3-Step Custom AI Agent Development With
TechMonk's Agent Builder

How to Build An AI Agent With TechMonk’s AI Agent Builder Tool

Before we explore how to build an AI agent with TechMonk, let us understand the three components of TechMonk’s AI Agent Builder.

The Tool Library: Tools That Power Your AI Agents

The Tool Library is a collection of powerful tools designed to streamline different functions in your e-commerce business. Want to talk to customers, check your data, or connect with other services? TechMonk has a tool for each of these tasks.

But if you want a custom tool that performs a unique function, you can still build it with TechMonk.

Types of Tools in the Library

TechMonk provides a diverse set of prebuilt tools to help you handle a variety of tasks. Each tool within the library serves a unique purpose and can be customised further to meet specific business requirements. It empowers businesses to automate and streamline various processes.

And it also allows to to build different type of custom tools to build your own tool set that you can power your AI Agents with.:

• Text : The text tool is a straightforward utility that may integrate AI, but fundamentally, it functions as a simple wrapper for text. This tool is used for tasks like providing user-friendly greetings or responding to text-based interactions.
  
  An example would be engaging in a basic chat conversation, where the tool enables the exchange of text between the user and the system.
• API : API tools are designed to interact with external APIs. These tools leverage external services to perform specific actions. For instance, retrieving loyalty points from a system would involve calling an API that provides the relevant data.
• Code : The code option allows for the execution of custom code. Users can write and modify code according to their requirements, providing complete flexibility to design and implement specific tasks within the tool.
• RAG : The RAG (Retrieval-Augmented Generation) tool enables the upload of vector-based documents. This tool is ideal for knowledge retrieval and searching through uploaded content.
  
  For example, businesses can upload company or product documentation, and the RAG tool can search through these documents to answer queries, such as identifying the founders of a company.
• Form : The form tool is designed to capture specific data points and manage updates. It is especially useful for collecting information, such as generating leads. The form tool allows businesses to customise the fields and structure to suit their data collection needs.

In-Built Tools

These are ready-to-use tools already built into TechMonk’s Agent Builder. They help power the AI agents that come pre-built in your system. You can also use these built-in tools to create new AI agents that fit your business needs.

Here are some examples of pre-built tools in TechMonk’s library.

Custom Tools

The custom tools section shows all the AI agents you've created within TechMonk's system. When you build new AI agents, you can see it being listed here. Merchants can create tools of any type, including Form, RAG, Code, or Text.

Agent Library: AI Agents to Streamline Every Workflow

TechMonk lets you create agents using the tools in the system. You can build these AI agents to handle specific tasks and workflows. In addition, you can create custom AI solutions from the agent library for any purpose.

In-Built Tools

Here are some examples of TechMonk’s pre-built agents:

Here are some examples of pre-built tools in TechMonk’s library.

• • About Company Agent : This agent is used for any queries related to the company. It gives responses based on company information like contact details and general questions. It helps users get consistent and accurate answers about the company.
• • Bot Challenge Agent : This agent is made for casual chats. It can answer questions like "Who are you?" or handle simple greetings and farewells. It manages basic conversations between the user and the system.
• • Lead Generation Agent : This agent is focused on capturing leads. It helps collect user details to create business opportunities. You can use it for sign-ups, contact info, and other lead-related tasks.
• • Order Management Agent : This agent looks after anything related to orders. It helps users track, cancel, and view past orders. It can also help schedule deliveries, making it easier for users to manage their purchases.

Custom Agents

TechMonk allows e-commerce businesses to create custom AI agents tailored to specific needs. For example, one such custom agent our users built is the loyalty program agent. This agent is dedicated solely to handling inquiries related to the loyalty program. It is designed to respond to any questions users may have regarding loyalty points, rewards, redemption processes, or any other aspects of the loyalty program.

Each agent has its tools; you can integrate custom tools as needed. If you wish to enhance an agent’s capabilities, you can add new tools anytime. These tools can be connected to your existing software systems by wrapping them in an API, extending the agent’s functionality.

Agent Flow

TechMonk’s Agent Flow is a system comprising multiple AI agents designed to run specific workflows within a business. For example, when a user interacts through a chatbot or sends an email, Agent Flow can trigger the correct set of agents to complete the requested task or action.

When creating any workflow, you can incorporate AI Agents to manage different functions within that flow. For instance, in an e-commerce flow, specific agents correspond to various tasks, such as Ticket Management, Product Management, Product Purchase, and others.

You also have the option to create custom AI agents, such as a Loyalty Manager agent, which can be added to your existing e-commerce flow. Custom AI agents can be incorporated into any flow, offering flexibility in the design and functionality of your workflows.

When a request is made, the orchestrator, the central processing component of the agent system, determines which agent or set of agents is most suited to handle the task. The orchestrator selects the appropriate agent and can even use multiple agents in sequence to fulfil the request.

If a user poses several questions, the orchestrator ensures that the relevant agents respond accordingly. This flexibility allows you to evolve and customise workflows to meet changing requirements.

TechMonk’s AI sales agents and AI support agents represent prebuilt workflows. These are dedicated flows to automated e-commerce sales and support operations.

3-Step Custom AI Agent Development With TechMonk

Step 1: Building Custom Tools From The Tool Library

1. From the left sidebar of TechMonk, select Agent Builder and launch the Tool Library.

2. Open the Custom Tools section and select the Create New Tool button.

3. On the pop-up screen, enter the name and description of the tool. Also, choose the type of tool you wish to build.

4. Select the Next button at the bottom of the pop-up.

5. On the next screen, fill in the required fields and click the Next button that you see at the bottom of the pop-up.

6. Type in the custom code for the tool if needed and select Next.

7. Finally, add more buttons, enter the button details in the respective fields and click Done to save the tool.

Step 2: Building Custom AI Agents From The Agent Library

8. From TechMonk’s Agent Builder open Agent Library.

9. Open Custom Agents and select the Create New Agent button.

10. You can see a pop-up on your screen. In this window, enter the agent’s desiredName, Description, and Output.

11. Next, choose the pre-built and custom tools from the Tools library to empower your AI agent. Finally, click on Done to save your AI Agent.

Step 3: Creating An Agentic Workflow

12. From the Agent Builder, choose Agent Flow.

13. Type in the Name, Description, and Input Channels for your agentic workflow. Once you've done this, select the Next button at the bottom of the screen.

14. Select the AI agents for your workflow from the screen and click Next. You can also use the search bar to find any AI agent.

15. You can also choose customised tools depending on the tasks you want your agents to perform. Once done, select the Next button from the bottom of your screen.

16. Finally, select the Save as Draft or Save & Publish option as needed.