Skip to content
Merged
Show file tree
Hide file tree
Changes from 3 commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
61 changes: 24 additions & 37 deletions collate-ai/ask-collate.mdx
Original file line number Diff line number Diff line change
Expand Up @@ -8,7 +8,9 @@ collate: true

# Ask Collate

Ask Collate is the AI assistant inside Collate. It lets you explore data assets, generate or fix SQL, visualize insights, and enforce governance using natural language. Responses are grounded in the metadata, lineage, and usage information ingested into Collate.
Ask Collate is the AI assistant inside Collate. It lets you explore data assets, generate and run SQL queries, visualize insights, and enforce governance using natural language based on Collate’s Unified Knowledge Graph.

Moreover, AskCollate runs with the permissions granted to the user, ensuring compliance with every performed action.

## What you can do

Expand All @@ -21,25 +23,17 @@ Ask Collate is the AI assistant inside Collate. It lets you explore data assets,
allowFullScreen>
</iframe>

- Ask plain-language questions to discover tables, dashboards, owners, or usage patterns.
- Generate SQL with explanations you can review before running in your warehouse.
- Fix or optimize SQL by pasting an existing query and requesting improvements.
- Summarize metadata to understand table purpose, lineage, quality signals, and owners.
- Create and manage glossary terms, tags, and data quality tests through prompts.
- Navigate your data with plain-language and understand what’s the best asset to handle your use case.
- Run SQL queries, discover insights and visualize results.
- Dive deep into assets’ lineage and health.
- Understand your company’s language with the glossary, create new terms, and apply metadata changes.

## Prerequisites

- Collate AI is installed and enabled in your workspace.
- Metadata, lineage, and usage ingestion jobs are running so Ask Collate has context.
- You have permissions to view and act on the assets you are asking about.

## How to use

1. Click the **Ask Collate** chat icon in the Collate UI.
2. Enter your question in plain language; no SQL is required to start.
3. Use follow-up prompts to refine the answer, adjust filters, or change the visualization.
4. Review the generated SQL or insights, then run or apply them in your preferred workflow.

## Core Capabilities

- **Discover**: Search certified datasets, BI assets, and databases with metadata filters.
Expand All @@ -49,16 +43,25 @@ Ask Collate is the AI assistant inside Collate. It lets you explore data assets,
- **Lineage**: Visualize upstream and downstream dependencies to perform impact analysis.
- **Govern**: Create or edit glossary terms, assign owners, and tag assets from the chat.

## Reliable Prompt Patterns
## Use cases

- "Show me top marketing tables by usage this month. Only certified assets."
- "What does the `orders` table store and who owns it?"
- "Write a query for the last 90 days of active users by region."
- "Here is my query. Can you make it faster and explain the changes?"
- "Which Tier 1 assets are missing descriptions?"
- "Show lineage for `customer_360` and highlight downstream dashboards."
Let’s imagine a company trying to assess sales for a new product launched to the market. Different personas will have different needs and goals, and they all can use AskCollate to help get their job done for this project.

## Use cases
### Data steward: glossary and tagging

<iframe
width="800"
height="450"
src="https://www.youtube.com/embed/5Qa3VwOd8To"
frameBorder="0"
allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture"
allowFullScreen>
</iframe>

- **Prompt**: "Do we have a definition for revenue?"
- **Result**: Searches glossary terms and suggests creating one if missing.
- **Prompt**: "Create a term for revenue calculation and tag relevant Snowflake tables."
- **Result**: Adds the term, links it to assets such as `order_items`, and tracks ownership.

### Data analyst: assess eco-friendly product performance

Expand Down Expand Up @@ -103,22 +106,6 @@ Ask Collate is the AI assistant inside Collate. It lets you explore data assets,
- **Prompt**: "Create test cases for the products table."
- **Result**: Suggests completeness, uniqueness, range, and integrity tests and records who requested them.

### Data steward: glossary and tagging

<iframe
width="800"
height="450"
src="https://www.youtube.com/embed/5Qa3VwOd8To"
frameBorder="0"
allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture"
allowFullScreen>
</iframe>

- **Prompt**: "Do we have a definition for revenue?"
- **Result**: Searches glossary terms and suggests creating one if missing.
- **Prompt**: "Create a term for revenue calculation and tag relevant Snowflake tables."
- **Result**: Adds the term, links it to assets such as `order_items`, and tracks ownership.

