Key takeaways:
AI's transformative potential will only be realized when organizations build trust, align around clear priorities, and redesign work around people.
- Without gaining employee's trust in both the organization and the technology, AI adoption creates anxiety and resistance rather than innovation.
- Use AI to rethink and redesign how work is done, rather than accelerating existing processes.
- Organizations need to slow down and focus on upskilling employees before they can use AI to move faster.
Before the invention of the elevator, not many buildings were more than five stories high.
It was just too much work to walk up all those stairs. The elevator gave us much more than an easy way to move between floors. The elevator prompted new ways to think about architecture, about engineering – even about cities.
In some ways, artificial intelligence is an elevator for knowledge work. But instead of using that elevator to rethink and redesign work – to build an entirely new architecture – we’re fixated on using it to move from the ground floor to the fifth floor really, really fast.
While the term ‘artificial intelligence’ has been with us since at least 1956, it’s only with the public launch of ChatGPT that organizations and individuals have seen its potential for knowledge work. If AI can help knowledge workers eliminate significant amounts of busy work, workers could repurpose that time to do work that is more innovative, meaningful, and conducive to human flourishing.
That’s the dream. But mostly, it’s not happening. Even when AI does remove repetitive work from highly skilled employees, those employees aren’t using their newfound time to innovate. Instead, they’re devoting more time to routine administrative tasks, such as email.1 According to the Kyndryl People Readiness Report 2026, only 49% of leaders say their AI implementations are resulting in either reduced costs or growth.2
For employees, there’s psychological safety in an overscheduled calendar and an overflowing inbox. There is no psychological safety in doing something new – especially if it’s something the organization has struggled with before.
For AI to bring value to organizations and to help us rearchitect work, employees need the knowledge, skills, and psychological cover to tackle the more strategic work their organizations need. That requires two distinct forms of trust: trust in the organization, and trust in the technology itself. These are different problems with different solutions, and few organizations are addressing them effectively.
The decline of trust
No generation enters the workforce friction-free. But GenZ is experiencing something new: a belief that the social contract between business, its workers, and society has been ruptured. Globally, 68% of workers believe that business leaders are purposely misguiding people,3 and 58% of Gen Z say they hold a moderate to strong level of grievance against business and financial institutions.4
Most leaders are not actively addressing these concerns. The complexity of the world they’re trying to lead in is unprecedented; they’re exhausted and burned out. A generation ago, a C-suite leader might have dealt with three to five economic shocks in a career. More recently, C-suite executives have had to navigate through seven such changes in the last 18 months – from layoffs to tariffs, inflation and war.
It’s in this environment that skeptical employees and overwhelmed leadership are being asked to redesign their work and their businesses to incorporate a new technology – one in which most have no specialized training, and one that is as unpredictable as it is transformative. Only 23% of leaders say their workforce is ready to successfully leverage AI. That’s actually lower – by six percentage points – than the number who thought the workforce was ready a year ago. Which may be why leaders don’t see things improving: 79% say that the speed of AI adoption will outpace their organization’s ability to adapt its workforce, governance, and operating model.5
Before addressing how organizations can best close these trust gaps, we should ask if some of that distrust is warranted. Recent research suggests that it is: A study from Stanford found that 41% of organizations are pursuing AI use cases that are either unwanted by workers or technically infeasible.6 In these cases, a trust gap is not a sign of irrational resistance. It’s a rational response to poorly designed use cases and implementation, and it’s an important signal to leadership. It’s difficult for leaders to build trust if they themselves are skeptical – and only 39% of leaders say they completely trust their own organization’s transparency about how AI is being used.7
Trust throughout the enterprise
Overall, C-suite leaders are still bullish on AI, seeing it as a powerful lever for cost-efficiency and productivity in a low-growth environment. Some 50% of leaders say executives are enthusiastically embracing AI.8
That enthusiasm wanes deeper in the organization, and leadership knows it. Leaders say only 36% of managers are enthusiastically embracing AI, and only 31% of individual contributors and front-line workers.9 Managers see AI being jammed into workflows without the proper groundwork, adding complexity but not improving results. Their employees, who were once productive and confident in their jobs, are now worried about being laid off. As are the managers themselves.
Employees often see AI being used to do work that they would prefer to do. Many software developers, for example, have deep technical expertise and enjoy writing code. Now, they’re being asked to be quality control for an AI tool. That’s the opposite of what they want: They want to do the work, and have AI be the quality check.
Even if trust levels in an organization are high, the AI itself is often not perceived as trustworthy. Many organizations don’t trust the companies that are developing large language models and AI tools. They’re apprehensive about how their data will be used, and they’re not confident in their ability to control the data leaving their organizations, especially given the proliferation of shadow AI. They’re also concerned about the fine-tuning of AI models, and the potential of synthetic data to degrade reliability and introduce drift.
Beyond vendor and data concerns, organizations need to evaluate whether AI systems perform reliably in their specific operational context. Trust in a vendor's reputation is not the same as trust in a system's output. The latter requires ongoing monitoring and measurement.
Managers are where AI strategy becomes real. If managers are unclear on the purposes, principles, or expected outcomes of AI use, employees will experience AI as yet another layer of ambiguity rather than a path to better work.
Going slow to go fast
In this environment the nature of leadership must change, rejecting blueprints that rely on command and control. Leaders need to navigate the complexity of compounding economic shocks, rebuild trust by keeping people at the center of their values and actions, and embrace new ways of working. All of this is required before an organization can even begin to think about integrating AI into their work.
Leadership teams can start by doing two things: setting priorities and cementing internal alignment. In an environment defined by cascading uncertainty, priority alignment is a core leadership requirement. Employees are not only listening to what leaders say; they are watching to see if leaders create a coherent operating environment.
