New Client Wizard — GrowthValet

⚠ Wizard Server is not running. The buttons on this page send commands to a local server on your Mac. To start it, open Terminal and run:

python3 "/Users/cjp/Documents/CJP Ops Brain/GrowthValet/gv_wizard_server.py"

Or to start it automatically every login, run this once:
launchctl load ~/Library/LaunchAgents/com.growthvalet.wizard.plist

Step 1 of 4

Who's the new client?

Fill these in once. Every folder name, command, and GHL KB entry generates automatically from here.

Client / Company Name

Full business name as you'd present it professionally.

Client name is required.

AI Assistant Name

The name callers will hear. First name, easy to pronounce.

Agent name is required.

GHL Knowledge Base Name

Exactly how it appears in GHL → AI Agents → Knowledge Base.

GHL KB name is required.

Folder Name (auto-generated)

—

Step 2 of 4

Create the client folder & run setup

One manual step in Finder, then one button click to build the full folder structure.

First: create this folder in Finder.
Navigate to Documents → CJP Ops Brain → GrowthValet and create a new folder named exactly:

—

Press ⌘+Shift+N to create a new folder, paste the name above.

Then click to build the folder structure (00_Intake through 07_Archive):

Step 3 of 4

Drop documents, then run the processor

Drop everything the client sent into 00_Intake, then let the processor do its work.

1

Open 00_Intake in Finder

Drag and drop the client's documents into this folder:
—
Supported: .docx · .doc · .pdf · .txt · .md — everything else is skipped safely.
2

Verify what's in Intake

Click to see a manifest of everything the processor will read:

Ready to Process

Run the 4-Pass Analysis

Reads all intake files, loads the KB masters, calls Claude, and runs Classification → Deduplication → Conflict Detection → Net New Extraction. Takes 30–90 seconds.

First run for this client? It will pause to propose a taxonomy and ask you to approve it in the output below. Type y in the Terminal window that opens, or wait — a future version will handle this inline.

Step 4 of 4 — Final Step

Upload updated masters to GHL

The processor just wrote new versioned master files. Upload them to GHL to make the AI assistant live.

1

Open GHL → AI Agents → Knowledge Base

Find or create the KB named:
—
2

Upload each updated master

Delete the old version first, then upload the new one from:
—
The terminal output above listed which files were written.
3

Wait for green Processed dot

Each file must show Processed status before the AI can use it.
4

Test the assistant

Use the GHL chat preview or place a test call to confirm it's responding with the new client's information.

✅ Client onboarded. AI assistant is live with a full knowledge base. When the client sends new documents, drop them in 00_Intake and run the processor again from Step 3.

← Back to Command Center