## Best practices

- Be specific with filters (time ranges, owners, schemas) to narrow results.
Expand Down
7 changes: 4 additions & 3 deletions collate-ai/documentation-agent.mdx
Original file line number Diff line number Diff line change
Expand Up @@ -13,9 +13,10 @@ The Documentation Agent utilizes generative AI to automatically generate metadat

## Key Features

- **Automated Schema Labeling**: Generates clear and concise descriptions for tables and columns, reducing manual documentation efforts.
- **Natural Language SQL Query Generation**: Enables users to create SQL queries by interacting with the MetaPilot chatbot using everyday language.
- **Intelligent Query Assistance**: Assists in building, refining, and optimizing SQL queries, including handling table joins and relationships.
- Automated description generation: based on the context of your assets.
- Human in the loop: The agent will either update or suggest descriptions, keeping your teams the ability to accept or deny the agent’s requests.

<img noZoom src="/public/images/collate-ai/collate-ai-agent-feature.png" alt="suggest descriptions" />

## Setup Instructions

Expand Down
84 changes: 17 additions & 67 deletions collate-ai/index.mdx
Original file line number Diff line number Diff line change
Expand Up @@ -10,98 +10,48 @@ collate: true
# Collate AI

<div className="collate-ai-iframe">
<iframe width="800" height="450" src="https://www.youtube.com/embed/6glMYLzxNqk?start=0&end=260" frameBorder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture" allowFullScreen></iframe>
<iframe width="800" height="450" src="https://www.youtube.com/embed/XauK0jEr3jI?start=0&end=260" frameBorder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture" allowFullScreen></iframe>
</div>

Collate AI is an AI-powered tool within Collate that simplifies and enhances metadata management. By integrating generative AI, Collate AI assists users in automating the documentation of data assets, writing and optimizing SQL queries, and interacting with data through natural language. This first-of-its-kind data copilot improves productivity by automating tedious tasks and providing intelligent insights into your data environment.
Collate AI brings generative AI to metadata management, making it effortless to document data assets, generate and optimize SQL, and explore data using simple language. It boosts productivity by automating routine workflows and empowering all users from analysts to engineers to get more value from their data.

## Key Features

<Columns cols={2}>
<Card icon='/public/images/icons/mcp.svg' title="MCP Server" href="/how-to-guides/mcp" horizontal>
Enable AI agent communication with metadata to add context and take intelligent and automated action
</Card>
<Card icon='/public/images/icons/askcollate.svg' title="AskCollate" href="/collate-ai/ask-collate" horizontal>
Your conversational interface for querying metadata, documentation, lineage, and governance controls.
An AI agent that provides conversational analytics and data management capabilities built upon Collate’s Knowledge Graph.
</Card>
<Card icon='/public/images/icons/documentation.svg' title="Collate AI Documentation Agent" href="/collate-ai/documentation-agent" horizontal>
Automate description generation based on the context of your data.
</Card>
<Card icon='/public/images/icons/tier.svg' title="Collate AI Tier Agent" href="/collate-ai/tier-agent" horizontal>
Classify Business-Critical Data with Collate AI Tier Agent.
Classify business-critical data assets by usage & impact.
</Card>
<Card icon="/public/images/icons/qualityagent.svg" title="Collate AI Quality Agent" href="/collate-ai/quality-agent" horizontal>
Enhance Data Quality with Collate AI Quality Agent.
Automatically create Data Quality tests to ensure health and stability.
</Card>
<Card icon='/public/images/icons/queryrunner.svg' title="Collate AI SQL Agent" href="/collate-ai/quality-agent" horizontal>
Copy link

Copilot AI Jan 7, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The link href points to "/collate-ai/quality-agent" but the card title says "Collate AI SQL Agent". This appears to be a copy-paste error - the href should point to "/collate-ai/sql-agent" instead.

Suggested change
<Card icon='/public/images/icons/queryrunner.svg' title="Collate AI SQL Agent" href="/collate-ai/quality-agent" horizontal>
<Card icon='/public/images/icons/queryrunner.svg' title="Collate AI SQL Agent" href="/collate-ai/sql-agent" horizontal>

Copilot uses AI. Check for mistakes.
Understand and optimize the queries running in your sources.
</Card>
<Card icon='/public/images/icons/autopilot.svg' title="Autopilot" href="/collate-ai/collate-ai-autopilot-workflows" horizontal>
Streamlining Metadata with Collate AI and Autopilot Integration.
Streamline metadata ingestion capabilities for every service in your Data Platform.
</Card>
<Card icon='/public/images/icons/documentation.svg' title="Collate AI Documentation Agent" href="/collate-ai/documentation-agent" horizontal>
Automate Metadata Descriptions with Collate AI Documentation Agent.
<Card icon='/public/images/icons/mcp.svg' title="MCP Server" href="/collate-ai/mcp" horizontal>
An embedded, secure and enterprise-ready MCP server to bring metadata context to your AI agents.
</Card>
</Columns>

