
After years of consulting startups and boards on technology, I often hear one trope: that technology is just a set of dumb machines. As a result, most executives focus on faster responses, lower costs, and greater output. These are legitimate objectives. Yet they leave out a question that deserves equal attention: What kind of organization are we building when we change how people work? AI is bringing this question to the discussion we see on employee websites like Blind, LinkedIn, or the Y Combinator forum.
AI implementations are not simple technology decisions. The projects redistribute responsibility, change expectations, and signal which contributions leadership values. A company that celebrates automated output while overlooking human judgment teaches employees something very different from one that rewards thoughtful verification. Technology cannot be separated from the environment in which people use it.
Consider an executive who announces that AI will give employees more time for meaningful work, then immediately raises production targets without discussing workloads or expectations. The announcement promises empowerment. The operating model delivers pressure. Employees will judge the implementation by the latter.
That’s why many of my conversations start with the corporate culture that should guide these decisions, not the communication plan that defends it after it’s too late. Unless leaders examine how the proposed system will affect autonomy, collaboration, learning, and accountability before deciding what successful adoption looks like, your team will start losing its identity and trust.
AI Cannot Install Trust
Think of the Employee experience. Our team members do not experience AI as an abstract strategic opportunity. They experience a new tool, a changed performance standard, or uncertainty about their future. Often all three combined. Ignoring any of these perspectives creates an avoidable gap between leadership’s intentions and employees’ interpretation.
In 2023, the OECD surveyed employers and workers in manufacturing and finance across seven countries and found generally positive views of AI’s effects. They also found that training and worker consultation were associated with better outcomes. This is an association, not proof that consultation alone produces success, but it gives leaders a practical reason to involve employees.
Meaningful involvement must allow people to influence implementation before plans are finalized. Asking for feedback after a workflow has been implemented is not shared decision-making. Employees should be able to identify where automation helps, where it creates additional checking, and where a human relationship remains essential.
Honesty matters as much. Leaders should not promise that roles will remain unchanged when they cannot support that claim. Nor should they describe workforce reductions as something a machine independently decided. Mass layoffs are a core responsibility of boards and executives. You cannot outsource the blame to technology while maintaining your culture.
Trust requires ownership of difficult choices. An algorithm does not relieve management of that obligation.
Protect the Judgment Behind the AI Output
A second cultural risk emerges when companies confuse producing work with understanding it. An employee can generate a persuasive recommendation without having tested its assumptions. A manager can receive a polished report without knowing whether its author can defend the conclusion.
AI overreliance happens when we replace human thinking, judgment, and accountability with faster output. Any company must retain the ability to question its tools and function when they fail.
That concern should shape implementation. If AI handles routine analysis, leaders should ask how junior employees will learn to recognize an unusual result. If automated drafting becomes standard, managers should preserve opportunities to develop reasoning rather than merely edit machine-generated prose.
The answer is not to protect every manual task. It is to distinguish unnecessary effort from developmental experience. Some work should disappear. Some should change. Some remains valuable because doing it builds the expertise needed for more consequential decisions.
Training should therefore extend beyond prompting techniques. It should teach employees how to evaluate outputs, identify limitations, and decide when not to use the tool. The OECD similarly places training within a broader framework of transparency, accountability, worker dialogue, and privacy protection.
What Leaders Reward Will Become the Culture
Formal policies explain how employees are supposed to behave. Incentives reveal which behaviors actually matter. An organization can require careful AI review while making that review practically impossible through deadlines and output targets.
Imagine a team measured entirely on the number of customer cases closed. An AI assistant may help resolve straightforward requests. But if employees receive no recognition for escalating ambiguous cases, the measurement system discourages the judgment that responsible implementation requires.
Leadership should therefore review performance management alongside technology deployment. Speed matters, but so do customer outcomes and the willingness to challenge an apparently convincing answer. These measures should reflect the purpose of the work rather than the software’s capabilities.
Managers also need permission to discuss failures without turning every incident into an accusation. Employees cannot serve as an effective line of defense if reporting a problem threatens their standing. Fearful or disengaged employees ultimately weaken organizational resilience during AI adoption.
It means making accountability useful. A concealed error protects nobody. A reported error can expose a faulty workflow, an unrealistic target, or inadequate training. Leaders should reward that information rather than punish its arrival.
The Board Must Look Beyond AI Adoption
Boards should resist treating license purchases, usage rates, and projected savings as a complete account of AI progress. None establishes whether employees retain confidence, understand their responsibilities, or feel able to question an automated recommendation.
Management reporting should connect commercial results with evidence about learning, workload, trust, and decision quality. Directors should also ask who owns the consequences when a system performs poorly. Responsibility must remain identifiable.
AI can strengthen a company without weakening its culture. But that outcome requires deliberate choices about participation, incentives, and human development. The decisive question is not whether people use the technology. It is whether using it makes the organization more capable and whether leadership can explain why.

Leave a Reply