When the CEO Becomes the Lab Rat
Jim VandeHei, co-founder and CEO of Axios, published something unusually candid last week. He spent a year using AI obsessively — Claude, ChatGPT, agents, the works — and came back with a verdict most people aren't saying out loud: AI is way better than most think, but it's colliding with hard human and business realities, making it confusing, clunky and chaotic for most employees. His second takeaway hit particularly hard: it takes work. You can't wing it. You need to work at it daily. You have to feed it information, tell it what works and what doesn't. This feedback loop is where the magic happens. He's right. And what he's describing is exactly the problem we're solving right now with a client.
What We're Actually Building
We're in active conversations with a client to design an AI transformation program for their existing non-technical employees — people who are going to be using Claude CoWork as part of how they do their jobs going forward. Not developers. Not engineers. Knowledge workers. The kind of people VandeHei is talking about when he says most of his employees find AI confusing, clunky, and chaotic in its current form. Here's what's different about how we're approaching it — and why it matters.
Hands-On Workshops, Not Classroom Training
The old model of corporate training — sit in a room, watch a presenter, take a quiz, get a certificate — doesn't work for AI. Full stop. VandeHei said it himself: you need to go beyond chat. The life-changing work happens in projects and with agents. You learn by doing, not by watching. This program is built entirely around hands-on workshops. Every session puts participants in direct contact with Claude CoWork on real tasks that mirror their actual work. They're not learning about AI. They're learning to work differently, with AI as a core part of how they operate. The difference sounds subtle. The outcomes are not.
Office Hours With Practitioners
Here's the piece most training programs completely miss. What happens between the sessions? What happens when someone gets back to their desk on Tuesday morning and hits a wall? In this program, learners get access to bILTup practitioners outside of class through custom office hours and collaboration sessions. Not a helpdesk. Not a chatbot. Actual practitioners — people who use Claude CoWork professionally, who deliver real work with these tools every day — available to work through the hard problems with your team when they come up. This is the feedback loop VandeHei is talking about. The magic doesn't happen in the training session. It happens when someone is trying to do real work and has a practitioner in their corner helping them figure out why it's not working the way they expected.
Four Hours a Week. That's It.
One of the things that kills AI training programs is the time commitment. Two-day bootcamps. Full-week intensives. Curricula that pull people out of their actual jobs for so long that they come back to a pile of catch-up work and immediately deprioritize everything they just learned. This client had a clear requirement: four hours per week maximum in instructor-led sessions. That's it. The program is designed to work within that constraint — because the constraint is right. Real learning happens when it's integrated into the work week, not when it's bolted on top of it.
You Have to Be Nimble
VandeHei's biggest lesson after a year of experimenting: be more curious than ever and more humble than ever. No one has a clue how AI will unfold inside organizations in the coming months. That's exactly the operating philosophy at bILTup. We don't have a packaged curriculum we drop into every engagement. We have practitioners who understand what's changing in real time — because they're working with these tools every day — and a methodology built around meeting learners where they are to get them to where they need to be. The client we're working with has employees at completely different starting points. Some have been experimenting with AI for months. Others have barely opened Claude.ai. A program that treats them the same way will fail half of them. You have to be nimble enough to meet people where they are — and build the path from there.
The Bottom Line
VandeHei ended his piece with something worth repeating: a new class of superworker is being born. Rank-and-file workers whose brains are wired for AI, who become AI enablers for their entire team. He said it's been easier than expected to find and train them — but only with the right approach. The right approach isn't a two-day bootcamp. It's not a vendor webinar. It's hands-on practice with real tools on real tasks, with practitioners available when the work gets hard, delivered in a cadence that people can actually sustain. That's what we're building. If your organization is trying to figure out how to do this well, that's worth a conversation.
Talk to us about your AI transformation program → View the Public Schedule → See the full Claude CoWork program →