What is the Adaptive Taxonomy?

Enterpret's Adaptive Taxonomy is a multi-level classification system that automatically organizes customer feedback into precise, actionable insights. Built from your unique business context on day one, it continuously evolves as your product grows—no manual maintenance required.

Why Adaptive Taxonomy Matters

The problem it solves:

Generic taxonomies can't keep up with business growth
Manual tagging consumes hours of team time
Flat keyword lists create overlaps and gaps
Insights lose trust when classifications don't make sense

How Adaptive Taxonomy is different:

Self-learning: Continuously evolves as your products and feedback patterns change
Multi-level hierarchy: Mirrors your actual product structure (L1 → L2 → L3)
MECE-compliant: Mutually Exclusive, Collectively Exhaustive—no overlaps or gaps
Transparent: Every classification is explainable and editable
Business-aligned: Reflects how your organization actually discusses products

The result: 70% less time managing taxonomy, with higher confidence in insights.

Understanding the Taxonomy Structure

Adaptive Taxonomy has two primary components that work together:

1. Keywords (L1/L2/L3) - What the feedback is about

Keywords represent aspects of your product organized in three hierarchical levels, from broad to specific.

2. Themes & Sub-themes - Why the feedback was given

Themes capture the underlying insight or intent, with Sub-themes providing granular detail.

Keywords: Three-Level Hierarchy

Keywords organize your product structure into three levels of specificity:

Level 1 (L1): High-level features or feature groups. Broad product areas or major functionalities.
Level 2 (L2): Features or feature aspects. More focused sections within the high-level feature.
Level 3 (L3): Specific Sub-features or Actions. Precise functionalities or specific user actions.

Example: Ridesharing Service Keywords

Level 1: Booking
├─ Level 2: Ride Scheduling
│  ├─ Level 3: Schedule future rides
│  └─ Level 3: Modify scheduled rides
└─ Level 2: Driver Assignment
   ├─ Level 3: Driver matching
   └─ Level 3: Driver arrival time

Level 1: Payment
├─ Level 2: Payment Methods
│  ├─ Level 3: Add credit card
│  └─ Level 3: Payment method errors
└─ Level 2: Billing
   ├─ Level 3: Receipt generation
   └─ Level 3: Refund processing

Keyword Components

Each Keyword consists of:

Name: How the Keyword is recognized across Enterpret
Description: Details that help the system understand and predict it reliably
Level: The hierarchical level it belongs to (L1, L2, or L3)

Special Keyword Types

Not Specified: Captures feedback that mentions only the parent concept without additional detail.
- Example: Feedback says "I love the playback feature" (mentions Playback but no L2/L3 specifics) → tagged as "Playback > Not Specified"
Misc (Miscellaneous): Contains feedback with specific patterns that don't fit existing child Keywords but aren't frequent enough to warrant their own dedicated Keyword yet.
- Example: Feedback mentions a niche playback scenario that appears occasionally but not enough to create a new L3 Keyword → tagged as "Playback > Offline Listening > Misc"

Themes & Sub-themes: Understanding Customer Intent

Themes are high-level insights that emerge from customer feedback, replacing the previous "Reasons" concept. They answer: Why was this feedback given?
Sub-themes provide specific details within a Theme, enabling deeper analysis while maintaining hierarchical organization.

Theme & Sub-theme Components

Each consists of:

Name: How it's identified across Enterpret
Category: Maps to one specific Category (Help, Improvement, Complaint, Praise)

Relationships

Theme → Sub-theme (One-to-Many):

Each Theme can have multiple Sub-themes
Each Sub-theme belongs to exactly one Theme

Theme ↔ Keyword (Many-to-Many):

A Theme must map to at least one Keyword but can map to multiple
A Keyword can have multiple Themes associated with it

Sub-theme ↔ Keyword (Many-to-Many via Theme):

Sub-themes inherit Keyword associations from their parent Theme

Special Sub-theme Types

General Sub-theme: A catch-all for feedback that matches the Theme but contains no additional specificity.
- Example: Feedback says "I had a billing issue" without details → tagged as "Billing Problems > General"
Misc Sub-theme:
- Captures feedback with specific patterns that don't fit existing Sub-themes but aren't frequent enough to warrant their own dedicated Sub-theme yet.

Example: How It All Works Together

Customer Feedback: "I deactivated my account, but I'm still being charged. I just wanted to make a few videos. I'm so frustrated and can't get anyone on the phone! Please give me a refund."

Enterpret identifies:

Keywords (What it's about):

L1: Account
L1: Payment → L2: Billing → L3: Refunds

Themes & Sub-themes (Why it was given):

Theme: Issues with unexpected charges → Sub-theme: Refund request to unexpected charge post account deactivation → Category: Complaint
Theme: Issues with Support availability → Sub-theme: Frustration with no response on phone support → Category: Complaint

Notice "Videos" isn't tagged as a Keyword because it's not central to the identified Themes. Enterpret considers both keyword presence and intent, ensuring accurate, relevant predictions.

The New Taxonomy Page

Overview

The Taxonomy page now provides a visual, hierarchical view of your entire classification system, making it easy to:

See L1/L2/L3 Keyword relationships at a glance
Understand Theme and Sub-theme mappings
Track how taxonomy evolves over time

Managing Your Taxonomy

The new management interface enables common operations with minimal effort:

Create: Add new Keywords, Themes, or Subthemes
Edit: Modify names, descriptions, and mappings
Move: Move keyword nodes across the hierarchy
Merge: Combine Themes and Subthemes
Archive: Remove outdated taxonomy items

Keyword management is available now, with Theme/Subtheme management coming soon later in Q4 2025.

[Todo: Image for managing Keywords]

Taxonomy Manager Agent (Coming in H1 2026)

A chat-like experience where you collaborate with an AI agent to manage taxonomy:

Context-aware suggestions: Leverages your product knowledge sources
Automated edits: Generate new nodes for product launches
Bulk operations: Promote, demote, or restructure taxonomy nodes
Immediate feedback: Visual preview of changes and their impact
Edit logs: Track taxonomy evolution over time

Example use case: When launching a new feature, describe it to the agent, and it will suggest all relevant L1/L2/L3 Keywords and associated Themes—no manual creation required.

Taxonomy Predictions in the Feedback Drawer

Taxonomy predictions are organized by Keyword paths (hierarchy nodes), making it easy to see:

The full path from L1 → L2 → L3
Associated Themes and Sub-themes
Category classifications

Visual organization improves:

Scannability: Find relevant classifications faster
Context: Understand where each prediction sits in your product structure
Clarity: See parent-child relationships at a glance

Keyword, Themes and Sub-Themes in Quantify

Plot or Breakdown by Keywords or Themes / Sub-Themes

Plot or breakdown by L1/L2/L3 keyword nodes or Themes and Sub-Themes with visual path selection

Deprecated Filters:

Old "Keywords" filter (replaced by hierarchical Keyword filter)
Old "Reasons" filter (replaced by Themes/Sub-themes filter)

Getting Started with Adaptive Taxonomy

Your Adaptive Taxonomy is created during white-glove onboarding based on:

Your existing Tracked Keywords
Knowledge base resources (Help Center, changelog, documentation)
Your product's unique business context

Once deployed, the taxonomy:

Arrives accurate on day one
Continuously evolves as feedback patterns emerge
Requires minimal manual maintenance
Provides explainable, editable classifications

Questions? Contact your Enterpret Customer Success team for taxonomy optimization guidance.

Taxonomy Management

Enterpret Features Explained

What is the Taxonomy?