System Tools
System tools are built-in capabilities that empower Amplifi agents to perform specialized tasks beyond natural language generation. These intelligent tools enable agents to search the web, query databases, analyze documents, create visualizations, and interact with various data sources seamlessly.
These tools act as modular plugins that agents can dynamically call to retrieve, process, and analyze information. They are configurable, reusable, and designed to solve complex business problems with precision and context-awareness.
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 real-time information from the internet to supplement your internal knowledge base.
Use Cases: Market research, competitor analysis, current events, regulatory updates, and any information not available in your organization's data sources.
Example Queries: "What are the latest industry trends in fintech?" or "Current exchange rates for USD to EUR"
Sample Response:
🔍 Web Search Results for "latest fintech trends 2024":
1. **AI-Powered Fraud Detection** - Banks are increasingly using machine learning algorithms to detect fraudulent transactions in real-time, reducing false positives by 40%.
2. **Embedded Finance Growth** - Non-financial companies are integrating financial services into their platforms, with the market expected to reach $7.2T by 2026.
3. **Digital Wallet Adoption** - Contactless payments now account for 65% of in-store transactions, driven by convenience and security features.
4. **Regulatory Technology (RegTech)** - Financial institutions are investing heavily in compliance automation to meet evolving regulatory requirements.
Source: Forbes Financial Technology Report 2024
🟪 Vector Search Tool
Purpose: Perform semantic search across all ingested documents and data sources using advanced AI embeddings.
Capabilities: Searches through PDFs, Word documents, presentations, meeting transcripts, images, and more using meaning-based matching rather than exact keywords.
Example Queries: "Find discussions about budget allocation" or "Locate product specifications mentioned in any document"
Sample Response:
🔍 Vector Search Results for "budget allocation strategies":
📄 **Q4 2023 Budget Planning Meeting** (Relevance: 94%)
"During our Q4 planning session, we discussed allocating 40% of the budget to product development, 30% to marketing initiatives, and 20% to operational improvements. The remaining 10% should be reserved for contingency planning."
📊 **Annual Financial Report 2023** (Relevance: 87%)
"The budget allocation for FY2023 prioritized digital transformation projects, with $2.3M allocated to infrastructure upgrades and $1.8M to employee training programs."
📈 **Marketing Strategy Document** (Relevance: 82%)
"Budget allocation recommendations: Increase digital marketing spend by 25% while reducing traditional advertising by 15% to align with shifting consumer behaviors."
Found 3 relevant matches from 1,247 indexed documents.
🟨 Text to SQL Tool
Purpose: Convert natural language queries into SQL statements and execute them against your connected databases.
Supported Databases: PostgreSQL, MySQL, Amazon Redshift, and other SQL-compatible data sources.
Example Queries: "Show revenue trends by product category over the last 6 months" or "Which customers have the highest lifetime value?"
Sample Response:
Hi List me top 10 movies with their rating?
Top 10 Movies with Their Ratings
• Antitrust Tomatoes
◦ Rating: NC-17
• Arabia Dogma
◦ Rating: NC-17
• Anonymous Human
◦ Rating: NC-17
• Alien Center
◦ Rating: NC-17
• Adaptation Holes
◦ Rating: NC-17
• Apache Divine
◦ Rating: NC-17
• Alley Evolution
◦ Rating: NC-17
• Alice Fantasia
◦ Rating: NC-17
• Aladdin Calendar
◦ Rating: NC-17
• Ark Ridgemont
◦ Rating: NC-17
Context Retrieved
SQL Query 1
Generated SQL Query:
SELECT title, rating
FROM nicer_but_slower_film_list
ORDER BY rating DESC
LIMIT 10;
title | rating
------------------|--------
Antitrust Tomatoes | NC-17
Arabia Dogma | NC-17
Anonymous Human | NC-17
Alien Center | NC-17
Adaptation Holes | NC-17
Apache Divine | NC-17
Alley Evolution | NC-17
Alice Fantasia | NC-17
Aladdin Calendar | NC-17
Ark Ridgemont | NC-17
🟧 Visualization Tool
Purpose: Generate interactive charts, graphs, and visual representations of data insights automatically.
Chart Types: Line charts, bar graphs, pie charts, scatter plots, heatmaps, and custom visualizations.
