Create an Agent

Overview

At UnifyApps, we've built our AI Agents to work smartly by breaking down complex problems into simple, manageable steps. By following a logical sequence of reasoning, the agents can provide transparent and explainable solutions. To function effectively, agents require access to knowledge, use prompts to format responses, topics to execute actions and guardrails to ensure safe and accurate actions.

To effectively understand AI Agents by UnifyApps, let's go over the basic building blocks.

Knowledge

Think of Knowledge in an AI Agent like a skilled employee's training and experience. Just as an employee needs the right information to help customers, an AI Agent needs proper knowledge to give accurate answers and solve problems.

Within UnifyApps, this knowledge can be pre-loaded or dynamically gathered in real time from sources such as applications, documents or websites. This may include an organization’s internal knowledge articles, external applications like Sharepoint, Jira, Gitlab etc.

For Example, Insurance AI Agent can be fed with policy details, premiums, and coverage options, enabling it to provide quick, accurate answers and improve customer support.

UnifyApps honors application source Role-Based Access Control (RBAC). For example if a Google Doc is added as a knowledge source for an agent then only the users who have access to that particular document will be receiving answers based on the contents of the document.

Topics

Topics are like specialized skills that allow the AI Agent to perform specific tasks. If knowledge is what the agent knows, topics are what it can do. Each topic is a set of steps the agent follows to complete a task or a journey.

For example, an IT support agent manages password reset requests through a simple sequence: verify user identity, generate a temporary password, send a secure reset email, and log the activity. This ensures secure and efficient password management.

Prompts

Prompts are instructions or queries that help an AI agent determine its next course of action and formulate its responses. The clarity and precision of a prompt directly impact the accuracy and relevance of the agent's output, ensuring proper task completion.

For Example, in the response generation prompt of an AI Agent you can define the formatting of the final response generated by the agent i.e use emojis to make the answer rich, try to divide in bullet points by sections etc.

Guardrails

Guardrails ensure your AI Agent operates within safe, ethical, and business-aligned standards through predefined rules and boundaries. These guardrails are important for maintaining the reliability and integrity of AI Agents by preventing unintended actions, blocking inappropriate content, and ensuring compliance with company policies and regulations.

Why Are Guardrails Important?

Guardrails are essential for ensuring that AI Agents:

Provide accurate and reliable responses.
Avoid harmful or inappropriate actions.
Respect privacy and data security regulations.
Stay aligned with company policies and ethical standards.

Prerequisite Actions

Prerequisite actions for AI Agents are simple steps needed before the agent can start. These tasks include collecting necessary information, like user data or access permissions, to ensure the agent is ready to work.
Prerequisite actions are by default executed at start of each new session and it's essential that the task gets completed before the agent can execute any topic or actions.

For Example, a customer support agent can have a prerequisite action defined to publish a contact details form to capture the user information before actually triggering a topic or a journey based on the issue type.

UnifyApps enables you to create and manage multiple AI Agents, each designed for specific business needs and user personas. Whether you're building a customer support agent, an IT helpdesk assistant, or a sales support tool, our platform makes the creation process simple and straightforward.

Step-by-Step Creation Guide

Let’s create your first AI Agent by following these easy steps:

Click on the "Agents" option in the left sidebar of your UnifyApps dashboard in the Unify AI category.
On the Agent page, click on the ”+ New AI Agent” button on the top right. This will prompt you to create a new AI Agent.
Write the name of your agent & describe its role. Then, Click on the “Create” button.
Here’s how your AI Agent's landing page will look like.
You can also add multiple users as collaborators to the AI Agent by clicking on "+ Add Collaborator" where you will get an option to provide Admin, Viewer or Editor access to the users.