Our Collaborative Process

Working together as human and AI collaborators requires intentional process design. Here’s how we actually get things done.

Daily Workflow

Context Loading: Each session begins with Claude reading current project status, daily notes, and any overnight developments.

Collaborative Planning: We discuss what to work on, often using todo lists to track multi-step projects and maintain focus.

Parallel Execution: AJ handles domain decisions and user experience choices; Claude manages implementation details, documentation, and systematic tasks.

Continuous Documentation: Everything gets documented in real-time - decisions made, problems solved, lessons learned.

Project Management

Obsidian as Command Center: All project documentation, context, and coordination happens in AJ’s Obsidian vault, which Claude can read and update.

Git for Everything: Every project lives in version control from day one, with detailed commit messages tracking both technical changes and collaborative decisions.

Living Documentation: README files, setup guides, and process docs get updated continuously rather than written once and forgotten.

Communication Patterns

Explicit Decision Points: When we hit a technical fork in the road, we pause to discuss options rather than having Claude make assumptions.

Learning in Public: We document failures and wrong turns, not just successes, because that’s where the interesting lessons live.

Role Clarity: AJ makes final calls on user experience and strategic direction; Claude handles implementation consistency and technical research.

Quality Assurance

Cross-Review: AJ reviews Claude’s technical implementations; Claude reviews AJ’s documentation for clarity and completeness.

Real-World Testing: Everything gets tested in actual use cases, not just theoretical scenarios.

Iterative Improvement: We regularly review what’s working in our collaboration and adjust our process.

Context Management

Session Logs: Every work session gets documented with decisions made, problems solved, and next steps identified.

Change Tracking: Technical changes get logged separately from conversational context, so we can trace both what changed and why.

Memory Maintenance: Claude’s knowledge base gets updated with new insights, tool discoveries, and process improvements.

The Meta-Project

This collaborative process is itself an ongoing experiment. We’re not just building technical projects together - we’re figuring out how human-AI collaboration works in practice and documenting what we learn.

Every workflow improvement, communication pattern that works, and coordination mechanism that fails becomes part of our shared knowledge about effective human-AI teamwork.


Want to understand the technical framework behind this collaboration? Read about how AI collaboration works.