Agentic AI is mostly discussed at the individual level: one person, one agent, shipping faster than ever. That story is real. But moving a whole team to agentic ways of working is a completely different challenge.
Something I confirmed again when our AI team recently gathered in Budapest for a two-day all-hands: the more autonomous the team becomes, the more intentional you need to be about alignment. After six months of parallel shipping, we were finally all in one room.
Autonomy needs shared models
We run three parallel workstreams: one building our agentic DBOS and AI SDLC factory, one focused on Financial Intelligence, one driving company-wide adoption. Everyone moves fast, makes their own decisions and uses AI their own way. That's exactly what you want.
But autonomous individuals solving parallel problems will each optimize their own lane. A team needs shared mental models: a common intuition about what we're building, why it matters and how the pieces fit together. You don't get that from async updates.
Autonomous individuals optimize their own lane. A team needs shared intuition, not just a shared roadmap.
When you bring people together, something shifts. You stop presenting outputs and start sharing thinking. You discover the person two workstreams over solved the problem you're still wrestling with. You realign not just on roadmaps but on instincts.
Human trust forms in person
And there's one more layer. We added six new people in recent months. Professional trust forms over calls. Human trust forms in person. In a team that operates autonomously, human trust is the connective tissue that holds it together.
The more work we hand to agents, the more it matters to protect what stays uniquely human: shared direction, mutual trust and the ability to realign fast. Technology accelerates the individual. People hold the team together.
If you're thinking about how to move a whole team, not just individuals, to agentic ways of working, get in touch. I'm happy to share what works for us and what doesn't.
