Skip to main content

Semantic Grouping

Learn how Autonify uses AI to automatically organise your data into business domains and concepts. The semantic grouping agent analyses your entire data source and groups related tables into meaningful business categories.

VIDEO TUTORIAL⏱️ 2 min

📹 Organising Data with AI

Discover how semantic grouping helps you understand data relationships and business concepts across your database.

Understanding Semantic Grouping​

The semantic grouping agent provides high-level organisation of your data:

  • Automatic Discovery: AI analyses all tables and their relationships
  • Business Concepts: Groups data into meaningful business domains
  • Team Onboarding: Helps new team members understand data structure
  • Data Understanding: Reveals how tables relate to business functions

The Challenge It Solves​

Teams often face significant challenges when working with data:

  • Lack of Understanding: What data exists and what does it mean?
  • New Project Onboarding: Teams parachuted into unfamiliar projects
  • Application Complexity: Understanding data in new applications
  • Business Context: Connecting technical tables to business purposes

Semantic grouping addresses these challenges by providing an incredibly high-level view of your data organised by business meaning.

How It Works​

Agent Analysis​

When the semantic grouping agent runs:

  1. Scans the entire data source
  2. Analyses table names, columns, and relationships
  3. Identifies common business patterns
  4. Groups related tables into semantic categories
  5. Provides descriptions for each group

Example Business Domains​

In a banking database, the agent might identify:

  • Account Management: Customer accounts and related services
  • Audit & Compliance: Audit trails and compliance tracking
  • Customer Identity: Customer profiles and authentication
  • E-commerce: Online shopping and payment processing
  • Standing Orders: Recurring payment management
  • Loan Servicing: Loan accounts and payment tracking
  • Card Services: Credit/debit card management
  • User Access: Authentication and authorisation

Viewing Semantic Groups​

Accessing the View​

  1. Navigate to your data source
  2. Click on Agents in the sidebar menu
  3. View the Connected Agents tab (default tab)
  4. Find the Semantic Grouper agent card
  5. Click the View button on the agent card
  6. Opens the Agent Activity page

Agent Activity Page​

The page displays:

  • Page Title: "Agent Activity"
  • Agent Statistics Card: Shows:
    • Agent name (e.g., "Semantic Grouper")
    • Total Runs count
    • Success Rate percentage
    • Last Run date
    • Current Status badge
  • Semantic Groups View: Below the statistics card

Semantic Groups View

Display Options​

The interface provides:

  • Title: "Semantic Groups"
  • Description: "Semantic groups represent collections of tables that are related conceptually"
  • View Toggle Buttons: Cards and List modes

Card View​

  • Grid layout (1 column mobile, 2 columns tablet, 3 columns desktop)
  • Each card displays:
    • Layers icon with group name
    • Group description (if available)
    • Database section with database icon and name
    • Related Tables section with table badges
    • Schema labels shown separately when present

List View​

  • Vertical list with bordered containers
  • Each group shows:
    • Layers icon with group name inline
    • Group description below name
    • Database name with icon
    • Table badges in compact format (schema.table format)

Interface Features​

Search Functionality​

  • Search bar with placeholder "Search groups or tables..."
  • Real-time filtering as you type
  • Shows match count: "Found X groups matching 'query'"
  • Searches across:
    • Group names
    • Group descriptions
    • Database names
    • Table names
    • Schema names
  • Automatically shows all results when searching

Group Information​

Each semantic group card displays:

  • Group Name: Business concept with layers icon
  • Description: Explanation text (if provided by AI)
  • Database Section:
    • Label: "Database"
    • Database icon with database name
  • Related Tables Section:
    • Label: "Related Tables"
    • Table badges with outline style
    • Schema shown in grey label next to table badge
    • "No tables associated" message if empty

Understanding Group Details​

Table Relationships​

For each semantic group, you can see:

  • Primary Tables: Core tables for the business function
  • Supporting Tables: Related lookup or reference tables
  • Table Count: Number of tables in the group

Example Groupings​

Loan Servicing Group​

  • Tables:
    • loans - Loan account information
    • loan_payments - Payment history and schedules

Card Services Group​

  • Tables:
    • cards - Card account details
    • card_transactions - Transaction history

Benefits for Teams​

Quick Understanding​

  • Immediate overview of data landscape
  • Business context without deep technical knowledge
  • Visual organisation of complex schemas

Onboarding Efficiency​

  • New team members understand data faster
  • Reduces time to productivity
  • Clear business domain mapping

Data Discovery​

  • Find related tables quickly
  • Understand table purposes
  • Identify data dependencies

Documentation Aid​

  • Automatic business categorisation
  • Supplements technical documentation
  • Provides business perspective

Working with Groups​

Navigation​

  • Initially displays first 6 groups
  • Show More/Less Button:
    • Shows chevron down icon with "Show More (X more)" text
    • Click to reveal all groups
    • Changes to chevron up icon with "Show Less" text
    • Click again to collapse back to 6 groups
  • Search automatically shows all matching results
  • Toggle between Cards and List views using icon buttons

Integration with Other Features​

  • Groups complement the data catalog
  • Use alongside documentation agent
  • Reference for data quality rules

Best Practices​

Regular Updates​

  • Re-run agent after schema changes
  • Review groups after new table additions
  • Update when business logic changes

Team Collaboration​

  • Share group insights with team members
  • Use for onboarding documentation
  • Reference in technical discussions

Data Governance​

  • Use groups to identify data ownership
  • Apply consistent policies per group
  • Monitor sensitive data by business domain

Common Use Cases​

Project Onboarding​

  • Quick orientation for new developers
  • Understanding existing data structure
  • Identifying relevant tables for features

Data Migration​

  • Understand logical data groupings
  • Plan migration by business domain
  • Maintain relationships during transfer

Documentation​

  • Generate business-oriented documentation
  • Create data dictionaries by domain
  • Support non-technical stakeholders

Impact Analysis​

  • Assess changes by business area
  • Understand cross-domain dependencies
  • Plan updates systematically

Empty States​

No Agent Runs​

If the agent hasn't run yet:

  • Message: "No agent runs available"
  • Use the Run Agent button to start the first analysis

No Semantic Groups Found​

If the agent ran but found no groups:

  • Message: "No semantic groups found for this run"
  • May occur with very small databases or single-purpose schemas

No Search Results​

When searching returns no matches:

  • Message: "No semantic groups found matching your search criteria"
  • Try different search terms or clear the search

Limitations​

Agent Accuracy​

  • Groups are AI-generated suggestions
  • May require manual refinement
  • Based on available metadata

Complex Relationships​

  • Some tables may fit multiple groups
  • Cross-domain tables may be ambiguous
  • Requires business knowledge validation

Next Steps​

After reviewing semantic groups:

  1. Browse the data catalog
  2. Review auto-generated documentation
  3. View sensitivity classifications
  4. Track data changes