Welcome to the Flyflow Prompting Guide. This guide will help you design effective prompts for your agents to ensure they respond in a way that aligns with your application's goals. Whether you’re creating a customer support bot, a companionship application, or an auto-calling system, this guide will provide you with the best practices for setting up your agents' personalities, response styles, and scripts.

Agent Personality

The personality of your agent plays a crucial role in how users perceive and interact with your application. A well-defined personality can make interactions more engaging and human-like.

Defining Personality

When defining your agent's personality, consider the following attributes:

Tone: Friendly, professional, empathetic, humorous, etc.
Formality: Formal, semi-formal, informal
Energy Level: Enthusiastic, calm, neutral

Example

For a customer support bot:

Your personality is professional and empathetic. You respond in a calm and patient manner, always looking to help the user with their issues. You are formal in your communication.

For a companionship application:

Your personality is warm and friendly. You are empathetic and supportive, always looking to make the user feel heard and understood. You are informal and use a conversational tone.

Response Style

How your agent responds to user queries is critical for maintaining a coherent and satisfying user experience. Flyflow supports various response styles, from concise to detailed.

Concise Responses

Concise responses are short and to the point. They are useful for quick answers and when the user's time is limited.

Example

Keep your responses brief and to the point. Answer the user's questions directly without adding unnecessary information.

Medium Responses

Medium responses provide more detail than concise responses but avoid being overly verbose. They are suitable for most interactions, offering enough information to be helpful without overwhelming the user.

Example

Provide enough information to be helpful, but avoid going into too much detail. Aim for clarity and usefulness in your responses.

Avoid Repetition

Ensure that your agent does not repeat information unnecessarily. Repetition can frustrate users and make interactions less efficient.

Example

Do not repeat information that you have already provided unless explicitly asked to do so by the user.

Initial Message

The initial message is spoken immediately when the user picks up the phone. This message sets the tone for the interaction and provides context for the call. If you do not want to use an initial message, simply leave this field empty.

Crafting the Initial Message

Consider the following when creating your initial message

Greeting: A friendly introduction to start the call.
Purpose: Briefly explain the reason for the call.
Action: Indicate what the user should do next or what the agent will do.

Example

For a customer support call:

"Hello! This is Flyflow customer support. I'm calling to assist you with your recent query. How can I help you today?"

For an auto-calling application:

"Hi, this is an automated call from Flyflow. I'm here to remind you about your upcoming appointment. Please press 1 to confirm."

Script

Scripts help guide the interaction flow and ensure your agent provides consistent and relevant information. Scripts can include greetings, responses to common queries, and closure statements.

Creating a Script

Greetings: Start with a friendly and appropriate greeting.
Responses to Common Queries: Predefine answers to frequently asked questions.
Closures: End the conversation with a polite closure or offer further assistance.

Example Script for a Customer Support Agent:

Greeting:
"Hello! Thank you for calling Flyflow customer support. How can I assist you today?"

Common Query Responses:
- If the user asks about their account status:
  "I can help with that. Please provide me with your account number or email address."
- If the user asks for technical support:
  "I'm here to help with any technical issues you might be experiencing. Can you describe the problem in more detail?"

Closure:
"Thank you for contacting Flyflow customer support. If you have any more questions, feel free to call us again. Have a great day!"

Example Script for a Companionship Application

Greeting:
"Hi there! How are you feeling today?"

Common Query Responses:
- If the user talks about their day:
  "That sounds interesting! Tell me more about it."
- If the user mentions feeling down:
  "I'm sorry to hear that. I'm here to listen if you want to talk about what's bothering you."

Closure:
"It was great talking with you. Remember, I'm always here whenever you need someone to chat with. Take care!"

Conclusion

By carefully defining your agent's personality, response style, initial message, and script, you can create a more engaging and effective interaction experience for your users. Use this guide as a starting point and customize your prompts to best suit your application's unique needs. Happy prompting with Flyflow!