New matter scaffolding (when intake comes in)
Sally drops an intake PDF; ten minutes later the matter folder is scaffolded with brief, walling, and precedent links, instead of three hours of hand-rolling.
Three sections below. Pantry = your raw data. Prep table = the weekly briefs Claude assembles. Plate = the decisions you actually serve up. Click any card or recipe for plain English.
The mental model in one picture. Your data moves left to right: raw sources feed weekly silver platters, platters feed the briefs and drafts Claude writes for you, and those land in the channels you actually use (Slack, your inbox, your iPad). The AI layer sits on top of the data prep, never the other way around. Hover any line to trace it. Click a node to open the details.
If a Pantry box has no line leaving it, that data is sitting unused, those are usually the loudest quick wins. Dashed boxes on the right are channels we suggest adding.
Aldrich & Stone LLP, these are the briefs and drafts your team can stop building manually. Each one shows the ingredients, the time saved, and the back-and-forth Claude does on your behalf.
Sally drops an intake PDF; ten minutes later the matter folder is scaffolded with brief, walling, and precedent links, instead of three hours of hand-rolling.
Carol stops chasing partners on Friday. Every late time entry has a drafted narrative ready for the timekeeper to approve.
Marcus opens his iPad in the morning and sees the focus-areas brief for tomorrow's MSJ argument, three deposition admissions, the carve-out doctrine, the count-by-count framing.
When a matter closes, lessons get distilled into firm knowledge, separated from ephemeral case content, redacted, partner-approved. Knowledge stops dying at the bar.
Aldrich & Stone LLP, here's the order to build this in. Each step builds on the last, total: Bedrock approval is the long pole (1-2 weeks). Once Bedrock is live, the rest is a focused weekend.. If a step looks like a cliff, hand it to @claude-code-guide.
Install the Claude Code CLI. It's the tool that runs everything below. If you've never opened Terminal, that's fine, this is the only command-line moment until you're set up.
npm install -g @anthropic-ai/claude-code claude --version # should print a version number
claude --version in Terminal prints a version number instead of an error.
Setup time: 5 minutes.
Configure Claude Code to invoke models via AWS Bedrock instead of Anthropic.com. Sign Anthropic BAA. Get IT/security committee approval.
This is THE unlock. Without Bedrock, IT pulls the plug on every other step. No matter content can leave the firm tenant.
# AWS account + Bedrock model access (request via AWS console) # Set environment: export ANTHROPIC_BEDROCK_BASE_URL=https://bedrock-runtime.us-east-1.amazonaws.com export AWS_PROFILE=aldrich_stone # Configure Claude Code to use Bedrock backend
Setup time: 1-2 weeks (mostly committee review). Community has the deeper walkthrough.
Write rules/matter_handling.md with paths: data/matters/**. Defines that an agent reading one matter cannot access another in the same session.
Ethics walling is non-negotiable for litigation + M&A. Without this rule, the firm cannot use AI on client content.
# write under .claude/rules/matter_handling.md # YAML frontmatter: # --- # paths: # - data/matters/** # - silver_platters/matter_briefs/** # --- # Then the rule body (template in demos/02_boutique_law_firm/after/.claude/rules/)
Setup time: 30 minutes
SessionStart hook that converts pleading PDFs, deposition DOCX, intake PDFs.eml correspondence into markdown.
Every matter document arrives as PDF or DOCX. Claude can't read these natively. This hook unblocks Step 4.
cp -r skill/silver-platter/scripts/templates/.. ~/.claude/hooks/ brew install pandoc poppler pip install xlsx2csv
Setup time: 5 minutes
Define four subagents under .claude/agents/. Case-Launch is the chief-of-staff. The other three are matter-scoped specialists.
The orchestrator hierarchy is what makes the system usable. Sally and Marcus only talk to the orchestrator; specialists report up.
# write under .claude/agents/: # case-launch-orchestrator.md (model: opus, tools: [Read, Glob, Grep, Task]) # billing-assistant.md (model: sonnet) # deposition-prep-specialist.md (model: opus, tools include reading whole matter folders) # post-mortem-distiller.md (model: opus, redaction-aware) # Hand to @claude-code-guide for the templates.
Setup time: 2-3 hours
PostToolUse hook on Edit|Write that appends every action to outputs/audit_log.md. Stop hook that warns the partner at session end if a draft is unacknowledged.
For a regulated practice, the audit log is the difference between Bedrock approval and IT pulling the plug. Receipts.
# Already in business_os_hooks.md reference doc. # Paste the three hook entries into .claude/settings.json. # chmod +x the three scripts in .claude/hooks/.
Setup time: 10 minutes
Next: hand @claude-code-guide a one-line ask for your next skill, or join the Early AI Dopters community to see how other operators evolved theirs.