
AI adoption has moved beyond experimentation and into its proof phase. IT leaders are being asked to deliver measurable returns on their investments, yet many organizations are struggling to translate AI adoption into meaningful business outcomes. Part of the struggle is that many leaders don’t yet have a clear way to measure what AI success looks like — making the question of how to track outcomes a gap in its own right.
One of the main reasons is a perception gap between leadership and employees on AI readiness. Recent research shows that while 77% of business leaders believe they have set their teams up for success with AI, fewer than one in four employees feel equipped to use the tools their organizations have given them.
The tools themselves aren’t the problem, and throwing money at newer, faster tools isn’t the solution. Employees are left to figure out the nuances of these AI tools on their own without guidelines on how to judge final outputs. That disconnect suggests AI deployment is moving faster than workforce readiness.
When AI Management Becomes a Second Job
For many employees, AI has created an unofficial second role: AI manager. Alongside their existing responsibilities, they must decide which tasks are appropriate for AI, provide the right context, refine prompts, monitor the process and review the results for accuracy.
The burden grows as the speed of AI increases the volume of work. Employees may be expected to take on more tasks, even if they don’t have the expertise to assess the output. That added cognitive load can contribute to burnout and make careful review harder to sustain. If someone uses AI to complete work beyond their own understanding, weak assumptions can look convincing and pass without scrutiny.
That is how AI-generated workslop enters a workflow: output that appears finished but lacks the context, judgment or accuracy needed to be useful. It moves downstream, leaving the next person to verify the work, correct it or determine how the conclusion was reached.
There’s a retention cost hiding in all of this. When people spend their days babysitting output they don’t fully understand and cleaning up work that only looks finished, the job stops feeling like the one they were hired to do. The most capable employees notice first. They came to solve problems and build expertise, not to serve as quality control for a system no one can fully explain. Left unaddressed, that erosion of engagement shows up as quiet disengagement long before it shows up as attrition.
The risk can remain hidden until a developer needs to trace a bug through a codebase, a client questions a recommendation or a regulator asks how a decision was made. By then, the initial productivity gain may be outweighed by the cost of rework, delays and lost trust.
For IT leaders, the problem spans adoption, workforce readiness, governance and the engagement of the people doing the work. Their responsibility cannot end once employees have access to secure, approved tools. It must also include the guardrails and capabilities people need to use those tools well without burning out in the process.
Skills Management as a Solution
AI has made critical thinking, collaboration, and creative problem-solving even more important attributes. And like any technical skill, they need to be continuously sharpened. Once employees begin to outsource these attributes to AI in the name of productivity, they risk losing the very qualities that got them hired in the first place.
Companies should manage workforce skills with the same rigor they apply to finances and customer relationships.
Too often, organizations treat skills management as an inventory of who can do what. Effective skills management starts with business priorities, identifies the workflows needed to achieve them and maps the human and AI capabilities those workflows require. That visibility helps leaders identify which tasks can be delegated to AI, where human oversight is needed and which skills gaps could affect performance. It also gives leaders a clearer way to measure progress — connecting the capabilities they build to the business outcomes they’re trying to reach, so AI success becomes something they can actually track rather than assume.
Developing these capabilities across teams also reduces reliance on any one employee to oversee AI-supported work or correct problems when they arise.
As AI models and workflows evolve, the skills needed to use and evaluate them must evolve just as quickly.
IT leaders are in a strong position to connect AI governance, deployment and skills management. By pairing each rollout with clear rules, hands-on practice and ongoing verification of workforce capability, they can reduce rework, keep technical talent engaged and give their organizations a better chance of realizing measurable returns from AI.



