Towards AI-Enabled Training Needs Analysis Using Dual AI-Agent Collaboration

Automated TNA System with Dual Agent Architecture

Nikilkumar Patel1,2 Peter J Barclay1 Janice McMillan2 David McGuire2
1School of Engineering, Computing, & the Built Environment
2The Business School
Edinburgh Napier University, Scotland

Abstract

A comprehensive Training Needs Analysis (TNA) is essential for effective HR development and organisational growth. However, traditional approaches often fall short due to limitations in scale, labour intensity, resource constraints, or expertise. To address these challenges, we propose an AI‑driven automated platform for conducting TNA at scale with unstructured data. Our prototype features an dual‑agent system powered by Large Language Models (LLMs). The Disseminator Agent performs knowledge extraction and deep data analysis, followed by the Formulator Agent producing novel intellectual ideas, actionable insights, and formatted TNA reports, facilitating final human verification, attestation, and decision‑making. We also outline a pragmatic plan for AI monitoring and platform evaluation—critical components for successful AI adoption in industrial settings. Our proposed design is currently being implemented for evaluation and for its future deployment in a business setting.

Step 1: Data Ingestion

Here are the set of 5, raw and unstructured documents we created for each of the different industries. The documents are AI generated and have been curated to reflect the noise, variations and complexities of real-world documents. Considering these as data inputs, the remaining part of the demo shows the outputs of the Disseminator and Formulator agents in their respective steps.

Step 2: Disseminator - Knowledge Extraction

The above unstructured documents are fed into the Disseminator agent, which extracts knowledge from the documents and creates a knowledge graph. The knowledge graph is then used to generate insights and recommendations for training needs analysis.

Step 3: Disseminator - Deep Data Analysis

OTP Framework Implementation

Our system implements the Organizational, Task, and Person (OTP) model (McGehee, W., & Thayer, P. W. (1961). Training in Business and Industry. John Wiley & Sons.) for comprehensive training needs analysis:

1

Extraction Phase (GPT4.1-mini)

  • Gather organizational goals and training initiatives
  • Identify key tasks and performance gaps
  • Extract employee skill profiles and training history
2

Research Phase (O3-mini)

  • Organizational Analysis: Assess goals, resources, and culture
  • Task Analysis: Identify required knowledge, skills, and abilities
  • Person Analysis: Determine individual training needs
  • Industry Context: Adapt analysis to domain-specific factors
3

Synthesis Phase (O3-mini)

  • Combine all analyses into coherent report
  • Generate actionable training recommendations
  • Provide industry-specific implementation guidance


Step 4: Formulator - Proposal Generation and Structured Output

Parameters & Stats

🤖

AI Models

Graph: Claude Sonnet 4

Entity extraction: GPT-4.1-mini

Deep reasoning & synthesis: O3-mini

Performance

Time: ~5-7 mins/full exection
Cost: $0.10-0.15 USD

🔬

Methodology

3-Phase: Extract → Research → Synthesize using OTP framework