Unfortunately, few leaders are effectively prioritizing or aligning. We know the human brain can manage three to five priorities; most organizational leaders have normalized a chaotic way of operating and are trying to manage 10. Over the long-term, stress leads to terrible decision-making. Most organizations would be well-served to slow down, set a small number of priorities, and stabilize their core business.
With too many competing priorities, it’s easy for organizations to slip into reaction mode and lose sight of the core questions: what are we working on, and why does it matter? Clear priorities reconnect people to a shared purpose beyond themselves. Once that alignment is in place, speed becomes possible. Teams can share resources more easily, reduce friction, make faster decisions, and reinforce the shared goals that give work meaning. At one technology organization, for example, the leadership team cascaded down 15+ strategic priorities for the year. Each priority made sense on its own, but together they created chaos. Functional leaders struggled to allocate resources, and cross-functional tension increased because different teams were optimizing for different priorities. Even investors were questioning the future direction of the company, unable to make sense of the organization’s true big bets.
Once the leadership team narrowed the list to three clear priorities, everything moved more smoothly. Teams understood what mattered most, tradeoffs became easier, and collaboration improved. The lesson this company learned is that alignment is not about giving people more priorities. It is about making the few that matter unmistakably clear.
When leadership teams clearly declare priorities and become internally aligned, employees experience less ambiguity, fewer mixed signals, and greater confidence that decisions are being made consistently. That predictability gives people a stable basis for action even in the midst of disruption. In that sense, clarity is kindness: it reduces the cognitive and emotional burden for employees, and helps them focus their energy on the work that matters most.
Alignment can't be relegated only to the C-suite. Managers are where AI strategy becomes real. If managers are unclear on the purposes, principles, or expected outcomes of AI use, employees will experience AI as yet another layer of ambiguity rather than a path to better work. Organizations that successfully build trust create structured feedback loops where employees can report what's working, flag what isn't, and see that their input changes decisions. Trust is built when influence flows in both directions. An example comes from a pharma company rolling out new AI-enabled ways of working. Early on, leaders realized that adoption data alone would not tell them whether the change was actually working. So, they moved beyond a one-time listening exercise and built in regular pulse surveys throughout the rollout. The goal was not just to ask, “Are people using the tool?” It was to understand where employees were getting stuck, what was creating confusion, and what support teams needed next. That feedback gave leaders a real-time view into the employee experience of change. Instead of waiting until frustration showed up as resistance, they could adjust communication, training, and manager support while the change was still unfolding.
The how and the why: Accelerating work with AI
Once leaders have committed to priorities and internal alignment, two key initiatives will allow them to move quickly with AI: upskilling and a strong statement of principles.
Organization-wide upskilling is critical, but only 31% of leaders say they have implemented a proactive AI-upskilling strategy.10 In many organizations, a wide swath of workers have access to Copilot or the enterprise version of ChatGPT, but they’ve had no training to go along with it. Yet leaders are still expecting to see cost savings.
That training should be preceded by a skills ontology – a clear picture of AI skills, and of the skills the organization needs to hire for. It also requires a definition of AI fluency as it applies to the organization, and an accurate picture of current AI use. But only 34% of leaders say their organizations even have an inventory of employee skills.11 Without this basic information, it’s difficult to design effective use cases.
Organizations also need a set of strong principles around the use of AI. Those go well beyond a policy that outlines what employees can and can’t do. Instead, principles lay out the organization’s philosophy around AI, its expectations for employees, and clear examples of what good looks like. For example, a Fortune 500 retailer rolling out AI in stores quickly encountered employee anxiety about job loss and changing expectations. Rather than letting that uncertainty spread, leaders defined a clear set of principles:
-
AI would be used to augment, not replace, retail associates.
-
Associates would be expected to spend more time creating value for customers.
-
AI would be used to increase personalization and improve customer experience.
That clarity changed the rollout. Employees understood what AI was meant to do, how their roles would evolve, and where their human contribution still mattered. As a result, the retailer was able to use AI as a source of competitive advantage rather than a source of fear and confusion.
AI principles dictate how the organization will make decisions in moments of ambiguity. Without them, AI adoption can feel reactive, uneven, or extractive, which erodes trust. Clear principles signal what the organization values, where human judgment must remain central, and how risks around fairness, privacy, accountability, and workforce impact will be managed. They show how tradeoffs will be made and what responsible use looks like in practice. They also create consistency across teams, so AI is used not just because it is available, but because it advances shared priorities in a way employees can understand and trust.
Effective principles are testable commitments, not documents. Organizations should be able to point to specific decisions that were made differently because of their AI principles, and specific outcomes that resulted from those decisions.
The organizations that move fastest with AI will not be the ones that normalize chaos. They will be the organizations disciplined enough to slow down where it matters: to set priorities, align leadership, define principles, build skills, and redesign work with people at the center. That is how trust becomes an accelerant. And without trust, AI remains a tool layered onto old work rather than the force that helps us re-architect it.
- Workforce Lab, Slack, 2024
- People Readiness Report, Kyndryl, 2026.
- Trust Barometer, Edelman, 2025.
- “Gen Z and grievance: A generation’s response to a world under threat,” Edelman Gen Z Lab, 2025.
- People Readiness Report, Kyndryl, 2026.
- “Future of work with AI agents: Auditing automation and augmentation potential across the U.S. Workforce,” Stanford Institute for Human-Centered AI and Digital Economy Lab, 2025.
- People Readiness Report, Kyndryl, 2026.
- People Readiness Report, Kyndryl, 2026.
- People Readiness Report, Kyndryl, 2026.
- People Readiness Report, Kyndryl, 2026.
- People Readiness Report, Kyndryl, 2026.
Have questions or want to learn more?
