In partnership with
You cannot build what you haven't lived.
Most organizations approaching AI transformation are doing it from a distance. Tool licenses get purchased. Training programs get launched. And people do adopt — but individually. Each person learns to use AI to speed up their own work. That is valuable. But it stops short of the deeper change.
The real gap is not adoption. It is coordination. The decisions shaping how teams work with AI together are being made by people who haven't actually built anything with AI themselves. Strategy from the outside of the work, not from inside it.
One company is doing something different — and the results are worth looking at closely.
In today's newsletter:
What Atlassian's AI Builders Week was, and why it produced results
The three principles that made it work — including a contrast with the hackathon era
A story from 2010 — this idea is older than AI
Where Most Companies Get Stuck With AI
According to BCG, 74% of companies struggle to achieve and scale value from AI — even when adoption is happening. (Source: BCG, October 2024)
Most companies approach AI adoption the same way.
They buy licenses. They run lunch-and-learns. They send a Slack message with a list of approved tools. Maybe they bring in an external speaker for an afternoon.
And then they wonder why nothing changes.
The problem is not the tools. I believe the problem is how teams are introduced to them. People are shown AI. They are not given space to experience it. There is a big difference between watching a demo and building something yourself.
Training programs create awareness. They rarely create capability.
What is Atlassian's AI Builders Week
Atlassian runs a quarterly event called AI Builders Week. Hundreds of product managers and designers spend a dedicated week not learning about AI — but building with it. Real agents. Real code. Real prototypes pushed to real repositories.
Their most recent week produced 298 working agents. Daily usage of their internal coding agent spiked 15–20x. 96% of attendees gave it positive ratings. And 95% said they left with concrete changes they could implement immediately.
But the numbers are not the point.
Most AI training programs ask people to watch, listen, and absorb. Atlassian's program asks people to build. That distinction sounds small. It isn't.
When a Loom designer used Cursor and Rovo Dev to push hover animation changes directly to Bitbucket, adjusting timing in real time with on-screen sliders — she didn't just learn something new. She changed how she works. The endless back-and-forth with engineers over "should this be 80ms or 120ms?" became something she could own and resolve herself.
More than a training outcome, it is a capability shift. And capability shifts only happen through doing.
The Three Principles Behind It
AI Builders Week is not a conference. It is a methodology — a deliberate set of rules for how learning happens. Three principles make it work.
Dedicated time and permission to build: The week exists specifically for building. Not learning about building. Building. Virtual mentors and technical support are present throughout so teams can push through problems in real time without losing momentum.
Cross-craft participation: The hackathon was the innovation format of the last decade. It had real value but it was built primarily for developers.
Designers were invited, often as a minimum. AI Builders Week inverts that. Product managers, designers, and engineers build together as equals, because when AI can close the gap between design and production, the old separation stops making sense.
Real artifacts, not just knowledge: The capstone is a Build Day followed by a Demo Day. Teams present working prototypes — including failed experiments. The celebration is of learning, not just output.
One iteration is not enough. Atlassian runs this every quarter, each one going deeper. The first focused on foundations. The second on measurable impact. The third will go deeper still. That rhythm is not incidental — it is the mechanism.
AI Is New. This Challenge Isn't.
This principle is not new. The technology is.
I joined my first company as an Ergonome — what we would now call a UX designer, though the term barely existed then. There was no design department. I was implementing Design Agile methodologies in a single project, figuring it out as I went.
That summer, the company was testing an idea for a new investment. Design thinking was beginning to germinate internally — quietly, without a formal mandate.
A product manager noticed what I was doing and asked if we could use the summer slowdown — when everything went quiet — to try it together.
Five people. No mandate. No one told what to think. We lived the practice instead of reading about it — and that changed everything.
That act of doing produced a whole new design team. It embedded UX process into the company's existing agile methodology. The organizational structure changed — not because a strategy said it should, but because a group of people had experienced something they couldn't unfeel.
That was 2010. Atlassian's AI Builders Week is the same move at scale, with AI. A team carves out time. They live the practice together. The organization changes.
Different technology. Same principle.
Key Takeaway
AI adoption is widespread. But adoption is individual — people are using AI to speed up their own work. The coordination layer, where the real value is, remains largely untouched.
Atlassian's AI Builders Week works because of three things:
Dedicated time to build,
Genuine cross-craft participation, and
Real artifacts — not just learning outcomes.
The biggest transformations often start with the smallest acts of doing. A protected week. A small team choosing to live a new practice rather than be told about it. That is where capability actually forms — and that principle is older than AI.
AI Tools I'm Using This Week
💬 Claude — for synthesizing research from long articles and YouTube presentations into a coherent narrative before I write.
🎨 Grok — I created 4 agents that find the best sources about the companies I am researching. If you have a SuperGrok account, I recommend trying agents — the difference in research speed is real.
Speak your PR description, bug reproduction, or Cursor prompt. Wispr Flow auto-tags file names, preserves variable names, and formats everything for immediate paste into GitHub, Jira, or your editor.
No re-typing. No context gaps. No mangled syntax. Works natively inside Cursor, Warp, and every IDE at the system level.
4x faster than typing. 89% of messages sent with zero edits. Used by engineering teams at OpenAI, Vercel, and Clay.
That's all for this week. See you in the next one.
If your team is starting to think about AI adoption beyond individual tools, I'd love to hear where you are. Reply to this email — I read every one.
P.S. The hardest part of AI transformation is not finding the right tool. It is creating the conditions for your team to actually use it together. That is where I focus.