AI tool for talent development and upskilling
Client
The customer is a US-based company that manufactures microelectronic products using manual, automatic, and semi-automatic equipment across multiple facilities worldwide.
- Location: USA
- Industry: Manufacturing
- Client since: 2024
Key achievements
- 67%decrease in instructors’ workload
- 2xROI on training investments
- 25%reduction in employee turnover
Challenge
The client, a large US-based manufacturer, faced multiple challenges in the direction of corporate learning.
According to the L&D department’s internal research, they included the following:
Traditional training is focused on academic-style testing rather than evaluating real-world skills, which are critical in manufacturing for safety regulations, technical compliance, and operational efficiency.
Employees were disengaged with conventional learning methods, reducing participation and skill retention.
Fewer young professionals were entering apprenticeships, while many opted out of formal education, exacerbating workforce shortages.
The declining effectiveness of corporate training directly impacted productivity, compliance, and operational efficiency. Without an effective way to assess and upskill employees, the company risked increased safety incidents, regulatory penalties, and production inefficiencies.
The client needed a cost-effective solution to engage and retain employees, accurately assess skills, and ensure continuous workforce development.
Solution
Aristek’s team proposed an AI-based tool as a solution that would solve multiple problems in multiple areas of corporate training at once:
1. AI-powered learning assistance:
- Answer employee queries on various professional topics (e.g., leadership, compliance, technical skills).
- Provide instant explanations, summaries, and clarifications on corporate learning materials in a trial testing format.
- Maintain conversation context for ongoing learning discussions.
2. Automated administrative support:
- Manage training schedules and notify employees of upcoming learning sessions.
- Track learning history and suggest relevant topics for upskilling.
- Assist HR/L&D teams by automating FAQs about training programs.
3. Seamless integration with corporate systems:
- Connect with existing LMS (Learning Management Systems).
- Sync with internal knowledge bases, wikis, and document repositories.
- Support API-based integration with performance management tools.
Project scope
For the project, the Aristek team applied its AI and corporate education expertise through these steps:
The Aristek team collected training materials from various departments, ensuring full corporate knowledge coverage. The data was cleaned to remove inconsistencies, eliminate duplicates, and standardize formats for seamless integration.
AWS services facilitated processing, with AWS Glue handling ETL operations and AWS Lambda managing granular transformations and downstream tasks. This scalable, serverless approach ensured well-structured data for advanced analysis. NLP techniques were then applied to categorize information into structured clusters.
A database of materials from different departments was created using AWS OpenSearch, enabling efficient semantic search and retrieval. All segmented materials were stored with extensive metadata, including categories, clusters, and detailed tags, allowing precise filtering by topic, complexity, and relevance.
For question generation, the system leverages the OpenAI API with GPT-4o, using advanced prompt engineering and structured output. It dynamically generates questions based on parameters such as type (open/closed), difficulty level, and topic relevance, ensuring high-quality, contextually relevant assessments.
Tests are dynamically formulated for both practice (formative) and final (summative) assessments, reducing the chances of employees relying on pre-prepared answers. During the test, the AI can adjust difficulty levels in real time, ensuring an objective evaluation of an employee’s knowledge.
Thanks to the cluster-based storage system, a test on one topic can also include contextually related questions, assessing an employee’s ability to connect concepts across different areas. This is achieved by querying OpenSearch, using semantic similarity search and metadata filtering. The Retrieval Augmented Generation (RAG) approach fetches relevant information based on the user’s query and the test’s topic.
The system evaluates responses by comparing them to two key benchmarks:
- Golden answer – A reference answer curated by subject matter experts.
- AI-generated response – An answer dynamically generated from structured information clusters within the internal database.
If no predefined answers exist for a test question, the system autonomously generates them using relevant content from the internal knowledge base. This eliminates the need for manual content updates, significantly reducing costs and optimizing processing time.
Role-based analytics provides tailored insights based on user roles, ensuring both employees and managers can track progress effectively. While employees receive personalized feedback to enhance their learning, checkers and managers gain access to detailed reports for informed decision-making.
For employees:
A concise report summarizing performance.
Insights into strengths and areas for improvement (optionally).
Suggested learning materials for further development.
For checkers & managers:
A detailed analytical dashboard with real-time performance insights.
Downloadable PDF reports for each employee.
Key metrics of the response provided include:
Accuracy & correctness – 96% accuracy of provided answers.
Efficiency – decreased response time to 15 seconds and time taken per question.
Benchmarking – comparison against team averages and peer performance.
Support & accessibility – 24/7 availability for continuous learning.
Learning outcomes – 20% faster course completion and a 10% increase in correct final test answers.
How it works: practical cases
1. Automated skill assessment & adaptive testing
Previously, test questions were manually created and analyzed. Now, AI dynamically adjusts question difficulty in real-time based on employee performance, ensuring a more precise and efficient evaluation.
2. AI-driven qualification for high-precision equipment
Before assigning employees to operate complex machinery, the company evaluates problem-solving abilities using AI-powered assessments with scenario-based questions and open-ended troubleshooting tasks.
3. Personalized training & progression paths
Based on assessment results, employees are either advanced to hands-on training or provided with targeted learning modules to strengthen knowledge gaps before proceeding.
4. Role-based induction programs
AI customizes onboarding programs based on an employee’s role, ensuring they grasp essential company policies, operational guidelines, and safety protocols from day one.
5. Interactive safety simulations
Employees engage in AI-driven simulations to practice responding to hazardous situations, enhancing workplace safety and compliance. HR teams also leverage these insights to make data-driven decisions about probationary periods.
Tools & Technologies
- Python
- OpenAI
- GPT 4o
- RAG
- LlamaIndex
- OpenSearch
- AWS
- Jira
- Github
Project results
The Aristek team leveraged AI to transform corporate learning by developing a personalized, adaptive training and assessment system. The solution processes corporate materials, industry reports, and policies to align content with required competencies while assessing employees’ skills to create tailored learning paths.
Since different people process and remember information uniquely, an AI solution helps personalize the learning experience regardless of the employee’s background. This includes educational level, industry experience, job role, technical proficiency, and even language skills.
The system ensures comprehensive test coverage with automated answer generation, even without expert-prepared responses. Contextual matching evaluates meaning rather than keywords and adaptive learning provides personalized resources for improvement. Real-time difficulty adjustments optimize assessments, enabling precise knowledge evaluation and supporting data-driven decisions for employee development and promotions.
The client significantly reduced training costs by leveraging internal talent, identifying employee potential, and fostering long-term workforce growth – achieving up to 2x ROI on training investments. AI-driven insights accelerated decision-making, providing a comprehensive workforce skills overview while reducing reliance on external hiring by up to 40%.
Employees now spend 30% less time on training and assessments while gaining confidence through personalized learning paths and adaptive evaluations. Immediate feedback and targeted learning resources ensured continuous professional development.
The L&D department improved employee retention rates by 25%, enabling career development planning through a role-based system that allowed non-technical users to access critical insights without IT support. Workforce management became more efficient, saving time by eliminating the need to create, check, and adapt tests and surveys, while also reducing time spent on FAQs, support, and content creation