Kyndryl’s latest Readiness Report, published for the third consecutive year, finds that companies are increasingly capturing value from artificial intelligence (AI) while simultaneously feeling less prepared to manage the external risks that AI introduces. The study summarises responses from 3,700 senior executives across 21 countries.
Key findings
- 64% of surveyed leaders report positive returns from their AI investments, a 10 percentage-point increase compared with 2025.
- The share of organisations that consider themselves fully prepared to address the external risks examined fell to an average of 24%, a 7 percentage-point decline year-on-year.
- Although 71% of leaders say their IT environments have been largely or fully modernised, 66% of CIOs believe substantial changes to technology architecture are still required to support enterprise-scale AI.
Risks and vulnerabilities
Readiness has declined particularly in these areas:
- Cybersecurity: only 26% of companies feel fully prepared to handle cyberattacks, an 11 percentage-point drop from the previous year. Three-quarters of respondents say AI can speed threat identification, yet confidence in their organisation’s ability to respond to attacks has fallen.
- Competitive disruption and changes to traditional business models: readiness to manage competition from outside traditional industry boundaries fell to 23%, a 10 percentage-point decrease from 2025. AI agents and other new technologies are blurring industry lines, forcing businesses to rethink product viability and business models in sectors such as retail, software and finance.
Deployment pace and strategy shifts
- The share of organisations still in an experimental AI phase dropped by 13 percentage points to 49% versus 2025.
- 48% of organisations significantly revised their AI strategies in the past year, with changes affecting technology choices, workforce issues and governance.
- 24% of respondents experienced a failed AI investment, and 21% stopped a major AI project.
Workforce, roles and change management
Organisational changes needed for broad AI adoption are lagging behind technology deployment:
- Only 27% have formally redesigned job roles and responsibilities to reflect AI usage.
- Just 29% apply comprehensive change management to support AI adoption. Those organisations making progress on both fronts are 28 percentage points more likely to consider themselves fully prepared for AI deployment and 8 percentage points more likely to report positive returns.
Digital sovereignty and trade-offs
Respondents show rising interest in digital sovereignty: 68% would accept at least a 10% premium for solutions that offer full operational control when needed. Additionally, 60% see increased control as advantageous even if it slows the adoption of the most advanced capabilities. Notably, executives in organisations operating in the EU reported taking concrete digital sovereignty measures slightly less often than peers in other regions.
Modern IT is necessary but not sufficient
For the third year running, leaders identify IT modernisation as a top means to mitigate external risks; 71% say their IT environment has been largely or fully modernised. Yet two-thirds of CIOs say the underlying architecture needs significant overhaul to truly support enterprise-scale AI. The results highlight that modern infrastructure alone does not equate to AI readiness—flexible architectures, updated roles, effective governance and strengthened cyber defences are also required.
Comment from Martin Schroeter
"Artificial intelligence creates significant opportunities, and organisations are now focused on ensuring early results can be sustained over the long term," said Martin Schroeter, Chairman and CEO of Kyndryl. He noted that AI ambitions are growing faster than the systems and capabilities needed to deliver them, and emphasised the importance of workforce readiness, technology foundations and governance frameworks.
Conclusion
Kyndryl’s report shows a clear tension: companies are accelerating AI adoption and reporting improved returns, but their ability to manage attendant risks is weakening. Achieving durable AI success demands more than technology investment alone—it requires people, processes, governance and architecture to evolve in step with AI’s expanding role in enterprise operations.
