Overview
Flow ID: onboarding-corpus-setup
Category: Onboarding & Initial Setup
Estimated Duration: 5-10 minutes (plus processing time if creating)
User Role: All Users (first time)
Complexity: Moderate
Purpose: Optional onboarding step that helps users upload or create their first dataset for AI-powered document querying. Can upload existing dataset file or start blockify job to create one from documents. This step introduces dataset query features.
Related Flows
- First-Time Application Setup - Parent onboarding
- Upload New Dataset - Full upload flow
- Create New Blockify Job - Full creation flow
Prerequisites
Before starting, users must have:
- Onboarding wizard active
- Embedding model available (or will upload during this step)
- Optionally: Dataset file or documents to process
- Understanding that this step is optional
Step-by-Step Flow
Main Path A - Skip Dataset Setup (Most Common)
Step 1: Dataset Setup Step Appears
- User Action: Arrive at dataset setup step
- System Response: Dataset introduction displays
- UI Elements Visible:
- Step header: "Add Knowledge Base (Optional)" or "Step 3: Datasets"
- Progress: "Step 3 of 4"
- Explanation text describing dataset features
- Two primary options:
- "Upload Existing Dataset" button
- "Create from Documents" button
- "Skip for Now" button (prominent secondary option)
- "Next" button
- Visual Cues:
- "Optional" clearly marked
- Skip option prominent
Step 2: Decide to Skip
- User Action: Click "Skip for Now" or "Next" without configuring
- System Response:
- Onboarding advances to next step or completion
- Dataset setup skipped
- No dataset configured
- UI Elements Visible: Next step or completion screen loads
Final Step: Skipped Dataset Setup
- Success Indicator: Advanced past dataset step
- System State Change: Dataset feature not configured; can add later
- Next Actions: Complete onboarding, configure datasets manually later if needed
Main Path B - Upload Existing Dataset
Step 1: Choose Upload Option
- User Action: Click "Upload Existing Dataset" button
- System Response: Upload interface appears within onboarding
- UI Elements Visible:
- File upload controls
- Dataset name field
- Embedding model selector
- Upload button
Step 2-5: Follow Upload Process
- User Action: Complete dataset upload (see corpus-upload.md for detailed steps)
- System Response: Dataset uploads and registers
- Note: Simplified upload interface within onboarding context
Step 6: Upload Completes
- User Action: Wait for completion
- System Response:
- Success message
- Dataset registered
- "Next" button enabled
Step 7: Proceed
- User Action: Click "Next"
- System Response: Advance to next onboarding step
Final Step: Dataset Uploaded via Onboarding
- Success Indicator: Dataset available in application
- System State Change: Dataset ready for use
- Next Actions: Complete onboarding, use dataset in chats
Main Path C - Create Dataset from Documents
Step 1: Choose Create Option
- User Action: Click "Create from Documents" button
- System Response: Simplified blockify job interface appears
- UI Elements Visible:
- Document upload area
- Basic configuration options
- Simplified from full blockify interface
- Explanation of what will happen
Step 2-4: Upload Documents
- User Action: Upload documents for processing
- System Response: Files upload and extract text
- Note: Simplified file upload within onboarding
Step 5: Start Processing
- User Action: Click "Start Processing" or "Create Dataset"
- System Response:
- Job starts in background
- Onboarding may advance while job processes
- Status indicator for job shown
Step 6: Complete Onboarding While Job Processes
- User Action: Click "Next" to finish onboarding
- System Response:
- Onboarding completes
- Job continues in background
- Can monitor via job status badge
Final Step: Dataset Job Started
- Success Indicator: Job processing in background
- System State Change: Blockify job created and running
- Next Actions: Monitor job, use dataset when complete
Alternative Paths & Strategies
Strategy A: Skip and Configure Later
When to use: Not ready to set up datasets yet
Steps:
- Click "Skip for Now"
- Complete rest of onboarding
- Add datasets later via Settings or Blockify page
Strategy B: Upload Multiple Datasets
When to use: Have several datasets ready
Steps:
- Upload first dataset
- Click "Add Another" if option available
- Upload additional datasets
- OR skip and add more later manually
Error States & Recovery
Error 1: No Embedding Model for Dataset
Cause: Embedding model not uploaded yet
User Experience:
- Error: "Embedding model required"
- Cannot upload or create dataset
Recovery Steps:
- Skip dataset step for now
- Complete onboarding
- Upload embedding model in settings
- Then add dataset manually
Error 2: Upload Fails
Cause: File error or system issue
User Experience:
- Upload error message
- Dataset not added
Recovery Steps:
- Try again
- Or skip and upload later
- Don't let this block onboarding completion
Version History
| Date | Version | Author | Changes |
|---|---|---|---|
| 2025-10-04 | 1.1 | Iternal Technologies | Initial comprehensive documentation |
Notes
Important Considerations:
- Dataset setup is completely optional
- Can be done anytime after onboarding
- Datasets not needed for basic chat
- Only needed for document-based querying
Best Practices:
- Skip during onboarding unless datasets ready
- Focus on completing basic setup first
- Add datasets after understanding chat basics
- Use full dataset workflow later for better control
Common User Questions:
- "Do I need a dataset?" - No, optional; only for document querying
- "Can I add later?" - Yes, anytime via Settings or Blockify
- "Should I skip?" - Yes if not ready; easy to add later
- "What's the difference between upload and create?" - Upload: existing file; Create: process documents with AI
Trigger
What initiates this flow:
- User manually initiates
- System event (reaches dataset setup step in onboarding)
Specific trigger: Onboarding wizard advances to optional dataset setup step (typically Step 3 or 4).
User Intent Analysis
Primary Intent
Optionally configure dataset features during initial setup to enable AI conversations grounded in specific documents.
Secondary Intents
- Understand dataset query capabilities
- Set up knowledge base early
- Enable advanced features from start
Subintents
- Learn what datasets are for
- Decide if feature is needed immediately
- Complete setup if relevant to use case