companies considered themselves fully prepared to deal with cyberattacks, down 11 percentage points from 2025.

This comes despite three in four executives reporting that AI helps their organisations detect security threats more quickly.

The figures reveal a growing contrast between the technology’s ability to identify potential threats and companies’ confidence in their capacity to respond effectively when attacks occur.

Businesses are also becoming less confident about their ability to respond to changes in the competitive environment.

Readiness to deal with competition from businesses and technologies outside their traditional sectors fell by 10 percentage points to 23 per cent, the second-largest decline recorded in the report.

The growing use of AI agents and other emerging technologies is increasingly challenging established boundaries between industries, forcing companies to reconsider existing products, services and business models.

Retailers, for example, are examining how AI-powered shopping assistants can be incorporated into their operations without losing direct contact with customers.

Software companies, meanwhile, are reassessing their software-as-a-service models as AI changes the balance between developing technology internally and purchasing ready-made solutions.

Financial institutions are also facing new lending models, adding to the pressure on established ways of doing business.

Another concern is whether existing IT infrastructure can support AI on a large scale.

For the third consecutive year, upgrading technology infrastructure remained the most important measure businesses reported taking to reduce their exposure to external risks.

Although 71 per cent of executives described their IT systems as largely or fully modernised, 58 per cent reported that significant architectural redesign was needed to support AI at scale.

The distinction becomes even clearer among chief information officers, with around two-thirds of CIOs saying their technology infrastructure would require substantial redesign before artificial intelligence could be fully deployed on a larger scale.

The findings underline the difference between having modern IT systems and having infrastructure capable of meeting the additional demands of artificial intelligence.