Overview
The AI agent is an autonomous, non-deterministic program that can gather input and act on its knowledge to achieve a specific goal. Unlike simple AI chatbots that primarily converse, our AI agents can take real-world actions on your behalf, such as placing an order, booking a reservation, or fetching live information.
Core characteristics that define our AI agents include:
🚀Autonomous: They can operate independently to complete a task without constant human guidance.
⚡Action-Oriented: They independently execute different tools and functions.
🌐Interface Agnostic: They can interact through various means like voice, text, or other systems, not just chat windows.
🎲Non-Deterministic: They use reasoning and can adapt their approach based on new information, so they don't always produce the same output from the same input.
How Do The AI Agents Work?
The AI agents typically follow a cyclical process to complete tasks. One powerful way to think about this process is as a continuous loop:
Think➡️Act➡️Observe➡️Repeat ♾️
A more detailed breakdown of their workflow includes these steps:
🎯Define a Goal: The process starts with a user-defined objective, like a customer calling for help with a return.
📚Retrieve Information: The agent collects relevant knowledge and data needed for the task.
🧠Reason and Plan: The agent parses the collected information and creates a plan of action.
⚙️Take Action: The agent uses tools like APIs, databases, or webhooks to execute steps in its plan.
💾Utilize Memory: The agent can use memory about the person, like previous chat logs or purchase history, to personalize the interaction.
These steps are repeated until the goal is successfully achieved.
The Core Components of the AI Agent
Building a functional AI agent involves integrating several key components, much like building a capable employee. The table below breaks down these core parts:
Component | Role & Technical Mapping | Description |
|---|---|---|
Persona | The Agent's Role / Job Title | Defines the agent's core identity and traits. It is technically mapped to the system prompt of an LLM and persists throughout the interaction. Example: "You are a friendly customer support engineer." |
Instruction | The Job Description | Provides broad guidelines and the scope of tasks the agent is expected to handle, ensuring flexibility across different scenarios. |
Tools | The Agent's Hands | The external capabilities an agent can use, such as performing API calls, retrieving data via RAG, executing code, or connecting to webhooks for real-time information. |
Memory | Historical Context & Knowledge | Allows the agent to access historical data, such as past interactions or customer profiles, to provide context-aware and personalized service. |
Planning | Problem-Solving Strategy | The agent's ability to use reasoning techniques (like Chain-of-Thought) to break down complex problems, create multi-step plans, and adapt its approach. |
Where the AI Agents Shine
AI agents are particularly well-suited for tasks that are repetitive, require 24/7 availability, or involve open-ended, natural language inputs.
Common use cases include:
🤝Customer Support: Handling routine inquiries like checking order status, answering FAQs about products, or booking appointments, while seamlessly transferring complex issues to a human agent.
👤Personalized Interactions: Using customer memory to provide tailored experiences, like remembering specific delivery instructions.
🔄Dynamic Information Fetching: Using webhooks to pull real-time, context-specific data, such as product availability or weather updates.
The AI Agents options
See below for a list of options available to make your AI phone agent perfect for your use-case.
📋Instructions: Set clear and structured instructions for your AI agent to handle various customer interactions.
🔧Functions: Enhance your agent’s capabilities by integrating various functions like scheduling and handling service requests.
🔗Links: Give your agent access to external sources like FAQ or product pages to provide accurate information.
🎤Voice: Choose from a variety of AI TTS voices to best represent your brand and communicate effectively.
👋Greeting: Define what your agent should say when it answers the phone or chat to set a welcoming tone.
🔤Language: Select the language your agent should understand and respond in, with options including English, French, Spanish, German, Hindi, Russian, Portuguese, Japanese, Italian and Dutch.
➡️Next, see how to access the ✅AI Agent configuration page