Skip to content
Fast Company Executive Board

The Fast Company Executive Board is a private, fee-based network of influential leaders, experts, executives, and entrepreneurs who share their insights with our audience.

AI without culture change is just a failed proof of concept

Use these six people-centric practices to make it work.

[Photo: Getty Images]

Profile photo of Ismail Amla

Senior Vice Preside, Kyndryl Consult · Kyndryl

In every one of my client engagements for the past two years, AI and its impact on the nature of work have been critical topics. My experience is not singular. Virtually every piece of research on AI I have seen, from G2 to Gallup, tracks rapid growth in usage and adoption. Our research found that 95% of organizations report using AI across multiple functions.

Yet, a stark disconnect remains between the promise of AI and its actual results. Our research also found that while three-quarters of tech teams believe employees are embracing AI, 71% of C-suite executives admit their workforces aren’t ready to leverage it effectively. More troubling, 45% of CEOs even report active resistance or hostility toward AI initiatives. Until this hostility is gone and workers are curious and open to AI, the success rate will remain spotty. Ultimately, even though AI can give people superpowers, it’s people who drive significant changes and productivity gains. AI is only an accelerant.

Rand notes that more than 80% of AI projects fail. That’s double the rate of more traditional IT project failures. A recent S&P Global Market Intelligence study found that 42% of businesses scrapped some or most of their AI initiatives prior to production in 2025, up from 17% in 2024. Gartner forecasted last year that by the close of 2025, 30% or more of generative AI projects will be abandoned after proof of concept, in part due to unclear business value. Taken together, these statistics underscore a systemic failure to translate AI’s potential into operational reality.

According to our internal findings, only 14% of organizations have progressed beyond pilots to genuine AI-powered transformation. We call these organizations AI Pacesetters. What distinguishes these organizations isn’t fatter budgets, cutting-edge models, or better data. Rather, the AI Pacesetters exhibit these six important people-centric practices.

1. Rethink talent management for a blended workforce

Leadership teams and the HR business partners supporting them must reinvent their mindsets to adapt to this new reality of autonomous intelligence. Evolve talent management into a dual-capability function that onboards, defines, coaches, and retires both humans and AI agents, and treats digital workers as full members of the workforce. This requires working closely with frontline employees to define “jobs to be done,” guardrails, and where humans must remain in the loop.

2. Modernize roles to bake AI into daily work

Closely related to the HR and talent management shift, companies must reimagine job descriptions and team structures to better align with their evolving needs. Among our Pacesetters, 53% update role profiles to embed AI responsibilities or create hybrid human-AI teams. By making AI a native element of workflows and job roles, they prevent the “pilot-only” syndrome.

3. Embed change management from day one

Only 27% of non-Pacesetters report having a formal change management strategy for AI, compared to 88% of Pacesetters who have fully implemented one. Even world-class AI languishes without a roadmap aligning leadership, frontline teams, and HR around the ways AI will reshape roles and workflows.

4. Involve front-line employees in the project and technology design

Too often, AI initiatives are driven from the top down, but the best outcomes happen when frontline employees help shape them from the start. The 2025 Kyndryl Readiness Report shows that only 29% of organizations believe their workforce is truly ready to leverage AI effectively. Meanwhile, 87% of leaders expect that AI will “completely” reshape jobs within the next year—a near-universal shift. This combination of high disruption and low perceived readiness underscores a critical gap: Employees may not understand how AI will affect their work, and may feel unprepared or uncertain.

By involving frontline workers early in shaping AI deployments—designing workflows, weighing where human judgment is essential, and defining practical “jobs to be done”—organizations give themselves the best chance of bridging that gap. Early involvement builds trust, increases clarity, and helps ensure that AI tools solve real problems rather than creating misaligned processes or resistance.

5. Rigorously inventory skills

Understanding who can do what—both employees and AI agents—matters as much as what model is available. Among Pacesetters, 70% rigorously catalog existing employee skills and identify gaps crucial for scaling AI responsibly. They will need to do this for onboarded AI roles as well, to align with real needs.

6. Invest in personal enablement, not just e-learning

Rather than relying solely on the virtual, Pacesetters combine mentorship, real-world projects, and industry certifications. Forty-nine percent have dedicated AI-enablement teams, 39% provide access to external training, and 38% run coaching or personalized apprenticeship programs.

In the end, putting people first will pave the way for AI to do its job best. For the C-suite, the timeless lesson is that great tech withers without great people. To win the AI game, enterprises must invest as much in the human element as in data centers and GPUs. Otherwise, the proofs of concept and testbeds that look so promising in lab environments will never gain sufficient altitude to attain a tangible ROI. AI without cultural change is merely a pilot project that will likely never take off in the real world.

Ismail Amla is senior vice president of Kyndryl.