OpenAI integration: Taxonomy Generator

Overview

The AI-powered taxonomy generator helps Admins quickly create or refine a learning taxonomy using OpenAI. With just a few prompts, the system suggests roles, associated skills, and even proficiency level definitions—saving time and improving consistency across your skills framework.

Whether you're starting from scratch or enhancing an existing taxonomy, this tool accelerates the process while maintaining quality.

Functionality Breakdown

What does the feature do?

• Uses OpenAI to generate a structured learning taxonomy based on your input

• Automatically suggests:

  • Job roles

  • Skills linked to each role

  • Proficiency level definitions per skill

• Enables exporting of the generated taxonomy as a CSV file

• Supports refinement of existing taxonomies through AI suggestions

Key Benefits

• Speed – Generate a comprehensive taxonomy in minutes rather than weeks

• Consistency – AI ensures logical connections between roles and skills

• Flexibility – Refine and customise the output to match your organisation's needs

Pre-requisites

To use the AI Taxonomy Generator, you'll need:

• Skills feature enabled on your account

• OpenAI integration enabled (contact your Customer Success Manager)

• Access to Company Settings → Learning Taxonomy

Required User Roles

Quick Start Guide

Step 1: Enable the Taxonomy Generator

1. Click your profile image in the top navigation bar

2. Select Company Settings

3. Navigate to Learning Taxonomy

4. Ensure the Taxonomy Generator toggle is enabled within the OpenAI integration page

Step 2: Launch the AI Workflow

1. Click on "Create skills taxonomy with AI"

2. This opens a guided workflow where you'll input prompts

Step 3: Enter Prompts

Add relevant inputs such as:

• Industry (e.g. Financial Services, Healthcare, Technology)

• Job roles (e.g. Sales Manager, Software Engineer, Customer Support Lead)

• Business functions (e.g. Marketing, Operations, Human Resources)

The AI will then:

• Suggest relevant roles

• Associate skills to each role

• Provide proficiency definitions for each skill level

💡 Tip: The more specific your prompts, the more relevant the output. "Skills for junior data analysts in retail" will yield better results than just "data skills".

Step 4: Review and Refine

1. Review the AI-generated suggestions

2. Make adjustments as needed—add, remove, or modify roles and skills

3. Ensure the taxonomy aligns with your organisation's terminology and structure

Step 5: Export as CSV

1. Once the workflow is complete, select "Export as CSV"

2. The file will include:

   • Role names

   • Associated skills

   • Skill-specific proficiency level definitions

3. Use this CSV to import the taxonomy into Learn Amp or share with stakeholders for review

Step 6: Refine Existing Taxonomy (Optional)

If you already have some data:

1. Use AI to improve your existing taxonomy

2. Provide further prompts or adjustments

3. Export an updated CSV with additions or enhancements

FAQs

Q: Do I need to know how to write prompts?
Not at all—simple terms like job titles or industry names are enough. The guided workflow supports you throughout the process.

Q: Will AI overwrite my existing taxonomy?
No, AI-generated suggestions are provided separately. The AI offers a CSV file with suggestions, which you can review and adjust before deciding to adopt or merge them into your current setup. This ensures you have full control over any changes made to your existing taxonomy.

Q: Is this available to learners?
No. Only Admins with access to the Learning Taxonomy area can use the AI features.

Q: Can I generate only specific skills or roles?
Yes. You can control the scope by tailoring your prompts accordingly (e.g. "Skills for junior data analysts in retail").

Q: What if the AI suggestions don't match our terminology?
The AI provides a starting point. You can edit, rename, or restructure any of the suggestions to match your organisation's language and culture.

Q: How accurate are the AI suggestions?
The AI provides well-informed suggestions based on industry standards, but you should always review and validate the output against your specific organisational needs.

Troubleshooting