AI shame is real, and it’s quietly holding your organization back
Everyone says they're past the basics. Almost no one is.
Welcome to FullStack HR, and an extra welcome to the 32 people who have signed up since last week.
If you haven’t yet subscribed, join the 11400+ smart, curious HR folks by subscribing here:
Happy Wednesday!
I know, I know. Claude Fable 5 dropped yesterday, so you probably expected me to write about that.
And I will. Tonight I’m recording a podcast episode about it, once I’ve had proper time to test it. Because my first reaction was a simple “oh my god.” I get the hype, at least to some extent. Especially for non-coding work.
But I’m getting ahead of myself. That’s what the podcast is for. Drop a follow and I’ll give you the full breakdown of the latest release from Anthropic there.
Today, I want to talk about something else entirely.
Shame.
Let’s get to it.¨
Today I want to talk about AI shame.
What do I mean by that? When I’m out in organizations, I keep meeting people who carry a bit of AI shame, and they try to cover it up. Let me explain.
Over this past winter and spring I’ve done roughly 110 workshops. I’m not saying that to brag. (I’m not even sure it’s a good thing.) If you do the math on how many workshops that is per day, it’s a lot. It’s a bit too much. I’ve been spread too thin, and I’m spending the autumn figuring out how to do less. I’ll personally do fewer workshops going forward, and we’ll hire people so we can still help just as many organizations.
But here’s the huge upside of all that volume. Data. Every organization I meet gives me data points, and I think that’s marvelous. I’ve written about this in the newsletter before. I see it as a real privilege to sit down with people and ask, “Hey, what’s happening in your world of AI?” I’m good at capturing meeting notes and transcribing conversations, so I’ve built up quite a ledger on what AI adoption looks like right now. And it’s from almost everywhere; there’s barely a part of the world I haven’t worked with, so I feel reasonably well informed on a global level.
Which brings me back to AI shame, because it comes straight out of that data. This isn’t about singling anyone out. It’s a summary of a lot of conversations with a lot of organizations.
The pattern I keep seeing
When I walk into a workshop, I always ask the same thing. “How would you rate your own AI proficiency? Where do you think you’re at?” It’s a hard question. People interpret it differently, and you can argue endlessly about what AI knowledge even is.
But here’s what happens more and more, and it’s happened far more often in 2026 than in 2025. The buyer tells me up front, “We’re quite AI proficient here. You can skip the basics, go straight to processes, team benefits, organizational impact.”
The first few times I heard that, I was thrilled. Finally I get to talk about the stuff I’m most passionate about!
Then I started pushing a little. “Okay, show me. How are you using AI today? What’s your proficiency really at? Tell me more about what you’re doing.” And you realize people are using AI, sure, but they’re using a sliver of what’s possible.
Why it happens, and why I get it
I understand why we ended up here, just look back at the support people have been given. Very little education, few real processes and policies that scare people rather than support them. Across this whole ChatGPT era, the help has been minimal. So how would anyone be proficient when they’ve had almost nothing to lean on?
The problem shows up the moment you try to move forward. You start talking about building agentic processes and adopting an agentic mindset, and you get blank stares back. The conversation gets harder because you’re trying to skip a step. You’re trying to leap past the grunt work.
And there’s no way around that grunt work. Using ChatGPT, Copilot, or Claude for your own work takes effort. You have to understand what the model can do and how it can help you. Skip that, and every conversation after it gets harder.
This is more common at the top
I’m a little hesitant to say this, but it’s slightly more prevalent in leadership teams. Conceptually, they know they should have done the work. They just haven’t, usually because there’s no time, and sometimes because they quietly feel a bit too senior to get back down in the weeds and play around.
It’s a combination, and I’m not generalizing about everyone. But it matters, because in a leadership setting it’s hard to lead an organization through this if you don’t know the true nature of the models yourself. You can’t suddenly set sensible expectations about AI in the workplace if you don’t understand what these tools are.
So this is really one observation. Don’t skip the step.
I know you’d love to be further ahead. That ambition is great but you still have to do the work to get there.
What to do then?
I’ve been saying the same thing for three years now. Educate your people! Give them help and support. Make time for them to change how they work, to play with the tools, to genuinely evaluate what the models can do. Otherwise AI adoption becomes a token gesture, something nice you say you’re doing but not really doing.
This is not rocket science. We need to support our people, and right now, that is getting a basic understanding of what the tools can do.
I’ve run programs for truck drivers who don’t even have a computer in front of them, for obvious reasons (usually a bad thing when you drive a truck). We still trained them on what AI can do, because they play a role in the organization and need to understand what’s happening across it. They also use AI in their spare time. They ask it for advice on salary, on how to handle a manager or a coworker. They’re being informed by AI off the clock, so their employer decided it was important they understand it. Include everyone.
The cure for AI shame is education, and it starts at the top.
One caveat about my data
All this said, my data set is skewed, and I know it. When people bring me in, they want someone practical, because that’s still my biggest selling point. I love being practical. That doesn’t mean I can’t be strategic; I’ve done plenty of that too, but I lean practical because so few people do. Most consultants in the AI space want to focus only on strategy.
So the people I meet are the ones who struggle with the practical side, or who’ve at least admitted they want to get practical. It’s easy to get pulled into “we need to think strategically about this” or “we need to stay high level.” I’d rather get practical and run practical sessions, because right now that’s exactly where you need to be.