Prompt

Create a customer support chatbot script for a [business type, e.g., e-commerce retail] that can handle common customer queries. Include a friendly greeting, and then for each frequent question or issue (such as [example query 1], [example query 2], ...), provide the chatbot's response. Ensure the tone is [desired tone, e.g., warm and helpful] and that the answers contain correct, helpful information (for instance, if a customer asks about order status, the bot should prompt for an order number and then provide an update). Also, include a polite fallback response for questions the bot cannot handle (e.g., offering to connect to a human agent).

How to Use

1. Define Your Inputs: List out the most common questions or issues your customers have. These could be things like “Where is my order?”, “How do I return an item?”, “What are your support hours?”, etc. Also clarify your business details (the kind of products/services you offer, any specific policy info like shipping times or refund conditions) because the AI’s answers should be accurate to those details. Decide on the chatbot’s tone/personality — should it be very formal and polite, or casual and friendly?

2. Customize the Prompt: Replace the placeholder examples in the prompt with your actual frequent queries. Specify your business type to give context (e.g., “for an online clothing store”). Ensure the tone descriptor matches your brand voice (warm and helpful, professional, witty, etc.). If there are any specific instructions for certain questions (like “if order status is asked, the bot should ask for an order number”), include those details so the AI builds them into the script. Don’t forget to ask for a fallback response for unhandled questions (the prompt already does, just make sure it’s clear).

3. Optional Add-ons: If you want the conversation to flow as a multi-turn dialogue, mention that. For example, for complex issues, the bot might ask a follow-up question before resolving. You can also specify format preferences: maybe you want the output as a numbered list of Q&A pairs, or as a dialogue script. If you plan to import this into a chatbot platform, you could ask for a structured format like JSON or a certain markup (only if you know how the platform expects it).

4. Run the Prompt: Use your AI tool to generate the chatbot script. The AI will likely give you a sequence of interactions: a greeting message, a list of common query responses (each possibly labeled by the question), and a fallback message for anything it can’t address. Each response will be phrased as the bot speaking to the customer.

5. Review & Select: Read through the script and verify the accuracy and helpfulness of the answers. Check that the bot’s answers align with your actual policies and information (the AI might make assumptions if not given specifics, so correct any mismatches like wrong return period or missing steps). Ensure the tone is consistent throughout. Also, see if the flow makes sense – for instance, if a response would logically come after a user’s question. If something is off, refine the prompt with the correct info or desired change and run it again. You might also test a few sample user inputs mentally against the script to see if the bot’s responses cover them appropriately.

6. Expected Outcome: You’ll have a draft script for a customer support chatbot that covers greeting the customer, answering the most common questions, and politely handling anything unknown. For example, it might start with “Hello! How can I help you today?”, then include Q&A pairs like Q: "How do I track my order?" A: "Sure, I can help with that. Please enter your order number...", and so on for other FAQs, and end with a fallback “I’m sorry, I don’t have that information, let me connect you to a human agent.” This script can be plugged into a chatbot framework, significantly cutting down the time and effort to program your bot’s knowledge base.