Generative AI is everywhere in 2025, but success is not. A new report from MIT's Project NANDA report reveals a stark truth: 95% of enterprise AI projects stall before showing results. The problem is not the technology, it’s the way businesses adopt, integrate, and govern it. For leaders, this is both a wake-up call and a roadmap.
The Reality Check: Why Most AI Pilots Fail
Despite the hype, MIT found that only about 5% of AI pilots achieve rapid revenue acceleration. For the rest, the results are flat. As Aditya Challapally, the report’s lead author, explains:
“It’s not the quality of the AI models, but the learning gap for both tools and organizations.”
Executives often blame regulation or technology, but the data points to a deeper issue: generic tools like ChatGPT don’t adapt to enterprise workflows unless given the right context, governance, and integration. This is exactly where the Kendall Project mission comes in. By raising AI literacy across the enterprise, leaders and teams learn how to frame problems clearly and provide AI with the right context. That literacy closes the “learning gap” MIT describes and makes implementation more efficient and ultimately more successful.
Where AI Really Pays Off: MIT Finds Hidden Value in Back-Office Automation
Here’s one of the most surprising findings: over half of enterprise AI budgets are going to sales and marketing. Leaders are betting on tools that promise faster lead generation, automated copywriting, and more personalized customer outreach. But MIT’s research shows the biggest measurable ROI is actually in back-office automation, where AI reduces reliance on business process outsourcing, slashes agency costs, and streamlines repetitive workflows.
This disconnect highlights a common leadership trap: investing in the areas that feel most visible rather than the ones that deliver the strongest financial outcomes. The Kendall Project helps companies shift that focus by spreading AI literacy inside the organization, showing leaders and teams how to give AI the structured context it needs to solve real operational bottlenecks. When AI literacy is strong, the investment follows impact, not hype.
The Human Side: Adoption, Not Just Tools
MIT points out that success depends on empowering line managers and team members, not just central AI labs. Change has to be distributed across the business, not confined to an innovation silo.
Workforce disruption is already underway. Companies are not conducting mass layoffs, but they are not backfilling administrative and outsourced roles as they become redundant. This quiet shift underscores why AI literacy matters: when people across the organization understand how AI fits into workflows, the transition is smoother, less disruptive, and more aligned with long-term value.
The Kendall Project believes that AI is a team sport. Success is not driven by a small group of technical experts, but by equipping employees at every level with the knowledge and confidence to use AI in their daily work. By spreading AI literacy, organizations can turn adoption into a collective effort, where managers, teams, and leaders all contribute to shaping how AI is integrated, governed, and scaled.