Fast Onboarding Chatbot

Overview

The Batcher.ai Onboarding Chatbot is an intelligent assistant designed to help new users overcome the "blank spreadsheet syndrome" and quickly create their first AI-powered workflow. It provides a guided, conversational experience that transforms user needs into ready-to-use spreadsheet configurations.

Getting Started

Accessing the Chatbot

The onboarding chatbot is prominently displayed on the main spreadsheet page. Look for the button:

"Not sure where to start? Let our AI guide you!"

Click this button to launch the chatbot and begin your guided workflow creation journey.

What You'll Need

• Basic understanding of your data and goals

• 3-5 minutes to complete the process

• No technical knowledge required

---

How It Works

The chatbot uses a conversational approach to understand your needs and generate a custom workflow. It:

1. Learns about your industry - Identifies your sector and data type

2. Suggests relevant use cases - Presents tailored options based on your industry

3. Helps craft prompts - Assists in creating effective AI prompts

4. Recommends workflow type - Suggests the best workflow structure for your needs

5. Generates your workflow - Creates a ready-to-use configuration for the spreadsheet

Progress Tracking

A progress bar at the top shows your advancement through the 5 steps (20% per step). The chatbot also displays dynamic loading messages like "Crafting...", "Orchestrating...", or "Synthesizing..." while processing your requests.

---

The 5-Step Journey

Step 1: Sector Selection

Choose your industry from 8 pre-defined sectors:

• E-commerce - Online retail, product catalogs, inventory

• Marketing - Campaigns, content, customer engagement

• HR - Human resources, recruitment, employee management

• Finance - Financial analysis, reporting, forecasting

• Healthcare - Medical records, patient data, health information

• Education - Course content, student data, educational resources

• Legal - Legal documents, contracts, compliance

• Real Estate - Property listings, descriptions, market analysis

Step 2: Workflow Type Selection

Based on your sector, choose from 4 workflow types:

Simple Workflow

• Best for: Quick, straightforward transformations

• Example: Generate product descriptions from product names

• Structure: Input → AI Processing → Output

Advanced Workflow

• Best for: Comprehensive content generation

• Example: Create complete SEO packages (title, description, keywords, FAQ)

• Structure: Input → Multiple AI Outputs

Compare AI Providers

• Best for: Quality comparison and optimization

• Example: Generate descriptions using multiple AI models and combine the best

• Structure: Input → Multiple AI Providers → Best Combined Result

Multi-Step Workflow

• Best for: Complex, sequential processing

• Example: Enrich data → Verify facts → Analyze competition → Generate final content

• Structure: Input → Step 1 → Step 2 → Step 3 → Final Output

Step 3: Use Case Selection

The chatbot presents 4-6 dynamically generated use cases specific to your sector and workflow type. Examples include:

E-commerce + Simple Workflow:

• 📦 Generate SEO-optimized product descriptions

• 🏷 Create compelling product titles

• 🔍 Generate product keywords and tags

• 📝 Write product specifications

Marketing + Advanced Workflow:

• 📧 Create multi-channel campaign content

• 🎯 Generate personalized customer messages

• 📊 Produce A/B testing variations

• 🌐 Create localized content versions

Step 4: Prompt Refinement

The chatbot offers tailored prompt templates based on your selection. You can:

• Choose a suggested prompt template

• Modify the suggested prompts

• Write your own custom prompt

Example prompts are provided with clear descriptions of what they'll generate.

Step 5: Workflow Generation

Two options are presented:

1. Preview Mode - See a sample of what will be generated

2. Generate Now - Create the full workflow and import to spreadsheet

Upon choosing "Generate Now", the chatbot:

• Creates your custom workflow configuration

• Imports it directly into the spreadsheet

• Shows a success message with next steps

---

Workflow Types

Simple Workflow Formula Pattern

Column A: Input Data                                                                                            
Column B: =llm(CONCAT(`Your prompt here `, A2), `default`)

Advanced Workflow Formula Pattern

Column A: Input Data                                    
Column B: =llm(CONCAT(`Generate title for `, A2), `default`)
Column C: =llm(CONCAT(`Write description for `, A2), `default`)
Column D: =llm(CONCAT(`List keywords for `, A2), `default`)    
Column E: =llm(CONCAT(`Create FAQ for `, A2), `default`) 

Compare AI Formula Pattern

Column A: Input Data                                    
Column B: =llm(CONCAT(`Your prompt `, A2), `chatgpt`)   
Column C: =llm(CONCAT(`Your prompt `, A2), `gemini`)    
Column D: =llm(CONCAT(`Your prompt `, A2), `claude`)
Column E: =llm(CONCAT(`Combine best: `, B2, ` | `, C2, ` | `, D2), `default`)

