I had four conversations this week with people in different industries and different roles. Every one of them expressed the same problem: AI access doesn’t equal AI adoption.
One was a new hire, bored out of her mind in a corporate AI training. The rollout started everyone at the beginning, even though the room held everything from first-timers to people who had been using these tools for years.
A day earlier, someone leading enablement at a tech services firm described the same room from the other side. Some of her people have never opened a chatbot. Others are running local models on their laptops. One curriculum fails both ends at once.
This matches what I saw running tech education at Airbnb, and it matches the numbers. Stack Overflow’s 2026 developer survey found that only 24 percent of developers say their organization provides approved AI tools and guidance on how to use them. Thirty percent say AI use is optional and left up to the individual. Everyone else is figuring it out alone.
Section’s proficiency report this summer found that workers who got agents plus training scored 47.5 out of 100. Workers who got agents and nothing scored 33.1.
Training moves the number. But the training I keep hearing about is built for beginners, walking through basic tool features in a session that has been taught in the same form a hundred times over.
Here are three tactics I’ve seen work across industries, company sizes, and levels of adoption:
Use AI to teach AI. You have learners at every level. You know what is excellent at meeting individuals where they are? AI.
Invest more in the people who are already moving. This goes against every L&D instinct. Don’t leave anyone behind! I agree. But if you want ROI this quarter, upskill the people moving fastest and let them pull the rest.
Make it job-specific. People have their aha moment when they see AI applied to their own workflow and their own problem. If you want beginners using it, make it worth their time.
I’ll be writing about each of these over the next three issues: how they’ve worked in different industries, and how to make them work for your organization.
This week’s People Who Know Things episode
This week I was joined by Cora Anthony, an AI enablement and L&D leader. Her workshops open with one question: what would you do with half your time? We talked about ROI for organizations and for the person whose skills go with them wherever they work next.
What I tried this week
I built a research coding sheet for the podcast. I included twelve things I want to learn from every guest, like how AI adoption started at their company, who drove it, what training existed, and what failed. Instead of asking guests a questionnaire, the answers get pulled from the transcript after the episode and logged in a Google Sheet.
My hope is that the podcast builds a research dataset over time. I’m also building an agent on top of my own transcripts for a talk at the end of the month, so the rubric doubles as its spec. Stay tuned to see how that build goes.
One question for you
When AI training showed up at your company, how did it start? Hit reply and let me know. I’d love to hear from you!
Molly