User guides, tutorials, and demo

Install

1. Double-click the installer and follow the on-screen instructions
2. Accept the End User License Agreement (EULA)
3. Launch Sema4.ai Studio from your Applications folder (macOS) or Start menu (Windows)
4. Wait for the initial environment setup to complete (first launch will take several minutes)
5. You will see a welcome screen like this:

Build a new agent

Now that we have Sema4.ai Studio setup, let's build your first AI agent! We are going to build one from the scratch (but worry not, all details are provided in the tutorial).

1. Navigate to the Agents screen from the main navigation
2. Under My Agents tab, click button.

   

3. Choose Conversational as the agent type.

   

4. Give your agent a name, for example Sales Data Analyst.
5. Give your agent a description, for example Excel expert that helps analyze sales data and provides insights.
6. Next, we'll create a runbook for your agent. Click Create Runbook button to open the editor. By default the editor shows an outline of a runbook that you can use to get started.
7. We've made this easy for you: Download the runbook.md (opens in a new tab) file first.
8. Then click the import button with icon and select the runbook file you just downloaded with the file browser.

   

9. Click Save button in the top right corner to save your changes, and then close the runbook editor with the button.

   

Set up your LLM configuration

Now it's time to set up the Large Language Model (LLM) that will power your agent's intelligence. Sema4.ai Studio supports multiple LLM providers, including OpenAI, Azure, and Amazon.

1. In the agent configuration screen, locate the "Large Language Model" section (step 4 in the sidebar).
2. Click the Setup button under the LLM provider of your choice (e.g., OpenAI).
3. You'll see the provider's configuration panel. Since you don't have any models configured yet, click Create new model.

   

4. In the "New Model" dialog that appears, fill in the following details:

• Name: Give your LLM connection a concise and clear name (e.g., "OpenAI-GPT4")
• Configuration Type: LLM should be selected by default
• Provider: Ensure your chosen provider is selected (e.g., OpenAI)
• Model: Select the specific model you want to use from the dropdown
• API Key: Enter your API key for the selected provider

5. Click Save Model to store your configuration.

This LLM configuration will be saved in your Sema4.ai Studio, allowing you to reuse it for other agents you create in the future without having to re-enter the details.

Add actions

Next up, we'll add actions to your agent. Sema4.ai Studio has a gallery of pre-built actions that make it easy to connect agents to your enterprise apps.

1. Click Add Action.
2. Scroll to or search for Sema4.ai / Excel action package.

   

3. Click button to add a new action to your agent - it may take a bit as the action package and the Python environment are automatically loaded.
4. Then click Add to Agent button at the top.
5. Click Confirm.
6. Click Deploy on the bottom left to bring your agent to life!

   

Test your agent locally

Once agent deployment is complete, you'll see the chat interface along with the agent details on the right panel. You can download our sample Quarterly Sales Data 2024.xlsx (opens in a new tab) file for testing.

1. Create new Chat from the left panel
2. Upload the excel file and ask agent a question releveant to data, for example:

   

Show me the top selling products in Q2

3. Observe how the agent processes your request. It calls the Excel action first to get schema information and then the data. It then identifies products with respective revunue and shows you a formatted list of products.

   

4. Now try the follow up action.

   Yes, please create a new sheet with this summary data

5. Observe how agent calls the Excel action to create Worksheet and Add Rows action to store the summary.