- **Automated Data Documentation**: Automatically generates metadata descriptions for tables and columns, saving time and effort for data owners and stewards.
- **Natural Language SQL Query Generation**: Allows users to interact with Collate AI through a chatbot to generate SQL queries by simply asking questions in plain English.
- **SQL Query Optimization and Fixing**: Capable of optimizing and troubleshooting SQL queries to improve their performance and efficiency.
- **AI-Driven Test Automation**: Automatically suggests and deploys relevant data quality tests based on table constraints and similarity with other assets, streamlining your data validation efforts.
- **Automated Tier Classification**: Uses data lineage and usage analytics to intelligently assign Tier levels, helping teams prioritize and govern critical data assets.
- **Natural Language Search**: Search for data assets using plain, user-friendly phrases — no complex syntax required. NLP interprets your query to return relevant results.

Examples:

> - **Tables owned by marketing** – Shows all assets owned by the marketing team.
> - **Tables with Tier 1 classification** – Lists all Tier 1 classified assets.

Click the icon again to switch back to standard search.

## Why Collate AI is Useful

### Metadata Management Challenges

Managing metadata across multiple data assets can be overwhelming due to the influx of new data and changing team dynamics. Collate AI addresses these challenges by:

- Automating metadata description generation.
- Simplifying the creation and optimization of SQL queries.
- Reducing manual effort and enhancing data quality.

### Time-Saving Features

Documenting thousands of tables manually is tedious and time-consuming. Collate AI automates metadata generation, allowing data teams to focus on high-value tasks and ensuring that data assets are consistently documented and understood across the organization.

## Use Cases

### 1. Automatic Data Asset Documentation

<img noZoom src="/public/images/collate-ai/reviewing-generated-metadata.png" alt="Automatic Data Asset Documentation" />

- **Problem**: Manually creating metadata descriptions for large datasets is labor-intensive and error-prone.
- **Solution**: Collate AI’s generative AI automates the process, analyzing database structures and suggesting accurate descriptions for tables and columns.
- **How It Works**: After configuring Collate AI, it scans the database schema and generates metadata descriptions automatically. Users can review these descriptions and approve or reject them in bulk or individually.
- **Benefit**: Streamlines the documentation process, ensuring consistent and up-to-date metadata across all datasets.

### 2. Natural Language SQL Query Generation

<img noZoom src="/public/images/collate-ai/using-collate-ai-chatbot-2.png" alt="Natural Language SQL Query Generation" />

- **Problem**: Non-technical users often struggle with writing SQL queries to extract insights from databases.
- **Solution**: Collate AI’s chatbot allows users to ask questions in natural language, generating SQL queries and providing explanations to help extract the required data quickly and easily.
- **How It Works**: Users can click on the Collate AI chatbot widget and ask questions like "Show me sales data from last quarter." Collate AI generates the corresponding SQL query and explains its logic.
- **Benefit**: Democratizes data access by enabling users of all technical levels to interact with data without deep SQL knowledge.

### 3. SQL Query Optimization and Troubleshooting

<img noZoom src="/public/images/collate-ai/fixing-sql-queries-3.png" alt="SQL Query Optimization and Troubleshooting" />

- **Problem**: SQL queries can become complex and inefficient, leading to performance issues and increased costs.
- **Solution**: Collate AI optimizes inefficient queries to improve performance, saving both time and resources.
- **How It Works**: If a query runs too long, users can request Collate AI to optimize it. Collate AI provides a more efficient SQL query version that can be implemented immediately.
- **Benefit**: Enhances query performance, reduces costs associated with inefficient queries, and speeds up data processing.

### 4. Fixing SQL Queries
## Why Collate AI

<img noZoom src="/public/images/collate-ai/fixing-sql-queries-4.png" alt="Fixing SQL Queries" />
Data only gains value when it is contextualized within your organization and your business. This requires understanding not only the data, but how the data is being used by different user personas, who owns it, and what is the business impact of each specific asset.

- **Problem**: Complex queries can cause issues even for SQL experts.
- **Solution**: Collate AI can fix problematic SQL queries, ensuring they run correctly and efficiently.
- **How It Works**: Users can ask Collate AI, "Can you fix this query for me?" It analyzes the query, detects issues, and returns a corrected version ready for use.
- **Benefit**: Simplifies writing and maintaining queries, allowing data teams to focus on analysis rather than troubleshooting.
Collate manages a Unified Knowledge Graph encompassing both your data and your organization’s semantics, which is the foundation to let your teams scale with your data by bringing automations and bridging the gap between people and data.

