What Happens in an Apparatus 202 Workshop
A transparency piece: what the session covers, what participants build, and what they leave with.
202 is designed for firms that have already done foundational AI training - either our 101 program or equivalent. Everyone in the room can prompt reliably, understands how AI tools work, and has some experience applying them to their work. The question 202 answers is: what do you build next?
The format is either a half-day workshop or six 60-minute sessions spread across two to three weeks. Both cover the same six modules. The half-day format is more intensive; the session format gives participants time to apply what they learn between meetings.
What happens across the six modules
01
From Prompts to Skills
What a Skill is, how it differs from a prompt, and what makes a Skill useful to someone who did not write it. Participants identify one prompt from their current work that is a strong Skill candidate and begin the conversion process during the session.
02
Building a Firm Skills Library
The organizational design of a Skills library that actually gets used: naming conventions, use-case tagging, ownership, and governance. By the end of this module, the firm has the structure of its Skills library in place and a clear system for how it grows.
03
Cowork - Longer Tasks, Bigger Outputs
How to brief and manage multi-step AI sessions for complex tasks: due diligence, research synthesis, complex reporting. Participants identify a task type in their work that is a strong Cowork candidate and draft a briefing template for it.
04
Connecting to Internal Data
What MCP connectors are and what they make possible. How to evaluate which internal sources are worth connecting first, what to think through on permissions and scope, and how to prompt AI that has access to your firm's files. By the end of this module, at least one internal connection is set up and tested.
05
MCP Connectors to External Data
How to evaluate external data sources for connection value, how to set up connectors to databases the firm already subscribes to, and how to build Skills that use external data with appropriate sourcing and citation requirements.
06
Building Firm-Level Assets
Documenting workflows in a form that survives turnover, building onboarding assets that transfer institutional knowledge systematically, and identifying what is ready for the next level of investment - and what is not yet. Every participant leaves with at least one firm-level asset drafted by the session's end.
What participants leave with
The artifacts from a 202 engagement are more substantial than the deliverables from a foundations-level program, because the work involves building rather than just learning.
Firm Skills library
Seeded with real Skills from your team's workflows, organized and ready to use from day one after the session.
Mapped data connections
A documented plan for which internal and external sources to connect first, with at least one connection set up and tested during the session.
Documented workflows
At least one workflow written in a form any team member can follow - and that serves as a starting point for custom development at the 303 level.
Who it is for
202 is for firms whose teams have a working AI baseline - everyone can prompt, the firm has some shared infrastructure in place, and AI is already part of how regular work gets done. The typical 202 participant is not learning to use AI for the first time; they are learning to build something lasting with the capability they already have.
Prerequisites are practical rather than credential-based: a working prompt library, an active Claude Project with firm context, and a team that uses AI regularly enough to have opinions about what works and what does not. If those are in place, 202 is the right next step.
Full curriculum details are on the 202 curriculum page. If you want to talk through whether your firm is ready and what the engagement would look like, the best place to start is a 30-minute call.