Make the work legible
A transcript is not shared understanding. Capture intent, decisions, tradeoffs, and evidence so the rest of the team can pick up the thread.
01 An independent field guide
The future of work is something we have to learn.
Practical guidance for building with AI in teams—so speed doesn't come at the cost of shared context, trust, or human judgment.
A note from the edge
This is a place for people figuring it out in public—without certainty theater, replacement fantasies, or fear dressed up as strategy.
Field notes onAI × teams × learning
02 The real challenge
AI can make one person remarkably fast.
The harder question is what happens next.
A developer can plan, prompt, and ship with an agent on their own. But the context behind that work often stays trapped in a private conversation. Code arrives faster while understanding arrives later—if it arrives at all.
Teams need new ways to see one another's intent, follow decisions, challenge assumptions, and bring agent-assisted work back into a shared product. Collaboration is not a feature we can bolt on after acceleration. It is the work.
Working model / 001Live question
Shared contextWork everyone can see,
question, and continue.
question, and continue.
Individual intelligence becomes organizational learning only when the context travels.
“The unit of change isn't the tool. It's the team.”
03 A practice, not a rollout
Build capability.
Not compliance.
Lasting change is learned together. It asks for clear systems and technical rigor, but also consent, curiosity, and enough psychological safety to admit what no one understands yet.
Preserve human agency
Useful adoption grows from curiosity, safety, and visible value. People need room to question the tools, shape the process, and choose to engage.
Rejoin the system
Agent-assisted work has to come back into the product and the team: reviewed, tested, explained, and connected to everything around it.
A working idea
Voluntary performance
People tend to bring more energy, care, and imagination to work they have chosen to enter. That does not mean change has no direction or accountability. It means participation is designed to be worth choosing.
The question for an AI initiative is not only “How do we get people to use it?” but “What would make people want to learn this way of working?”
04 Field notes
Questions worth
working on.
Early lines of inquiry for builders, technical leaders, and teams learning how to work with agents without losing the plot—or one another.
Shared context
In developmentContext is becoming product infrastructure
What should a team preserve when prompts, plans, code, and decisions are spread across people and agents?
Team practice
In developmentFrom solo acceleration to team coherence
Individual speed is useful. Collective understanding is what lets that speed compound instead of creating drag.
Change
In developmentWhat a humane AI pilot looks like
A practical starting point built with the people doing the work—not imposed on them after the decisions are made.
Engineering
In developmentHarness engineering for real teams
Constraints, tests, tools, and feedback loops that help agents contribute reliable work inside an existing system.
For teams in the middle of it
Learning in public.
Helping in practice.
I'm exploring how small companies and technical teams can adopt AI thoughtfully: finding useful opportunities, designing credible pilots, improving agentic engineering workflows, and developing capability inside the organization.
The goal is not dependency on another consultant. It is a team that can make better decisions and keep learning after the engagement ends.
Tell me what your team is working through05 Open a conversation
What are you trying to learn—or change?
Share a little about the team, the friction, or the possibility you see. You don't need to have the problem perfectly framed.
Open to thoughtful advisory, collaboration, and project inquiries.