Map
Choose one real workflow and write down the inputs, decisions, handoffs, exceptions, and review criteria.
The workshop is hands-on: participants learn the new interface by turning one real process into a folder, memory, agent workflow, and reusable output.
The goal is not a lecture about AI. The goal is a live shift in how the team thinks about work, context, tools, and durable output.
Choose one real workflow and write down the inputs, decisions, handoffs, exceptions, and review criteria.
Create the project folder, add source materials, and write the first project memory file.
Open Claude Code or Codex in the folder, brief the agent, supervise the work, and inspect outputs.
Turn repeated steps into reusable instructions, skills, review loops, and team-ready artifacts.
Short videos can make the operating model obvious: create the project, add context, open the agent, supervise the work, and preserve what matters.
Why the folder becomes the workspace for agentic work.
Create the project, add sources, and open the agent in the right place.
Move from vague chat to clear objective, context, constraints, and verification.
Show what the agent knows, what should be written down, and what survives later.
Package repeated expertise and split specialist work without losing the thread.
Produce a dashboard, memo, deck, or workflow from the same project context.
A good workshop leaves behind working materials, not just enthusiasm. The team should be able to reopen the project and continue.
A usable structure for sources, notes, memory, outputs, and review.
The undocumented process made visible enough for people and agents.
Instructions, examples, and review standards that make the next pass easier.
A project that someone else can reopen without starting from zero.