</div>
2 changes: 1 addition & 1 deletion how-to-guides/mcp/claude.mdx → collate-ai/mcp/claude.mdx
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
---
title: Getting Started with Claude Desktop
description: Set up MCP Server to connect Claude Desktop, generate tokens, and enable seamless AI-powered access to your data.
slug: /how-to-guides/mcp/claude
slug: /collate-ai/mcp/claude
sidebarTitle: Getting Started with Claude Desktop
---

Expand Down
4 changes: 2 additions & 2 deletions how-to-guides/mcp/connect.mdx → collate-ai/mcp/connect.mdx
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
---
title: Collate MCP Server Connection Guide
slug: /how-to-guides/mcp/connect
sidebarTitle: Connect
slug: /collate-ai/mcp/connect
sidebarTitle: MCP Server Connection
---

# MCP Server Connection Guide
Expand Down
2 changes: 1 addition & 1 deletion how-to-guides/mcp/goose.mdx → collate-ai/mcp/goose.mdx
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
---
title: Getting Started with Goose Desktop
description: Learn how to connect Goose Desktop with MCP Server, create tokens, and enable secure AI-powered access to your data platform.
slug: /how-to-guides/mcp/goose
slug: /collate-ai/mcp/goose
sidebarTitle: Getting Started with Goose Desktop
---

Expand Down
4 changes: 2 additions & 2 deletions how-to-guides/mcp/index.mdx → collate-ai/mcp/index.mdx
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
---
title: Data Discovery
slug: /how-to-guides/mcp
sidebarTitle: Overview
slug: /collate-ai/collate-ai/mcp
sidebarTitle: MCP Server
---

# Overview of MCP
Expand Down
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
---
title: Collate MCP Tools Reference
slug: /how-to-guides/mcp/reference
sidebarTitle: Reference
slug: /collate-ai/mcp/reference
sidebarTitle: Collate MCP Tools Reference
---

# Collate MCP Tools Reference
Expand Down
53 changes: 53 additions & 0 deletions collate-ai/sql-agent.mdx
Original file line number Diff line number Diff line change
@@ -0,0 +1,53 @@
---
title: Classify Business-Critical Data with Collate AI SQL Agent
slug: /collate-ai/sql-agent
sidebarTitle: Collate AI SQL Agent
collate: true
---

# Collate AI SQL Agent

Quickly understand and improve the performance of the queries that are executed in your system via a chat interface!

## Key Features

- Query explanation and breakdown
- Query performance improvements

<img noZoom src="/public/images/collate-ai/collate-ai-sql-agent-feature.png" alt="Ask Questions related to SQL Queries" />

## Setup Instructions

1. Open a Database

- Navigate to the database section in the platform.

- Open any database.

<img noZoom src="/public/images/collate-ai/collate-ai-sql-agent.png" alt="Open a Database" />

Example: In our case, we opened Customers.

2. Go to the Queries Tab

Inside the database view, locate and click the Queries tab.

<img noZoom src="/public/images/collate-ai/collate-ai-sql-agent1.png" alt="Go to the Queries Tab" />

3. Click the Ask Collate Icon

In the Queries tab, click on the Ask Collate icon.

<img noZoom src="/public/images/collate-ai/collate-ai-sql-agent2.png" alt="Click the Get Collate Icon" />

4. Redirect to Ask Collate

After clicking the icon, you will be redirected to the Ask Collate interface.

<img noZoom src="/public/images/collate-ai/collate-ai-sql-agent-feature.png" alt="Ask Questions related to SQL Queries" />

5. Ask Anything Related to SQL Queries

In Ask Collate, type your question or request.

You can ask anything related to SQL query creation, optimization, explanations, or filters.
3 changes: 3 additions & 0 deletions collate-ai/tier-agent.mdx
Original file line number Diff line number Diff line change
Expand Up @@ -14,6 +14,9 @@ The Tier Agent analyzes data usage and lineage to suggest appropriate tier class
## Key Features

- **Automated Tier Suggestions**: Assigns tier classifications to assets based on their usage patterns and lineage information.
- **Human in the loop**: The agent will either update or suggest tiers, keeping your teams the ability to accept or deny the agent’s requests.

<img noZoom src="/public/images/collate-ai/collate-ai-tier-agent-feature.png" alt="update or suggest tiers" />

## Setup Instructions

Expand Down
Loading