The Agent Builder is a streamlined alternative to the Workflow Builder that lets you create an AI agent in just a few clicks. In this video you’ll see how to connect private files and knowledge bases, pick the right LLM (cloud or on-prem), and brand the UI for your team. By the end, you’ll have a shareable, fully-featured financial assistant that can answer questions, cite sources, and even set reminders.

Summary

• Launch the Agent Builder via Quick Start directly from the dashboard

• Brand your chatbot: name, avatar, and theme

• Select an LLM provider (Anthropic, OpenAI, Google, or local on-prem) and model (e.g., Claude 3.5 Sonnet)

• Drag-and-drop files or link cloud drives (SharePoint, Google Drive, Notion, Dropbox, etc.) to create a knowledge base

• Enable advanced response settings: citations, thinking steps, suggested follow-ups

• See live demo: ask the bot to summarize recent financial docs and jump to cited passages

• Add tools for real action (task reminders, web search, Yahoo Finance look-ups, LinkedIn queries, Salesforce updates, and more)

• Example: set a reminder about Q1 filings, or retrieve a 30-day stock-price chart with one prompt

• Publish the agent, share the link company-wide, or deploy to an on-prem StackAI instance

• Empower any teammate—no coding required—to build a tailored AI co-pilot for their workflow

────────────────────────────────────────────────────────────────────────────────────

Course Overview

At Stack AI, we’re committed to helping every customer build AI agents with confidence. That’s why we’re launching Stack AI Academy — a new educational series created to support you every step of the way.

Stack AI Academy lays the groundwork for your success, with guided, step-by-step courses that walk you through the platform and showcase what it can do.

Welcome to Stack AI Academy - Course #3 - The Agent Builder. In this course, you’ll learn how to use the Agent Builder in StackAI to quickly launch AI agents. 

How to Use the Agent Builder

The Agent Builder is a faster and easier way to build AI agents in Stack AI. These AI Agents are deployed as Chat Assistants. 

Imagine a custom ChatGPT or Perplexity, powered by any LLM (OpenAI, Anthropic, DeepSeek, Llama), knowledge base, or tools—all built in seconds.

You can easily customize the title, description, LLM, knowledge base, and agentic tools to trigger — then go live almost instantly.

Let’s build our own AI agent in the Agent Builder! 

Start from the Stack AI dashboard. On the lefthand panel, you’ll see that there are four options: 

Choose Projects. We’ll come back to the other options later. If you already have projects, you’ll see them in each of the individual folders. 

Click New Project to create a new project. 

The Quick Start option allows you to launch a ChatGPT-type of AI agent instantly. 

The Workflower Builder enables you to build an AI agent with custom logic and components with drag-and-drop functionality. 

Let’s create a Quick Start agent. We’ll use the Quick Start agent to build a simple financial agent. Once you click Create, you’ll land on the agent builder interface.

The agent builder interface has two sections. The first is the user interface. This is where users will interact with your AI agent. 

The second is the configuration panel. This is where you can customize the pre-built AI agent for your specific use case. 

Give your AI agent a name and a description. 

Next, choose your LLM provider. For this example, we’ll choose Anthropic. 

If you want to host your LLM locally, choose the Local option.

Next, choose your LLM model. For the following use case, we’ll select Claude 3.5 Sonnet. 

For now, we’ll keep the default instructions in place.

Now add a Knowledge Base to the agent.

You can upload documents directly to the Knowledge Base. Or you can connect to a saved Knowledge Base in Sharepoint and other management tools. 

From this interface, you can preview Knowledge Base documents.

View the metadata of the documents.

See the parsed data.

And enable you to look at the individual chunks of the document.

You can also select Advanced Settings for the Knowledge Base.

With the Advanced Settings, you can:

• Use citations - Choose if you want to use citations in the response 

• Show steps - Show the thinking and reasoning steps of the AI agent

• Enable follow-ups - Show suggested follow-ups after the agent responds

Now that we’ve done this basic configuration, let’s see the AI agent in action. Ask the AI agent a question about the financial documents.

The AI agent will provide a response based on the financial documents in the Knowledge Base. 

You can view the footnote citations for each response.

Click on the citation to see where in the document the answer came from.

Select pre-built, contextualized follow-up questions to continue the conversation.

Now let’s add a Tool to the AI agent. Tools allow AI agents to execute apps and assets when they decide it is necessary, rather than on every execution of the workflow. 

Click on the Tool button and add the Task Scheduler Tool. 

Now you can schedule calendar reminders through your AI agent.

Try other tools, such as Yahoo Finance.

You can add web search tools to increase the helpfulness of your answers.

And that’s how you launch your first Quick Start agent. You can continue to customize the agent as you see fit.