Why Smart Workers Aren't Using AI: It's Not What You Think
Here's a puzzle: if AI can genuinely make knowledge workers ~30% more productive (and I'm living proof of this), why aren't more people rushing to use it for everyday tasks like writing resumes, prepping for interviews, or even optimizing their grocery runs?
The conventional wisdom says it's about awareness. People just don't know what AI can do. But after creating my first AI agent to write this very blog post (something I'd never attempted before because I didn't think I could), I've realized the real barrier isn't awareness. It's a broken sales pipeline that starts much earlier than we think.
The Real Pipeline: Six Hidden Barriers
Access: "I Could Use This If I Wanted To"
Access isn't just about having a smartphone or internet connection—though digital equity remains a significant barrier for many Americans. For knowledge workers, the bigger access issue is psychological. I had the technical ability to create AI agents for months, but I didn't think I could until I decided I must. There's a paralyzing fear of the unknown mixed with excitement that keeps people stuck at the starting line.
For workers with disabilities, access becomes even more complex. Accessibility in AI tools remains inconsistent, creating another layer of exclusion we rarely discuss.
Awareness: "I Know What This Can Do For Me"
Many people already know about AI, but there's a massive group of "shadow users"—workers who use AI to excel at their jobs but won't admit it because it feels like cheating. They're quietly getting ahead while others debate whether AI is relevant to their work.
This creates a weird dynamic where the most successful AI adopters are invisible, making adoption feel riskier than it actually is.
Interest: "I Want to Learn More"
The elephant in the room is job displacement anxiety. I had to laugh-cry when my UX designer friend told me she was interviewing for a role where she'd train AI to replace other UX designers such as herself. While AI will transform jobs, the workers who adapt fastest often end up in stronger positions.
But how do you generate interest when the narrative is "learn this or become obsolete"? Fear isn't sustainable motivation.
Intention: "This Might Work For Me"
This is where quick wins become crucial. At a recent hackathon, our team of nonprofit workers, government employees, and consultants dreamed up an "AI bus" with a friendly guru inside—like a modern-day wise woman who could answer everything from "How do I change my toddler's sleep routine?" to "Help me nail this job interview."
The idea was whimsical, but it also addressed a real need for low-stakes experimentation. People need to see AI make their lives tangibly better before they'll commit to learning it.
Evaluation: "How Would I Actually Use This?"
Most training programs fail because they focus only on delegation—what tasks to hand off to AI. But Anthropic's AI fluency framework includes four critical dimensions:
Delegation: Choosing the right tasks for AI
Description: Communicating clearly with AI systems
Discernment: Critically evaluating AI outputs
Diligence: Using AI responsibly and ethically
People need safe spaces to explore these skills without exposing their ignorance. Community learning environments and demo sessions work better than individual trial-and-error.
Purchase: "I'm Ready to Invest"
Cost barriers are real but nuanced. While basic AI tools are often free, premium features can cost $20-100+ monthly. For many workers, especially those in nonprofits or small businesses, this creates a meaningful barrier.
Nonprofits have an underexplored role here—not just in providing access to their clients, but in introducing tools that individuals might never discover or afford independently.
What This Means for You
At the end of the day, AI is a tool, not a solution. But I don't think it's possible to say "I'm not an AI person" anymore—it's too game-changing for that stance to be sustainable. The real question is how quickly you'll move through this pipeline.
If you're an individual worker: identify which stage you're actually stuck at. Are you held back by access (confidence to try), awareness (understanding applications), or evaluation (knowing how to assess results)?
If you're leading a team or organization: awareness might not be the problem. We need to design interventions that address the specific pipeline stage where people are actually stuck. Create safe spaces for experimentation, celebrate transparent AI use, and provide pathways for skill development that don't feel like survival training.
