# Role You are a Customer Service Automation Specialist who designs AI chatbot systems that handle routine inquiries, improve response times, and reduce support costs for small businesses. # Task Design a complete AI chatbot system for [YOUR_BUSINESS_TYPE] that automates [PERCENTAGE_TARGET] of customer inquiries while maintaining quality and knowing when to escalate to human support. # Instructions **Your Business Information:** - Business Type: [YOUR_BUSINESS_TYPE] - Products or Services Sold: [WHAT_YOU_SELL] - Current Support Channels: [EMAIL_PHONE_CHAT_SOCIAL] - Average Support Inquiries Per Week: [NUMBER] - Most Common Questions: [LIST_TOP_5_10_QUESTIONS] **Current Pain Points:** [SLOW_RESPONSE_TIMES_HIGH_COSTS_OVERWHELMED_STAFF_OTHER] **Goals:** - Target automation rate: [PERCENTAGE_OF_INQUIRIES] - Acceptable response time: [SECONDS_OR_MINUTES] - Budget for chatbot tools: [MONTHLY_AMOUNT] Based on this information: 1. **Use Case Prioritization**: Identify which inquiries to automate first: - **Tier 1 (Automate Immediately)**: Simple, repetitive questions with clear answers - Hours of operation, location, contact info - Order status checks - Pricing and product information - Return and refund policies - Shipping times and costs - **Tier 2 (Automate with Escalation)**: Questions requiring some logic but have clear paths - Troubleshooting common issues - Account management tasks - Appointment scheduling - Product recommendations based on needs - **Tier 3 (Human Only)**: Complex issues requiring judgment - Complaints and disputes - Custom requests or exceptions - Technical issues beyond basic troubleshooting - High-value or emotional situations 2. **Conversation Flow Design**: Map out chatbot conversation paths: - **Greeting**: Friendly welcome message, set expectations for what chatbot can help with - **Intent Detection**: Ask clarifying questions to understand what customer needs - **Decision Tree**: Branch based on customer response to appropriate solution - **Answer Delivery**: Provide clear, concise answer with next steps - **Escalation Trigger**: Recognize when to transfer to human (confusion, frustration, complexity) - **Satisfaction Check**: Ask if question was answered, offer human if not 3. **Knowledge Base Development**: Build the content your chatbot will reference: - FAQ database covering all common questions - Product specifications and details - Troubleshooting guides with step-by-step instructions - Policy documents (returns, shipping, privacy) - Links to relevant help articles or videos - Organize by topic and searchability 4. **Escalation Rules**: Define when chatbot hands off to human: - Customer explicitly requests human agent - Chatbot confidence score below threshold (unclear what customer wants) - Customer repeats same question multiple times (sign of frustration) - Sensitive topics detected (legal, medical, complaints) - High-value customer identified (VIP status) - After 3 failed resolution attempts - Provide seamless handoff with conversation history to human agent 5. **Tone and Personality**: Define chatbot voice: - Friendly and helpful (not robotic or overly formal) - Empathetic to customer frustration - Clear and concise (avoid jargon) - Professional but approachable - Aligned with your brand personality - Sample responses for different scenarios 6. **Platform and Technology Selection**: Choose chatbot tools based on budget and needs: - **Budget Option** (Free to $50/month): Tidio, ManyChat, Chatfuel (good for small businesses) - **Mid-Tier** ($50-$200/month): Intercom, Drift, HubSpot (more features, better analytics) - **Enterprise** ($200+/month): Zendesk, Salesforce Service Cloud (full support suite) - Consider ease of setup, integrations with your current tools, and scalability 7. **Integration Requirements**: Connect chatbot to your systems: - CRM (customer history and context) - Order management system (track orders, shipments) - Scheduling software (book appointments) - Email platform (send follow-up emails) - Analytics (track performance metrics) - Payment processor (process simple transactions if applicable) 8. **Training and Improvement**: Plan for continuous optimization: - Review chat transcripts weekly to identify gaps in knowledge base - Track questions chatbot failed to answer, add to knowledge base - Monitor escalation rate and customer satisfaction scores - A/B test different response phrasing - Update for new products, policies, or seasonal changes - Gather customer feedback on chatbot experience 9. **Human Agent Backup**: Ensure smooth human support when needed: - Define support hours for live agents - Set up queue system when all agents busy - Provide agents with full chat history - Train agents on common escalated issues - Create scripts for handling chatbot handoffs - Measure first response time and resolution time for escalated cases 10. **Success Metrics**: Track chatbot performance: - **Automation rate**: Percentage of inquiries resolved without human - **Resolution rate**: Percentage of customers satisfied with chatbot answer - **Average response time**: How fast chatbot responds - **Escalation rate**: How often customers need human help - **Customer satisfaction score**: Post-chat survey rating - **Cost per interaction**: Total cost divided by inquiries handled - **Time saved**: Hours freed up for human agents 11. **Implementation Timeline**: Roll out in phases: - **Week 1-2**: Set up platform, build initial knowledge base - **Week 3**: Create conversation flows for top 10 questions - **Week 4**: Soft launch to subset of customers, gather feedback - **Week 5-6**: Refine based on feedback, expand to all customers - **Week 7+**: Monitor, optimize, add more automation Provide a simple conversation flow diagram showing how customers will interact with your chatbot and when escalation occurs.