Skip to main content

System Tools

System tools are built-in capabilities that empower Amplifi agents to perform specialized tasks — such as searching the internet, querying structured data, or generating charts — beyond just language generation.

These tools act as modular plugins that agents can call to retrieve or process information intelligently. They are configurable, reusable, and built to solve domain-specific tasks with precision.


Why Use System Tools?

System tools enhance agent capabilities by enabling them to:

  • 🔍 Search real-time information from the web
  • 📚 Find relevant content from large document sets
  • 🧠 Understand and query tabular data
  • 📊 Turn insights into visual summaries

By attaching tools to an agent, you enable it to handle queries using the right approach based on the task.


Core System Tools in Amplifi

🟦 Web Search Tool

Purpose: Fetch the most up-to-date answers from the live web.
Great for time-sensitive questions or topics not covered in your internal documents — like “What’s the current repo rate?” or “Who won the last Formula 1 race?”


🟪 Vector Search Tool

Purpose: Find semantically relevant content from ingested documents.
Perfect for retrieving answers from PDFs, meeting notes, screenshots, or transcripts — even when keywords don’t exactly match.


🟨 Text to SQL Tool

Purpose: Understand natural language questions and generate meaningful answers from your structured databases.
Useful for answering things like “What were last month’s total sales?” or “List the top 3 performing regions this quarter.”


🟧 Visualization Tool

Purpose: Convert insights into dynamic charts and graphs.
Great for turning summaries into visual formats — like “Show a trend of sales over 6 months” or “Plot leads by region.”


Note

Tools like the MCP Tool are external integrations you bring into Amplifi — they are not built-in system tools, but powerful ways to extend Amplifi with your own enterprise logic or APIs. See the MCP Tool page for details.


How Agents Use Tools

When a tool is enabled for an agent, it becomes part of the agent’s reasoning system. At runtime:

  1. The agent analyzes the user’s question.
  2. It chooses the right tool based on context.
  3. It runs the tool in the background and responds with a refined, insightful answer.

Best Practices

  • 🎯 Assign tools based on what the agent needs to solve — not all tools are always required.
  • 🧪 Test each tool with a few trial queries to fine-tune agent instructions.
  • 🧩 Combine tools with well-crafted system prompts to maximize value.

System tools transform agents into smart problem-solvers — capable of reasoning, searching, and responding with data-backed intelligence.