Companies are putting artificial intelligence (AI) into core business processes and expecting it to take a greater role in decisions, but few say they have achieved their main goals from the technology.
A Kyndryl survey of 1,100 senior business and technology leaders reveals that 57% of organisations have embedded AI in core processes or deployed it broadly across the enterprise. However, only 32% say they had achieved at least one of their two main AI objectives, while 11% had met both.
The gap is increasingly about how companies organise work around AI rather than whether they can deploy the technology. Just 23% of organisations Kyndryl surveyed believe their workforce was fully ready for AI, down six percentage points from a year earlier. At the same time, 79% say the speed of AI development would outpace their workforce, governance and operating models.
“AI’s ability to reshape work is challenging organisations to reshape their workforce more rapidly than ever before. The leaders pulling ahead are aligning skills, roles and decision-making with how work is actually changing,” says Mark Paulek, Kyndryl’s chief human resources officer.
Preparing workers for AI
The pressure is set to grow as companies use more AI agents, which are being introduced to carry out work that previously required employees to complete multiple tasks or make routine decisions. According to Kyndryl, 81% of organisations expect AI agents to make impactful decisions within the next year. Yet, only 25% completely trust AI systems to operate without human oversight.
See also: China pushes all school levels to teach AI in Xi’s tech drive
Separately, AvePoint’s survey of 750 leaders responsible for information management, data security or AI programmes states that 46.9% of employees already use AI agents daily or weekly. They are using the systems to reduce manual effort, shorten process times and redirect employees towards higher-value work. Cutting headcount ranked last among the reasons organisations gave for adopting agents.
The surveys point to a more immediate challenge for executives: converting AI use into measurable business outcomes while changing jobs, training staff and setting clear limits on what the technology can do.
Kyndryl’s study reveals that 61% of organisations have redesigned roles for AI-enabled work, while 24% are creating new roles focused on managing AI. Still, 52% of leaders cite difficulty in finding employees with the skills needed to advance their AI strategy. Only a third have fully implemented training programmes to help employees work effectively with AI tools.
See also: OpenAI limits release of new model under pressure from US
A smaller group of companies appears to be making more progress. Kyndryl identified 9% of respondents as “Pacesetters” because they had redesigned roles around AI, managed the workforce transition and built employee readiness. Those organisations were 1.5 times more likely to report AI-related revenue growth and 1.6 times more likely to report improved innovation in products and services.
Lagging guardrails
Many companies have yet to put in place clear guardrails for more autonomous systems. Only 33% of Kyndryl respondents say they had policies defining which decisions AI can and cannot make, while 27% say they used a registry and monitoring capability for all AI systems.
Meanwhile, AvePoint respondents share that weak oversight is already contributing to deployment delays. About 86% of organisations had postponed AI agent rollouts by almost six months on average because of data security and management concerns. The figure was 86.9% for generative AI deployments.
The survey reveals that 21.1% of organisations did not know whether employees were using unsanctioned tools to create AI agents. Another 17.6% lacked visibility into whether workers were using unauthorised generative AI tools, nearly triple the 6.3% reported in 2025.
Companies’ confidence in their controls did not always match reported incidents. While 82.7% of AvePoint respondents say they were very or extremely confident in their ability to prevent unauthorised AI-related data access, 72% of the “very confident” group and 62% of the “extremely confident” group say they had experienced an incident in the past year.
Almost nine in 10 organisations reported at least one AI agent-related security incident during the period, too, according to AvePoint's report.
To stay ahead of the latest tech trends, click here for DigitalEdge Section
Data adds to the challenge
The data challenge is becoming larger as AI contributes more to the information companies need to manage.
AvePoint states that 35.5% of enterprise data is now generated by AI assistants, a share expected to rise to 42.1% within 12 months. More than four-fifths of respondents manage at least one petabyte of data, while 78.1% say at least half of their information was more than five years old.
“Nearly half of global employees are already relying on AI agents weekly or daily, and organisations are deploying agents faster than they are building the foundations required to trust them. The constraint on enterprise AI is no longer model capability, but whether organisations have built a trust layer: the data visibility, governance, and enforceable control required to scale AI with confidence. Without it, speed of deployment becomes speed of exposure,” says Tianyi Jiang, AvePoint’s chief executive officer and co-founder.
