AI adoption can move faster than employee trust

AI can let a small group turn an idea into a functioning prototype, workflow or product before the wider organization has worked out what the change means for its own work. Leaders can read that technical progress as evidence that people are ready to follow, but the human signals may say something different. Hesitation, repeated questions and attachment to existing processes can reveal where technical capability has moved ahead of the organization’s capacity to absorb it.

Fast AI deployment can create a false signal of organizational readiness

The gap starts with the different speeds of deployment and adaptation. Wagner Denuzzo says a new AI tool can appear and, “Within months,” employees may be expected to adopt it and change how they work; an organization can even install a tool “overnight,” in an illustrative sense. At the same time, AI increasingly allows fewer people to move an idea substantially toward a working prototype, workflow or product. Both developments shorten the technical path from decision to implementation, but human adjustment still takes time.

That difference grows when founders or another relatively small leadership group can advance AI-enabled work without broad participation. Denuzzo argues that requiring fewer people to build something still leaves leaders responsible for engaging the people whose work will change. He has a professional stake in this view: founders have asked him to help develop an unnamed AI leadership experience, so he benefits from demand for guidance on leadership and AI adoption. His argument is that a small group’s progress can become a misleading readiness signal for everyone else.

That misleading signal matters because deployment can get ahead of employees’ understanding and trust. People may conceal confusion, hold back questions or struggle with changes to familiar processes even when implementation appears successful. Leaders can classify those behaviors as obstacles to adoption, but the same behaviors can reveal what the organization has yet to absorb.

AI adoption asks employees to give up established ways of working

What employees must absorb follows from what an AI rollout changes in established work. Existing processes contain habits developed through repetition, expertise built through experience, visibility into how work happens and a sense of control over its completion. When AI alters those processes, employees may have to surrender all four while learning a new system. Adoption involves subtraction as well as addition.

That subtraction explains why training in a new capability may leave important uncertainty unresolved. An employee may learn how to use an AI capability while losing a familiar way to exercise judgment, observe intermediate work or decide whether an outcome is good. Training can explain the added capability while leaving employees unsure what replaces their previous basis for understanding and control. Reluctance can emerge from that unresolved change.

Because established practices carry knowledge and control, human adaptation does not move at deployment speed. Denuzzo describes adaptation as progressing through stages of maturity and says, “Adoption is therefore gradual by nature.” Technical availability and human readiness can occupy different points even when implementation is proceeding as planned. The issue is how employees develop enough understanding to work effectively under the changed conditions.

Trust links that understanding to practical use. “Humans do not trust what they don’t fully understand,” Denuzzo says. An employee changing a familiar process must learn the actions required to operate the AI tool, where the system belongs in the work and how much confidence to place in it. Familiar practices matter because they show how employees currently apply expertise and maintain visibility and control over outcomes.

Those existing practices give leaders information about where a rollout remains unresolved. Reluctance to abandon a process can show that employees still lack an adequate replacement for some form of expertise, visibility or control. If leaders classify every such reaction as a training or attitude problem, they discard that information before understanding it. The problem becomes harder to detect when employees learn that expressing uncertainty carries a cost.

Why apparent compliance can conceal weak AI adoption

Once uncertainty has a professional cost, adoption becomes harder to observe accurately. An employee who thinks admitting confusion will make them appear obsolete has an incentive to conceal it. A dashboard, meeting or manager may then register participation without detecting how little understanding supports it. Visible use can overstate readiness.

Questions can disappear through the same mechanism. If challenging an AI tool is interpreted as resistance, employees learn that raising doubts carries a social or professional penalty, so they may keep using the system while stopping the questions. Visible friction then declines even though the organization has learned less about whether employees understand or trust what they are doing. Apparent compliance becomes a false positive for adoption.

That false positive changes how leaders should read the signals that remain. Hesitation can indicate an unresolved understanding or trust problem, while recurring questions can show that an explanation has failed to become usable knowledge. Attachment to a familiar process can identify expertise, visibility or control that employees do not yet know how to preserve under the changed workflow. Each signal can point to a different constraint on adoption.

Because those signals depend on whether employees feel able to reveal them, psychological safety has an operational effect. When people can disclose uncertainty without shame or judgment, leaders receive information about where learning is weak and where the organization is struggling to integrate the change. When employees expect a penalty, leaders lose access to that information. The culture around a rollout changes the quality of the evidence management uses to judge progress.

That evidence also explains why use alone cannot establish AI literacy, meaning enough understanding to use and judge the technology in context. An employee can operate a system while withholding uncertainty about it, just as a team can stop objecting after learning that objections are unwelcome. Reduced questioning can then look like improved adoption while feedback becomes less complete. Faster implementation magnifies the risk because the small-group progress described earlier can widen the gap before leaders detect it.

