Kyndryl Finds AI Returns Rising, Risks Persist

A Kyndryl report published Tuesday found that senior business leaders reported higher returns on AI investments year over year, while their readiness to manage external risks declined. Kyndryl surveyed 3,700 senior business leaders across 21 countries. The report said many enterprises are struggling to move from AI experimentation to broader adoption, particularly in workforce planning: fewer than one-third of technology leaders said they had formally updated job roles for AI responsibilities or introduced an AI adoption change management system. Kyndryl CIO Kim Basile said CIOs need to work with other C-suite leaders and bring employees along so they can make a difference with the technology.

Security was the leading external risk concern in the survey, as technology leaders said they felt ill-equipped to handle changing business models, technology foundations and threats such as cyberattacks. The report linked those concerns to incidents including a breach of open source platform Hugging Face and Anthropic Claude models escaping their test environment. Basile said cybersecurity is advancing quickly, while threat actors are also moving faster and using technology to increase the number of threats. She said increasingly interconnected systems and AI's greater operational role make visibility into systems and understanding vulnerabilities more important. Leaders also raised concerns about AI changing industries and business models beyond their own, as well as digital sovereignty's role in the evolution of technology stacks.

Enterprise AI adoption had matured during the year, the report said: fewer than half of surveyed technology leaders remained in the experimentation phase, and nearly half had substantially changed their AI strategies across technology, people and governance. However, 21% said they had canceled a major AI initiative after it failed. Although 71% said their IT infrastructure was nearly or fully modernized, two-thirds of CIOs believed their technology foundations still needed significant work before they were ready to scale AI. Kyndryl Consult SVP Ismail Amla said enterprises need security, transparency, auditability and updated infrastructure, while many of their main challenges involve non-technical work. He said organizations must consider whether employees working with new systems are properly incentivized and trained, and understand what working with agents means; he added that people, process and technology now involve a much more complex human-change layer.

A Kyndryl report published Tuesday found that senior business leaders reported higher returns on AI investments year over year, while their readiness to manage external risks declined. Kyndryl surveyed 3,700 senior business leaders across 21 countries. The report said many enterprises are struggling to move from AI experimentation to broader adoption, particularly in workforce planning: fewer than one-third of technology leaders said they had formally updated job roles for AI responsibilities or introduced an AI adoption change management system. Kyndryl CIO Kim Basile said CIOs need to work with other C-suite leaders and bring employees along so they can make a difference with the technology.

Security was the leading external risk concern in the survey, as technology leaders said they felt ill-equipped to handle changing business models, technology foundations and threats such as cyberattacks. The report linked those concerns to incidents including a breach of open source platform Hugging Face and Anthropic Claude models escaping their test environment. Basile said cybersecurity is advancing quickly, while threat actors are also moving faster and using technology to increase the number of threats. She said increasingly interconnected systems and AI's greater operational role make visibility into systems and understanding vulnerabilities more important. Leaders also raised concerns about AI changing industries and business models beyond their own, as well as digital sovereignty's role in the evolution of technology stacks.

Enterprise AI adoption had matured during the year, the report said: fewer than half of surveyed technology leaders remained in the experimentation phase, and nearly half had substantially changed their AI strategies across technology, people and governance. However, 21% said they had canceled a major AI initiative after it failed. Although 71% said their IT infrastructure was nearly or fully modernized, two-thirds of CIOs believed their technology foundations still needed significant work before they were ready to scale AI. Kyndryl Consult SVP Ismail Amla said enterprises need security, transparency, auditability and updated infrastructure, while many of their main challenges involve non-technical work. He said organizations must consider whether employees working with new systems are properly incentivized and trained, and understand what working with agents means; he added that people, process and technology now involve a much more complex human-change layer.