Configuring Your Data Classification Taxonomy

Overview

The data classification taxonomy is the foundation of how Relyance AI recognizes and categorizes data across all your integrations. Each data element in your taxonomy represents a type of data you want to track — such as Bank Account Number or First Name — and is paired with one or more match rules that tell Relyance how to detect it in your source code, APIs, vendor connections, and data stores.

There are two types of data elements:

• Relyance AI elements — built-in elements maintained by Relyance, covering common personal and sensitive data types. These cannot be deleted but can be disabled.
• Custom elements — elements you create for data types specific to your organization. These are fully editable and deletable.

Navigating to Classification Settings

1. From the main navigation, click Settings.
2. Select Classification Settings.

The Classification Settings View

The Classification Settings view shows all data elements in your taxonomy. Each row represents one data element with the following columns:

• Name — the display name for the data element
• Category — the parent category this element belongs to (e.g., Natural Human Data, Financial Data)
• Context — regulatory or operational context tags (e.g., GDPR)
• Obligations — compliance obligations mapped to this element (e.g., GDPR, CCPA)

Filtering and grouping

Use the filter bar at the top to narrow your view:

• Created by User — shows only custom elements your team has created
• Created by Relyance — shows only built-in Relyance AI elements
• Highly Sensitive — shows only elements marked Critical or High sensitivity

Use Group By: Category to organize elements into their parent category groups.

Click Save View to save your current filter and grouping configuration for quick access later.

Creating a new data element

Click + Create Data Element in the top right to open the creation form. Provide a name, description, sensitivity level, category, and any regulatory context or obligations that apply. Once created, add match rules to tell Relyance how to detect it (see Match Rules below).

Data Element Details

Click any data element to open its detail view. There are three tabs: Details, Match Rules, and Audit Log.

The Details tab shows:

• Description — a plain-language description of the data type
• Sensitivity — the risk level assigned to this data element: Critical, High, Medium, Low, or Unset. Sensitivity affects how findings are surfaced in the Data Classification view and in quick filters.
• Data Classification Category — the parent grouping (e.g., Natural Human Data, Financial Data). This can be edited.
• Context — regulatory context tags such as GDPR or CCPA
• Obligations — the specific compliance obligations mapped to this element (e.g., GDPR, CCPA). These are used in ROPAs and obligation tracking across the platform.
• Created By / Creation Date — who created the element and when. Elements created by Relyance.AI are built-in.

Changes to sensitivity, category, and obligations take effect immediately across the platform, including ROPAs and the Data Classification findings view.

Match Rules

The Match Rules tab lists all the rules Relyance uses to detect this data element across your integrations. Each rule specifies a matching strategy and the patterns to look for.

Rule types

• Relyance AI — semantic detection powered by Relyance's built-in AI model (SDSS). Relyance AI rules understand context and meaning, not just exact patterns. These rules are not editable, but can be disabled using the toggle if they are generating too many false positives for your environment.
• Fuzzy — similarity-based matching with a configurable confidence threshold. Useful for field names that vary across systems (e.g., "fname", "first_name", "given_name").
• Exact — precise string matching. Use when you need to match a specific, well-known field name exactly.
• Regex — regular expression pattern matching. Use for structured formats like IDs, codes, or custom identifiers.

Each rule shows its Match Patterns, Findings count (how many active findings were generated by this rule), and an enable/disable Status toggle.

Relyance AI rules are marked with a star icon (✦) and cannot be edited. All other rules can be edited or deleted.

Creating a custom match rule

Click + Create Match Rule to open the match rule editor.

Configure the following:

• Name — a descriptive name for the rule
• Description — a note explaining what the rule detects and why
• Integrations that will run this Rule — where Relyance will apply this rule:
  • Vendor, Data Store, API Integrations — runs against your vendor connections, databases, and API payloads
  • Source Code Integrations — runs against your connected code repositories
• Match On — what part of the data to evaluate:
  • Key — matches against field names (e.g., "bank_account_number")
  • Value — matches against the actual data values
  • Key and Value — matches against both
• Match Type — Exact, Fuzzy, or Regex
• Classification Level — optional; overrides the element-level classification for findings generated by this specific rule
• Match Patterns— the keywords or patterns to match. Click + Add to add multiple patterns.

Testing a match rule

Use the Test panel on the right side of the Create Match Rule dialog to validate your rule before saving. Paste a sample field name or payload, then click Test to see whether your patterns would match. Click Generate to have Relyance AI suggest match patterns based on the element's description.

Audit Log

The Audit Log tab records all changes made to the data element — including edits to details, match rule additions and deletions, and status changes — along with who made each change and when.

Considerations

• Changes to match rules affect all findings going forward. Existing findings are not retroactively updated unless a rescan is triggered.
• Disabling a match rule stops new findings from being generated by that rule but does not remove existing findings.
• Data elements and their match rule configurations are shared across all platform users. Exercise caution when modifying built-in elements, as changes affect ROPAs, findings, and all downstream reporting.
• Relyance AI (SDSS) rules are updated periodically by Relyance as the detection model improves. These updates apply automatically.
• Custom data elements (Created by User) count against your tenant's element limit.