Example Queries: "create a graph of patient with their disability"
Sample Response:
🟫 File System Navigator
Purpose: Navigate through connected file systems and cloud storage to locate, read, and analyze files.
Supported Sources: Local file systems, Azure Blob Storage, AWS S3, and other connected storage systems.
Example Queries: "Find the most recent financial report" or "Locate all Excel files containing budget data"
Sample Response:
📁 File System Search Results for "most recent financial report":
📄 **Q4_2024_Financial_Report.pdf**
- Location: `/company/reports/financial/`
- Size: 2.3 MB
- Modified: December 15, 2024 at 14:32
- Type: PDF Document
📄 **Q3_2024_Financial_Report.xlsx**
- Location: `/company/reports/financial/`
- Size: 1.8 MB
- Modified: September 30, 2024 at 16:45
- Type: Excel Spreadsheet
📄 **Q2_2024_Financial_Report.docx**
- Location: `/company/reports/financial/`
- Size: 956 KB
- Modified: June 30, 2024 at 11:20
- Type: Word Document
Found 3 matching files in `/company/reports/financial/` directory.
🟥 Graph Search Tool
Purpose: Navigate through Amplifi's knowledge graph to understand relationships between entities, data sources, and business concepts.
Capabilities: Discovers connections between people, projects, documents, and data sources within your organization.
Example Queries: "Who is Alice?" or "What are the different entities and their relationships?"
Sample Response:
Who is Alice ?
Alice: Profile Overview
• Person
◦ Assigned to the Marketing project
◦ Responsible for Market Research
◦ Project start date: 2024-01-01
◦ Project duration: 13 days
◦ Progress: 78%
FA
Who is alice
Alice: Profile Overview
• Person
◦ Assigned to the Marketing project
◦ Responsible for Market Research
◦ Project start date: 2024-01-01
◦ Project duration: 13 days
◦ Progress: 78%
FA
What are the different entities and their relationships
Entities and Their Relationships
• Person
◦ Alice — assigned to Marketing (Project)
◦ Bob — assigned to Marketing (Project)
◦ Charlie — assigned to Marketing (Project)
◦ Daisy — assigned to Marketing (Project)
◦ Ethan — assigned to Product Development (Project)
◦ Fiona — assigned to Product Development (Project), assigned to Consulting (Project)
◦ Gabriel — assigned to Product Development (Project)
• Project
◦ Marketing — has members: Alice, Bob, Charlie, Daisy
◦ Product Development — has members: Ethan, Fiona, Gabriel
◦ Consulting — has member: Fiona
🟩 Video Search Tool
Purpose: Search through video content using AI-powered transcript analysis and visual recognition.
Capabilities: Indexes meeting recordings, training videos, presentations, and demos to find specific moments or topics.
Example Queries: "Give me the summary of video?", "What is hadoop as per the video?", or "Provide the summary of video which has cars on the road"
Sample Responses:
- For "Give me the summary of video?"
Summary of the Video: BigData.mp4
• **Parallel Processing**: Introduces distributed computing concepts with visual examples
• **Data Generation**: Shows data generation rates across platforms like Snapchat, Google, YouTube
• **Healthcare Applications**: Explains the 5 V's of big data with medical use cases
• **Hadoop & Frameworks**: Covers HDFS, MapReduce, and distributed processing
• **Real-world Impact**: Includes disaster management and smartphone data examples
- For "What is hadoop as per the video?"
Hadoop Overview:
• Uses HDFS for distributed storage across multiple machines
• Implements data redundancy for fault tolerance
• Employs MapReduce for parallel processing
• Enables efficient handling of large datasets
• Supports various big data frameworks like Spark and Cassandra
- For "Provide the summary of video which has cars on the road"
Urban Traffic Summary (47s):
• Golden hour scene: Sunrise/sunset over a busy urban highway
• Key Features:
- Heavy traffic in both directions
- Mix of modern and traditional architecture
- Prominent streetlights and high-rises
- Lens flare from low sun angle
- Faint rainbow effect in corner
• Ambiance: Serene yet dynamic cityscape
• Audio: Brief "Thank you" at 0-2s and 2-4s
◦ The video maintains a consistent and serene pace, emphasizing the congestion and the busy atmosphere of the cityscape. Throughout the video, there are no significant changes or movements, retaining a still moment captured in time under the soft glow of the low sun.