Adoption is nearly universal, readiness lags far behind
Skillsoft surveyed 2,000 full-time employees across North America, the UK, and Germany between March and April 2026 and found that 86 percent of them now use AI tools as part of their day-to-day work. That number alone would suggest AI rollout has succeeded at most companies, and many boards are reading it exactly that way in quarterly updates. But only 24 percent of those same employees said they feel fully equipped with the skills to use AI well. Adoption and competence have become two entirely separate metrics, and most executive dashboards today are still only tracking the first one.
The gap between how leaders and employees perceive readiness runs even wider than the adoption numbers suggest. Seventy-seven percent of leaders in the survey believe their people are ready to work with AI, while just 24 percent of employees say the same about themselves. That 53-point spread reflects a structural blind spot in how most organizations measure the success of their AI programs, built on the comfortable assumption that access equals capability. If your executive team is reporting AI readiness up the chain based on tool adoption alone, the real figure is very likely less than half of what gets presented.
The training pipeline behind the gap is thin
The survey's most operationally useful finding is where the gap actually comes from. Only 11 percent of employees receive a formal skills assessment for AI, meaning most organizations have no measured baseline of what their workforce knows before or after a rollout begins. Just 16 percent get any training before a new AI tool is introduced to them. In practice, the overwhelming majority of enterprise AI deployment today happens with no measured starting point and no structured onboarding sequence, just access granted and usage assumed to follow naturally from there.
Fifty-nine percent of employees cited lack of time as the primary barrier to building AI skills, ahead of any complaint about content quality or tool availability. Sixty-nine percent said they lack clarity on which AI skills actually matter for their specific role, while only 43 percent of leaders believe their own guidance on this question is clear enough to act on. That combination, no baseline measurement, no structured onboarding sequence, and no organizational consensus on which skills matter most, explains why usage keeps climbing while confidence stays essentially flat quarter over quarter.
Distrust and job anxiety are quietly building underneath adoption
Twenty percent of employees told Skillsoft they remain cautious about or actively distrust AI tools, and 31 percent said the guidance they receive differs depending on which manager or team they happen to sit on, meaning inconsistent internal policy is actively compounding the measured skills gap. Twenty-nine percent expect AI to reduce the number of entry-level positions at their company, a figure CxOs should read as a leading indicator of retention risk among the junior talent they will need to have trained up over the next several budget cycles.
Ciara Harrington, Skillsoft's Chief People Officer, put the core finding plainly, saying organizations cannot afford to confuse AI adoption with AI readiness. Her prescription, treating skills as a business discipline instead of an assumption baked into a rollout plan, is the right frame for any CHRO to adopt. Forty-five percent of employees and 46 percent of leaders both described current training as mostly confidence-building for existing roles, a finding that suggests most current training budgets are aimed squarely at the wrong target given how fast those roles are actually changing underneath them.
What CTOs and CHROs should change this quarter
The fix Skillsoft's data points toward runs through measurement infrastructure that most enterprises have skipped entirely in their rush to grant tool access. That means running baseline skills assessments before a tool rollout rather than after adoption has already been declared a success internally. It means sequencing structured training before broad access rather than leaving it optional and self-paced once employees are already using the tool unsupervised. And it means publishing role-specific skill definitions centrally so managers stop giving their teams conflicting, ad hoc guidance that undermines consistent capability building across the organization.
For any CxO currently reporting AI readiness metrics up to the board, the honest move is resetting the baseline using assessment data instead of adoption data, even when the resulting number is uncomfortably low compared to what leadership has assumed all year. A 24 percent readiness figure makes for a worse headline than an 86 percent adoption figure in a board deck, but it is the number that actually predicts whether an AI investment will produce the productivity gains it was originally budgeted to deliver over the coming fiscal year.
The measurement gap will show up in your AI ROI numbers first
This readiness gap has a direct line to the AI return-on-investment conversations most boards are already pressing executives on. If 86 percent of a workforce uses AI tools but only 24 percent uses them competently, the productivity math finance teams built into last year's AI business case is very likely overstated, because it assumed capability tracked adoption. Skillsoft's data gives CFOs and CTOs a concrete reason to revisit those projections with harder assessment evidence before the next budget cycle locks in spending based on numbers nobody has actually verified against real skill data.
Organizations that get this right will treat the 53-point gap as an operating metric to close deliberately, not a survey footnote to acknowledge and move past. That means budgeting for assessment infrastructure alongside tool licensing costs, not after them, and it means holding vendors accountable for readiness outcomes rather than just seat activations. The companies that close this gap fastest will be the ones that actually capture the productivity gains their AI investments were supposed to deliver in the first place, while competitors keep reporting adoption numbers that mask a readiness problem underneath.
