Stack AI Academy: #6 - How to Use Agentic Tools

Apr 22, 2025

Kevin Bartley

Customer Success at Stack AI

Stack AI is dedicated to helping users confidently build AI agents. To support this vision, we are launching Stack AI Academy—a comprehensive educational program crafted to assist and guide you at every stage of your experience.

Through Stack AI Academy, you’ll gain essential skills with interactive, step-by-step courses that showcase the platform’s features and provide practical, hands-on learning opportunities.

Welcome to Stack AI Academy — Course #6 — Agentic Tools. In this course, you’ll learn everything you need to know about agentic tools in Stack AI.

What is an Agentic Tool?

AI agents have the ability to make autonomous decisions about whether or not to take an action. However, due to the way teams build agentic workflows, many agents are designed to take certain actions on every run. In the workflow below, the AI agent will perform a web search on each run.

Agentic tools, on the other hand, are only executed if the AI agent decides to do so. That means the tool isn’t necessarily triggered on each run. The tool is used only if the user prompt requires the tool.

In the example below, the Web Search function is used as a tool for the LLM to reference, as opposed to a node on the workflow.

The tool is only executed when the LLM believes it needs data from the internet to respond to the prompt. Now the AI agent can call tools and functions on its own, based on its reasoning skills.

Teams might implement agentic tools for many different reasons. They can make AI agents more efficient, and lead to faster results. They work well with conversational agents that receive a variety of queries. And agentic tools can also prevent 3rd party systems from getting overwhelmed with requests.

Choosing Tools in Stack AI

In Stack AI, you can add Tools to your LLM by clicking the “Tool” button on the far right side of the LLM.

Clicking this button will give you a collection of tools to choose from.

Tool categories that users can select from include:

Airtable
Algolia
Azure SQL
BigQuery
Exa AI
Gmail
GSheets
HubSpot
Jira
Knowledge Base
LinkedIn
Make
MongoDB
MySQL
Notion
Oracle
Outlook
Pinecone
PostgreSQL
Regex
Salesforce
SerpAPI
Slack
Snowflake
Stack AI
Stripe
Synapse
Time
Weaviate
Web Search
Wolfram Alpha
Yahoo Finance
YouTube
Zapier
Zendesk

Choose from this list of 35 different apps and functions to equip your LLM with the tools it needs to be effective.

Provide the necessary input and configuration parameters in the settings section to allow the tool to function properly.

Use Case: News Summarizer (with SerpAPI Web Search)

Let’s set up a tool as an example. Choose the Web Search function under SerpAPI.

The tool is designed with several input and configuration parameters. Click the gear icon to adjust the tool.

From here, you can adjust the configuration parameters.

Now we can write instructions for the LLM. For this example, let’s specifically mention the tool and what to use it for.

Now let’s run the workflow.

Then publish the workflow. Navigate to the Export tab.

Choose Chat assistant for a UI, and give the AI agent a name and description.

Now launch the AI agent. Let’s instruct the agent to gather 5 articles for summarization.

You’ll see the article summaries, along with an acknowledgement at the top that the AI agent used the Web Search tool.

Combine Tools for More Powerful AI Agents

You can also combine different tools in the same LLM. This makes your AI agents more powerful and helpful to human users.

Let’s consider our example from the previous section. Let’s add the Yahoo Finance Analytics Tool to our LLM.

With these two tools, we can both search the web, and gather finance analytics data from Yahoo. Let’s change our Instructions to take advantage of both capabilities.

Now let’s test out the tool from the user interface.

The Yahoo Finance tool now supplements the web articles with pertinent metrics.

By using these two tools together, you can gain more detailed insights into financial news stories.

Custom Tools

Stack AI also gives you the ability to create custom tools. Although many tools are available from 3rd party platforms, your team might need to connect a proprietary tool or one that is otherwise unavailable. To build a custom tool, click on the tool button, and then press the “Custom” tab.

Custom tools are defined through API services, allowing you to integrate external functionality into your LLM. When you create a custom tool, you’ll describe your API endpoints and their capabilities.

Each API endpoint becomes a distinct tool that represents a specific action or operation in your system. The LLM will automatically understand how to use these endpoints and fill in the required parameters (like body and query parameters) based on the context and user input.

To create a custom tool, you need to include:

Tool Provider Name: Give your tool provider a descriptive name that represents the service or system
OpenAPI Schema: Provide the OpenAPI specification that defines your API endpoints. The schema must include:
- Server URLs for the API endpoints
- Complete endpoint definitions with:
  - Important! Clear descriptions explaining what each endpoint does and its purpose to help the LLM understand how to use them correctly
  - HTTP methods (GET, POST, PUT, etc.)
  - Path parameters
  - Query parameters for GET requests
  - Detailed request body schemas for POST/PUT requests
  - Response schemas
  - Required headers specific to endpoints
Common Headers (Optional): Define headers that should be applied across all endpoints, such as:
- Authentication headers (e.g. API keys)
- Custom headers required by your API

Each API endpoint defined in your OpenAPI schema will be automatically transformed into an individual tool that you can use in your LLMs. Taking time to properly configure these settings will make your tools more user-friendly and reliable.

Your custom tool will now appear in the tools panel and can be dragged into any LLM.