In mid-June, Amazon CEO Andy Jassy sent a memo to all employees that caused a bit of a stir: It openly stated that AI adoption would shrink the size of Amazon’s workforce. Jassy is far from the first person to say that AI is coming for jobs. The CEO of Klarna was gloating about replacing workers with AI in 2024, before backtracking and supplementing AI with gig work. Elon Musk joined a 2023 letter calling for a pause in AI model development (before rushing to build Grok) due to fears about letting AI “automate away all the jobs.”
Perhaps Jassy’s letter caused a reaction because it’s far more grounded than past messages: AI is here, according to Jassy, and it’s already automating work. For executives rushing to implement AI, this is a problem. Employee rejection is one of the key risk factors for successful AI implementation, and fears of job loss are a major driver of rejection. For executives, our consumer insights research shows that employee attitudes about AI are already mature; the time for messaging has largely passed. Leaders must now focus on strategic rollout tactics that address these fears head-on.
Understanding Employee Fears Around AI
Employees hesitate to adopt AI because it makes work feel less secure, less fulfilling, and less human. Their concerns are not high minded or moralistic but are practical considerations about their future and opportunities. Employees fear being unable to adapt quickly enough to keep their jobs, or if they do keep their jobs, their day-to-day work will become menial and unrewarding: merely being an “AI babysitter” while they AI monitors them. In addition, employees worry about their career path. How can they follow in the footsteps of senior employees and advance in the company when they aren’t gaining the same skills or completing the same functions as previous employees? Worse, employees feel that they are training their eventual AI replacement. Even employees who embrace AI want to show they’re thoughtful, forward-thinking, and use AI to enhance their work — demonstrating that AI reflects and amplifies their own skills.
Our maturity curve shows that these beliefs are already in the realm of established ideas. Internal messaging about how a company is different or corporate pledges not to replace jobs with AI will do little to assuage these fears. This helps explain AI hesitancy: while surveys show many employees are using AI already, they are highly concentrated in areas like computer science and management, not sales or administration. Executives must avoid aggravating these fears by carefully selecting AI applications or else risk employee backlash and subversion. The use of AI as a technology of seeing stands out as the most risky and backlash-prone area. It directly triggers the fear of increased surveillance, and for many tools (like meeting summarization), it’s difficult to make the case that this is saving employees time, as it’s not automating an existing task but adding a new function. Plus, there’s no way for an employee to feel like a savvy user of AI, as they are largely the subject of the seeing, not the operator of the system.
Key challenges to AI adoption
1. Avoid automating away joy and pride
Employees don’t just fear job loss, they worry about losing the parts of their jobs they love. Some tasks are deeply tied to identity, while others are not. For example, customer service and sales professionals take pride in building strong client relationships. Automating these touchpoints with large language models risks alienating experienced sales leaders. In contrast, few engineers see sending emails as a core competency, even if it consumes much of their time. They’re more likely to embrace automation for these routine tasks.
Leaders must do the hard but necessary work: identify tasks that drive employee pride and delay or avoid automating them. While these tasks may appear ripe for efficiency gains, those gains are a mirage if employees resist adoption. Instead, focus on automating tedious or frustrating tasks first. This approach builds trust in leadership and helps employees gain confidence using AI tools.
2. Build new pathways for promotion
AI transforms knowledge work by breaking down complex, long-term tasks like coding into simple, iterative interactions with AI systems. This shift can unsettle junior employees, who often earn promotions by mastering skills demonstrated by senior colleagues. If proving one’s expertise is essential for advancement, AI can feel like a threat, as it limits opportunities to develop these skills and may render them less relevant
AI also reshapes daily work in ways that existing KPIs and performance metrics may not capture. Employees who use AI effectively will spend more time strategizing and planning and less time on the tasks themselves. Will an employee’s careful planning and foresight be captured when it comes to a traditional performance review? Executives must understand how AI changes career trajectories, especially for junior and mid-level staff. They should proactively adapt promotion practices to recognize and reward employees who demonstrate thoughtful, effective use of AI in their roles.
3. Retain AI champions
Successful AI adoption depends on internal champions: employees who spot opportunities for AI and inspire peers by showing its benefits in action. Identifying and retaining these champions is critical and goes beyond broader career pathway concerns. Our consumer insights research highlights a dilemma: The employees most eager to embrace AI and lead its rollout are also the most likely to leave for better opportunities. AI skills are in high demand, younger employees tend to switch jobs more often, and those drawn to the challenges of AI often crave new experiences elsewhere.
Executives must act proactively to identify these champions early, give them freedom to experiment with AI and the room to fail, and show them that they have more opportunities for growth within the company than outside it. Leaders should also cultivate unexpected champions: experienced, dedicated employees who may resist AI due to the fears described above. Encourage these trusted team members to lead with AI. Their credibility will help bring more skeptical colleagues along during the transition.
Tactics for successful AI implementation
Given these challenges, what does successful AI implementation look like? While the circumstances and approaches of each company will differ, there are common threads we can look to.
1. Treat AI implementation as a process change
AI is not fundamentally different from other new software tools: It changes how employees work. The difference is its impact; AI transforms day-to-day tasks more dramatically than most other software. Successful adoption requires new workflows, processes, and support. Moving beyond pilots demands real investment in human resources to ensure AI delivers value. Some teams already excel at managing these changes. For example, manufacturing groups and strong sales leaders often handle process shifts well. Others, such as heads of R&D or innovation, may lack these operational skills. Executives should plan for a shift from innovation leaders to implementation leaders to manage this transition effectively. AI will also create new bottlenecks, especially in decision-making. As AI automates routine work, junior employees will reach decision points faster. Without clear processes for delegation and oversight, managers risk becoming overwhelmed by constant requests for approvals. Leaders must build structures that empower employees to make decisions while maintaining appropriate oversight.
2. Leverage existing KPIs to measure AI success
One of the most common questions we are asked about AI is how to quantify its impact — or the lack of it — across the organization. Many AI functions are inherently hard to measure, much like other core software tools. For example, what’s the ROI on having email or Excel? While the value of these tools is obvious, their precise impact is difficult to quantify, yet no one would suggest restricting their use. Although it’s possible to measure the direct impact of AI automation, its broader role as a tool for insight and knowledge management will likely show up indirectly. Rather than building an entirely new measurement architecture, executives should rely on established, trusted KPIs to gauge AI’s contribution. This approach ensures consistency and keeps the focus on overall business performance.
3. Meet employees where they are
Employees can tell when leaders aren’t being candid. Acknowledge their concerns openly instead of dismissing them as unfounded. Avoid or at least postpone AI applications that risk provoking employee backlash. If certain uses remain in the roadmap, develop clear plans to ensure employees can still find satisfaction and growth in their roles. This approach may leave some potential efficiency gains unrealized, but it’s a worthwhile tradeoff for a smoother, more sustainable AI rollout. Building trust and addressing real fears will drive long-term success, and don’t try to convince them that those fears are unfounded.
Lux Take
Executives will need to meet employees where they are and build trust by changing what tools they roll out, how they reward AI use, and what structural changes they make. This means avoiding some AI applications most likely to provoke backlash, while building trust by communicating transparently, changing processes around decision-making and promotion, and leaning on familiar key performance indicators (KPIs) to drive success. Leaders should refrain from introducing AI-specific KPIs or demanding sharp productivity increases that alienate long-standing high performers with an excessive focus on AI.
For more strategies on successfully implementing AI, connect with us today.