(888) 881-1066
Login
ech
win
  • Pricing
  • Platform
  • White Label
Home
Quick Start
Getting Started
Deployment Checklist
Core Concepts
Crafting Instructions
Special Instructions
Tool Usage
Language Support
Awareness
Knowledgebase
Guides
Call Routing
Call Screening
Customer Service
Migration Guide
Tools
Send SMS
Send Email
Send Text Message To Caller
Create CRM Note
Change CRM Stage
Transfer Calls
Call Webhook

Crafting Instructions

Learn how you can craft the perfect instructions for your AI receptionist.

With the latest generation of our platform, we have hidden a lot of complexities of building an AI receptionist agent behind a simple document editor interface. While the instructions section might just look like a simple document editor, it is very powerful. In this guide, we will explore how we can craft the best instructions for our receptionists.

Screenshot 2024 12 04 at 1.31.39 Pm

Instruct Like It's For A Human Receptionist

When crafting instructions for your AI receptionist, approach it as if you were training a new human staff member. Use clear and concise language to outline tasks, protocols, and responses to common inquiries. Provide context where necessary, such as details about your company culture, customer service standards, and frequently asked questions. By communicating your expectations in a human-centric way, you enable the AI to better understand and emulate the nuances of personal interaction, resulting in a more natural and effective customer experience.

Instructions vs. Knowledge Base

When determining how your AI receptionist should handle calls, remember that instructions always take precedence over information found in the knowledge base. Use the instructions section to set the overall tone, protocols, and non-specific, high-level guidance on managing interactions. This ensures the AI consistently follows your policies and standards.

For any detailed or fact-specific information—such as product specifications, pricing, or team member contact details—move that content into the knowledge base. The AI will reference these details only when needed, avoiding confusion and ensuring that your primary call-handling directives remain clear and authoritative. By maintaining this separation, you create a streamlined, efficient system where instructions guide the overall approach and the knowledge base supports the finer details.

Keep It Short

When providing instructions to your AI receptionist, brevity is essential. Large language models already have a vast understanding of language and common interactions, so overly detailed instructions can sometimes lead to confusion or unintended responses. Focus on conveying only the crucial points needed for everyday interactions. For information that isn't required in every conversation, consider moving it to the knowledge base using the "Knowledge" tab. The agent will reference this information when necessary, keeping interactions smooth and allowing the AI to function more effectively without being bogged down by excessive details.

Greeting Matters

First impressions are vital, and the greeting your AI receptionist delivers sets the tone for the entire customer interaction. Instruct the agent to offer a warm and friendly welcome that not only greets the caller but also outlines the common tasks it can assist with. This helps set clear expectations and guides callers efficiently. Don't forget to include any current promotions or special offers in the greeting to capture interest and provide added value. A well-crafted greeting enhances customer experience and reflects positively on your brand from the very first touchpoint.

Set Proper Boundaries

Defining clear boundaries for your AI receptionist is essential to ensure professional and effective interactions. In your instructions, outline the types of queries the agent should handle directly, such as basic inquiries or appointment scheduling. Specify which matters should be delegated by transferring the call or taking a message—for instance, complex issues that require human expertise. Also, identify any topics that the agent should politely decline to address, especially sensitive or critical matters pertinent to your industry. By setting these boundaries, you help the AI distinguish between different types of requests, ensuring that customers are guided appropriately while maintaining compliance and upholding your company's standards.

Test Every Scenario

Thorough testing is key to ensuring your AI receptionist operates seamlessly in all situations. Utilize the chat tab to simulate different conversations and refine your instructions accordingly. Take advantage of the voice feature within the chat tab to experience interactions as they would occur over the phone, providing insights into the agent's spoken communication skills. You can also call your agent directly to assess its performance in a real-world setting. By testing every possible scenario, you can identify and address any gaps or issues, ensuring your AI receptionist delivers an optimal experience for every customer interaction.

Screenshot 2024 12 04 at 12.52.54 Pm

If you wish to make your AI receptionist even more powerful, we suggest you head over to our Special Instructions guide.

Again, remember that your AI receptionist has reasoning and logic capabilities, and is capable of handling many different types of phone calls. Good luck!


Recommended Guides

Quick Start

Deployment Checklist

Learn how to deploy your AI receptionist on it's first day of work....

Tools

Send Email

Learn how to leverage the Send Email tool....

Guides

Migration Guide

Learn how to migrate from traditional business phone systems....

Get A Personalized Consultation
We are always here to answer any questions you may have.
Schedule a call with us!
Email Support
About UsContact UsFAQCareerReferral Program
Terms of ServicesPrivacy PolicyLimited Use DisclosureUser Data Deletion Policy
Free Phone Number LookupFree Business Name Generator
ech
win
Twitter
LinkedIn
Facebook
Instagram

©2025 echowin Inc. All rights reserved.
• Built and refined in Boulder, Austin and Kathmandu.