Multi-Step Formula Pattern

Column A: Input Data                                    
Column B: =llm(CONCAT(`Step 1 prompt `, A2), `default`)
Column C: =llm(CONCAT(`Step 2 using `, B2), `perplexity`)
Column D: =llm(CONCAT(`Final step with `, C2), `default`)

---

Use Cases by Industry

E-commerce

• Product description generation

• SEO optimization for listings

• Category and tag creation

• Product comparison tables

• Customer review summaries

• Multi-language translations

Marketing

• Email campaign personalization

• Social media content creation

• Ad copy variations

• Content A/B testing

• Customer segment analysis

• Brand voice consistency

HR

• Job description writing

• Resume screening and scoring

• Employee performance summaries

• Training material generation

• Policy document creation

• Candidate evaluation reports

Finance

• Financial report generation

• Market analysis summaries

• Risk assessment documents

• Investment recommendations

• Budget forecasting narratives

• Compliance documentation

Healthcare

• Patient record summaries

• Treatment plan descriptions

• Medical report translations

• Insurance claim processing

• Health education content

• Clinical trial documentation

Education

• Course description writing

• Learning objective generation

• Student feedback analysis

• Curriculum planning

• Assessment question creation

• Educational content adaptation

Legal

• Contract summarization

• Legal document drafting

• Compliance checklists

• Case brief preparation

• Policy documentation

• Risk assessment reports

Real Estate

• Property descriptions

• Listing optimizations

• Market analysis reports

• Virtual tour scripts

• Neighborhood guides

• Investment summaries

---

Technical Details

Supported AI Providers

The chatbot can configure workflows using various AI providers:

• default - Batcher's optimized AI (Ollama)

• chatgpt - OpenAI's ChatGPT

• claude - Anthropic's Claude

• gemini - Google's Gemini

• perplexity - Perplexity AI for fact-checking

Formula Syntax Rules

1. Always use backticks (`) instead of quotes for text strings

2. Cell references (like A2, B3) don't need backticks

3. Use CONCAT() to combine text and cell references

4. The =llm() function requires two parameters:

   • The prompt (text or cell reference)

   • The AI provider configuration name

Example Formulas

// Simple formula
=llm(`Generate a product description`, `default`)                                                               

// With cell reference                                  
=llm(A2, `default`)                                     
                                                                                                                
// With concatenation
=llm(CONCAT(`Write SEO description for: `, A2), `chatgpt`)                                                                                                              
// Multi-cell reference
=llm(CONCAT(`Compare `, A2, ` with `, B2), `default`)

---

## Tips and Best Practices

### For Best Results

1. **Be specific about your data**
   - Instead of "I have products", say "I have an electronics product catalog"
   - Mention any special requirements upfront

2. **Start simple**
   - Begin with a Simple Workflow to understand the system
   - Graduate to advanced workflows as you gain confidence

3. **Test with sample data**
   - Use 5-10 rows of data initially
   - Verify results before processing large datasets

4. **Refine your prompts**
   - The chatbot's suggestions are starting points
   - Customize prompts for your specific tone and requirements

5. **Save successful workflows**
   - Export working configurations for reuse
   - Build a library of proven formulas

### Common Patterns

**For Product Descriptions:**

"Generate a compelling 150-word product description for [product] highlighting its key features and benefits"

**For SEO Content:**

"Create an SEO-optimized title (max 60 chars) and meta description (max 160 chars) for [topic]"

**For Data Enrichment:**

"Analyze [data] and add relevant categories, tags, and a brief summary"

---

## Troubleshooting

### Chatbot Doesn't Open
- Ensure JavaScript is enabled in your browser
- Try refreshing the page
- Clear browser cache and cookies

### Slow Response Times
- The AI is processing your request (normal for complex queries)
- Check loading messages for progress
- Wait up to 30 seconds for complex workflows

### Workflow Not Importing
- Ensure you're on the spreadsheet page
- Check that no other imports are in progress
- Try the "Reset" button and start over

### Unexpected Results
- Review your prompt for clarity
- Check that you selected the appropriate workflow type
- Try regenerating with slightly different wording

### Need to Start Over
- Use the "Reset" button in the chatbot footer
- This clears the conversation and returns to step 1
- Your spreadsheet data remains unchanged

---

## Support

If you encounter issues not covered in this documentation:

1. Check the main Batcher.ai documentation
2. Contact support through the platform

Remember: The chatbot is designed to be intuitive and helpful. Don't hesitate to experiment, you can always reset and try different approaches!

---