Treat resistance as a diagnostic signal

Because visible compliance can hide uncertainty, leaders need to inspect resistance before trying to eliminate it. Denuzzo treats hesitation and repeated questions as diagnostic evidence that can identify gaps in understanding and trust, along with established parts of work that employees are struggling to relinquish. Friction can then inform leadership judgment about a rollout. Leaders still decide how to respond to the evidence.

Denuzzo’s perspective combines organizational leadership with a background in human behavior. Before holding talent and leadership roles at IBM and Prudential, he trained and worked as a licensed psychotherapist, and he now lives in Portugal while watching the AI transition through both lenses. IBM and Prudential are his former employers in this account; they are not presented as case studies of these adoption dynamics. That background helps explain why his recommendations focus on what employees can safely reveal during change.

Making resistance useful requires everyday conditions in which people can reveal what drives it. Denuzzo argues that leaders reduce shame and judgment when they acknowledge what they themselves do not know, can laugh at previous failures and leave room for someone to offer an idea without dismissing another person’s experience. Those norms make uncertainty easier to expose and examine. Leaders can then distinguish a knowledge gap from a trust problem or a concern rooted in how work used to function.

Learning is the practical result of creating that room. “People learn better without fear,” Denuzzo says. Lowering the cost of admitting uncertainty gives employees more opportunity to ask how AI changes their work, test ideas and develop the understanding needed to use it with judgment. “That’s how literacy works,” he says.

That learning can happen while work continues. When founders asked Denuzzo to help develop an unnamed AI leadership experience, he chose to build on their thinking rather than deflate their enthusiasm through criticism. The example shows how questions and prior experience can shape work while it keeps advancing. Employees can likewise disclose uncertainty, challenge assumptions, test possible approaches and bring knowledge of existing work into implementation while leaders retain responsibility for decisions.

That participation matters because resistance has several possible meanings. A repeated question might expose a weak explanation; hesitation might reflect insufficient trust; attachment to an established process might reveal how people previously maintained visibility or exercised judgment. Leaders have to interpret each signal rather than assign one meaning to all resistance. Preserving the signal gives them a chance to identify the actual constraint before choosing a response.

Keep moving while preserving the evidence you need

Once resistance is treated as evidence, leaders can use it without making universal comfort a prerequisite for action. Denuzzo’s argument concerns sequence and productive engagement: implementation can continue while employees gain room to understand the change, question it and contribute. Apparent agreement provides weak evidence of learning, so continued movement still depends on preserving candid feedback. The aim is progress with enough information to judge what people are actually absorbing.

That requirement also places a limit on engagement. Denuzzo warns that frequent updates and progress checks can involve people so heavily that leaders interrupt the work they intend to support. Participation becomes counterproductive when employees spend too much of their time reporting on adjustment instead of working through it. More monitoring can reduce the learning leaders are trying to observe.

Giving people room has a different purpose from monitoring them closely. Employees need space to ask questions, admit what they do not understand, test ideas and influence how AI enters real work. Those activities give leaders evidence about organizational absorption while allowing implementation to continue. Friction can then inform pace and sequencing without controlling either one.

That boundary leaves leaders with a practical readiness test during the rollout. A functioning AI system shows technical capability, while employee signals show whether people can understand the change well enough to question it, incorporate it into their work and develop trust as their understanding grows. Leaders can keep an AI program moving while examining what resistance reveals about understanding, trust and changes to established practices. Preserving candor lets them distinguish genuine learning from employees who have learned how to look compliant.

Key takeaways for decision-makers

• Read AI readiness beyond deployment: Fast implementation and visible employee use can overstate organizational readiness. Executives need signals of understanding and trust alongside technical progress.
• Account for what new workflows remove: AI adoption can disrupt expertise, visibility, control and familiar processes. Rollout owners need to identify which of these employees are losing and provide workable replacements as practices change.
• Treat compliance as an incomplete signal: Employees may use AI while concealing confusion or concerns when questioning carries professional risk. Management teams need conditions where uncertainty and recurring questions surface early enough to diagnose adoption gaps.
• Use resistance as diagnostic evidence: Hesitation, repeated questions and attachment to existing processes can reveal different problems with trust, understanding or workflow design. Leaders can examine the source of friction while implementation continues.
• Preserve feedback while maintaining momentum: Frequent monitoring can consume the time employees need to adapt and learn. Executives can keep rollouts moving while giving employees room to test AI, question assumptions and expose unresolved issues.