The Bottleneck Was Never the Technology
After 200+ workshops, I'm done being polite about it.
Welcome to the 89 new FullStack HR readers who joined last week!
The ambition is high with this newsletter - to be the guide for organisations in this AI transformation that is happening!
If you aren’t yet subscribed, join the other like-minded people in this free newsletter by subscribing below:
This week’s FullStack HR is brought to you by…HR Tech Europe.
We need to talk to each other. For real. The pace is faster than ever, none of us can crack this alone, and we need rooms full of smart people who challenge how we think.
Amsterdam, April 22-23. Google, LEGO, Coca-Cola, Microsoft, Maersk, Oatly, and a hundred other companies!
If you’re in HR, it’s free.
Register here
Happy Tuesday,
Quick update before we get to this week’s article.
Despite my brilliant move of putting the wrong link in last week’s description, the interest for the Cowork Bootcamp for HR has been completely insane. So I’ve decided to run it in two rounds. Simple as that. If you missed the first round (signups closed April 11), you can still join the second. I’m offering 30% off as a beta 2 discount. I’m fully convinced you’ll be happy you signed up.
Sign up here
(Yes, the link is the correct one)
Next week I’m at HR Tech in Amsterdam. And I want you to come build things with me!
Bring your laptop. We’re not doing slides and theory. We’re building real HR tools, live, together. This is the most practical session I’ve ever done at a conference.
Two sessions,
April 22 (12:30–13:30, Ask the Experts & Think Tank) Ask the Experts: Using AI to Build Practical HR Outputs
April 23 (15:00–15:45, AI Labs) AI Lab: From Prompts to Products: Building Real HR Tools with AI
If you’ve been reading this newsletter and thinking “I should really try this stuff,” this is your moment. Show up. Bring your laptop. We build!
Now. To this week’s article.
I’ve lost count of how many workshops and keynotes I’ve done this year. I don’t even want to count. But every session is a data point, and every discussion sharpens the picture. Things I believed six months ago, I don’t believe anymore. The tech moved forward, and I got better at seeing what actually works in organizations and what doesn’t.
The thing that is clearest of all? Leadership. Again.
If your CEO isn’t crystal clear on why AI matters, what you’re actually going to do, and how you’re going to do it, almost nothing will happen. You get this half-hearted energy where the organization acts as AI might possibly become important someday. And if leadership isn’t prepared to make real changes to how work gets done, nothing moves. I’m more convinced of this every week.
The capability isn’t the constraint anymore.
The models we have right now, whether it’s GPT 4.5 Pro, Opus 4.6, or the agentic capabilities in Claude Cowork, don’t matter. They are already more capable than what most organizations are extracting. Most companies aren’t using a fraction of what’s available.
Do they still hallucinate? Yes. Do they make mistakes? Absolutely. But if we applied the same forgiveness factor to AI agents that we apply to human employees, the gap would be indistinguishable.
The constraints are the organization and the people within it. The CEO who says “AI is important” but doesn’t mean it. Who hears “we bought Copilot” and accepts that as a strategy. But for the 1023rd time in this newsletter, this isn’t a technology purchase. This requires working through your entire organization, rethinking how work gets done, making real technology choices, and getting every level to understand what’s actually possible versus what’s just old mythology that won’t die.
The excuse factory is running at full capacity.
I meet so many organizations where old beliefs still live. “This won’t work.” “It’s not safe.” “The models can’t do that.” Sometimes it’s fear, sometimes ignorance, sometimes professional pride, sometimes politics. People who won’t make decisions and hide behind GDPR concerns or security worries that sound legitimate but are really just shields for inaction. And sometimes, honestly, people just aren’t capable enough to evaluate what they’re looking at. I know that sounds harsh, but it’s true.
So the CEO isn’t confident enough to push through it. They listen to the IT director protecting their territory. They listen to the HR director who doesn’t feel confident in the technology. And nothing moves. “Let’s just go with Copilot. It feels safe.” You shouldn’t make reckless decisions. But the cost of not making decisions gets higher every single day.
Courage. Year three.
I’ve been banging this drum for three years. You have to dare. There are ways to make it safe, but that requires investment. So either you invest in making it secure, or you accept that it’s maybe 90% there and move forward anyway. Do your GDPR assessments. Be compliant. But know what is what and get good advisors.
On that note. I always want to focus on the positive, on what you can do. But there is so much bad advice floating around this AI era, from both internal and external sources. I’m not exempt; I have blind spots, too. But the gap between good and bad guidance right now is enormous, and the wrong advice at the wrong time can cost you a year or more. Get people around you who actually know what they’re talking about.
The real risk.
It doesn’t matter if you’re an industrial company, a consulting firm, or a tech company. Someone, somewhere, is sitting with Claude Cowork right now, automating their support processes, operating at lower margins than you. And they will come for your market share. Not overnight but steadily.
So either you believe you’re fast enough to pick this up when it matters, or you start now, or you accept you’ll invest like crazy later to catch up. But believing you won’t need to do this work at all? That’s Nokia. That’s Ballmer-era Microsoft. That’s standing outside the internet and saying “we don’t believe in mobile phones.”
I’ve been conservative about this. I’ve tried to balance hype with reality. But enough is enough.
$20. That’s it.
Claude Cowork costs $20 a month, and you spend maybe 1-2 hours with it. That’s all it takes to see what this technology is capable of. The potential to fundamentally change how work gets done is real, and there’s a real risk that this leads to jobs disappearing. Which is exactly why, if you’re a leader or in HR, it feels wrong not to test this. If you can’t do it at work, do it privately.
I’m not saying this so you’ll buy my course. YouTube is overflowing with free examples. The point isn’t the course. The point is that you start. And the worst that can happen? You come back and tell me I’m a hype boy. Fine. At least you tested it, at least you formed an opinion.
I’ve given a lot of presentations this year. I haven’t met a single person who, after actually trying Claude Cowork, said “that looks useless.” Not one.
And, yes, this is coming to Copilot too. Eventually. So start now, and it’s downhill. Wait a year, and it’s uphill.
We need to lead this. Not stumble into the future because we were too scared to decide. We need to steer our organizations using
Maybe this was too much of a rant, I don’t know, but I’m getting a bit tired of just sitting around waiting.