Chatbot Designer

Ecommerce customer service teams drown in repetitive tickets — "where's my order," "how do I return this," "does this come in blue" — while high-value conversations that need human nuance get buried in the queue. A well-designed chatbot handles the predictable inquiries instantly and routes the complex ones to agents with full context. But most chatbot implementations fail because the conversation flows are designed by engineers guessing at customer intent rather than being mapped from actual support patterns. This skill designs complete chatbot conversation architectures grounded in ecommerce-specific support workflows, so you launch a bot that genuinely deflects tickets instead of frustrating customers into demanding a human.

Use when

• You are setting up a customer service chatbot for your Shopify, Amazon, or TikTok Shop store and need conversation flow diagrams before building in your chatbot platform
• A CX manager says "we need to reduce our ticket volume by 40% without hurting our CSAT score — design me a chatbot that can handle the top inquiry types"
• You are migrating from a basic FAQ bot to a multi-turn conversational chatbot and need structured dialogue trees for order tracking, returns, exchanges, product questions, and escalation paths
• Your current chatbot has a high abandonment rate or excessive escalation rate and you need to redesign the flows to actually resolve inquiries

What this skill does

This skill takes your store's support context — top inquiry categories, product types, return policy, shipping carriers, and any platform-specific constraints — and generates a complete chatbot conversation architecture. It designs multi-turn dialogue flows for each major inquiry type, mapping out the decision tree from initial customer message through resolution or escalation. Each flow includes intent detection triggers (the phrases and keywords that route customers into the right flow), clarifying question sequences, API integration points where the bot needs to pull live data (order status, tracking numbers, inventory availability), response templates with personalization placeholders, and explicit escalation criteria that define when and how to hand off to a human agent. The skill also designs fallback handling for unrecognized intents, satisfaction measurement touchpoints, and a conversation analytics framework so you can measure and iterate on bot performance after launch.

Inputs required

• Top support inquiry categories (required): The most common types of customer inquiries your team handles, ranked by volume. Example: "1. Order status/tracking (35%), 2. Returns and exchanges (25%), 3. Product sizing questions (15%), 4. Shipping time estimates (10%), 5. Discount code issues (8%), 6. Other (7%)"
• Store policies summary (required): Your return policy, shipping policy, and any warranty or guarantee terms the bot needs to communicate accurately. Example: "30-day returns, free return shipping on defective items, customer pays return shipping on preference returns, 3-5 business day standard shipping"
• Product catalog context (required): A brief description of what you sell and any product-specific FAQ patterns. Example: "Women's athletic wear, sizes XS-3XL. Common questions: fabric composition, size chart accuracy, sports bra support level, washing instructions"
• Chatbot platform (optional): The tool you plan to build in (Tidio, Gorgias, Zendesk, Intercom, custom). If specified, the skill tailors integration recommendations and flow formatting to that platform's capabilities.
• Current escalation rate or CSAT score (optional): If you have baseline metrics from an existing bot or live chat, providing them helps the skill set realistic improvement targets and prioritize which flows to optimize first.

Output format

The output is structured as a Chatbot Architecture Document with five major sections. The first section is an Intent Map listing all supported customer intents with their trigger phrases, confidence thresholds, and routing priorities. The second section contains Conversation Flow Diagrams for each major intent — presented as structured decision trees showing each bot message, expected customer responses, branching logic, API call points, and terminal states (resolved, escalated, or abandoned). Each node in the flow includes the exact message template with personalization variables marked in brackets. The third section is an Escalation Rules Matrix defining the conditions that trigger human handoff, the context data passed to the agent, priority levels, and SLA targets for each escalation type. The fourth section covers Fallback and Edge Case Handling including unrecognized intent responses, repeated failure loops, profanity or frustration detection, and after-hours behavior. The fifth section is an Analytics and Iteration Framework specifying which KPIs to track (resolution rate, escalation rate, CSAT per flow, average turns to resolution) and a suggested A/B testing roadmap for optimizing underperforming flows. The document is designed to serve as a complete specification that a developer or chatbot platform administrator can implement directly.

Scope

• Designed for: ecommerce operators, customer experience managers, support team leads, and chatbot developers
• Platform context: Shopify, Amazon Seller Central, TikTok Shop, WooCommerce, and platform-agnostic DTC stores
• Language: English

Limitations

• Does not build or deploy the actual chatbot — this skill produces the conversation architecture and flow specifications, not executable bot code
• Does not integrate with your live order management or CRM system to pull real customer data during the design phase; API integration points are specified but must be connected during implementation
• Escalation rate reduction estimates are based on industry benchmarks for ecommerce chatbots and may vary based on your specific customer base complexity and product category