What challenges did HAYAT HOLDING face before AI implementation?

Quality was previously determined by laboratory tests with delays of up to several hours, slowing operator feedback. Glue usage needed to be optimized because excessive adhesive increases cost and waste while insufficient glue causes quality defects. The existing know-how for setting adhesive amounts was empirical and depended on experienced operators.

What was the impact of the AI solution?

The article states the laboratory results equate to savings of $300,000 annually. The solution reduces unnecessary chemical waste and carbon footprint. It provides a faster feedback loop to plant operators. The deployment is intended to expand to additional wood panel plants.

Source

Evidence: Low25/100

HAYAT HOLDING builds predictive quality and glue recommendation using Amazon SageMaker (plus edge inference)

Use case typeManufacturing quality monitoringUpdated Jun 13, 2026

HAYAT HOLDING and its KEAS wood-based panel operation built an end-to-end machine learning pipeline to predict MDF panel quality and recommend optimal adhesive usage. Plant sensor and process data are ingested from the production line, batch-aligned features are engineered, and SageMaker models predict multiple quality parameters as well as the minimum glue amount needed to meet quality thresholds. The solution supports near-real-time operator dashboards and edge inference for production use.

Organization: HAYAT HOLDING
Industry: Manufacturing
Location: Turkey
Published: March 2023

Reported outcomes

Strategic outcomes

New product / capabilityBuilt predictive quality and glue recommendationSpeed & agilityProvided near-real-time operator feedbackSustainability & ESGReduced chemical waste and carbon footprintMarket & geographic expansionPlanned expansion to additional plants

Why do we believe this deployment?Customer identity, provider attribution, maturity, and source checks

Customer: HAYAT HOLDING, KEAS
Provider: AWS
Maturity: Unknown
Linked source: AWS Machine Learning Blog

No explicit deployment-stage evidence found.

Customer identity supportedSource describes one deploymentMaturity evidence evaluated

Primary read

Use case focus

Showing 3 of 3

1Predictive quality management
2Manufacturing process optimization
3Edge inference

Quality was previously determined by laboratory tests with delays of up to several hours, slowing operator feedback.
Glue usage needed to be optimized because excessive adhesive increases cost and waste while insufficient glue causes quality defects.
The existing know-how for setting adhesive amounts was empirical and depended on experienced operators.

HAYAT HOLDING built an end-to-end pipeline with AWS Prototyping specialists and AWS Partner Deloitte.
Sensor data is streamed from the plant through AWS IoT SiteWise Edge Gateway and AWS IoT Greengrass.
The team prepared batch-aligned features and used Amazon SageMaker model training, automatic model tuning, and model deployment to build quality prediction models.
A second model recommends the minimum glue amount that satisfies required quality thresholds.
Amazon SageMaker multi-model endpoints and Amazon SageMaker Edge Manager support scalable deployment and edge inference on AWS IoT Greengrass devices.
Operator dashboards present product quality predictions and adhesive recommendations in near real time.

Technologies

Amazon SageMaker Amazon SageMaker Automatic Model Tuning Amazon SageMaker Multi-Model Endpoints Amazon SageMaker Edge Manager AWS IoT Greengrass AWS IoT SiteWise Edge Gateway

The article states the laboratory results equate to savings of $300,000 annually.
The solution reduces unnecessary chemical waste and carbon footprint.
It provides a faster feedback loop to plant operators.
The deployment is intended to expand to additional wood panel plants.

Architecture

Data is streamed from the plant via OPC-UA through AWS IoT SiteWise Edge Gateway and AWS IoT Greengrass. After batch segmentation and feature engineering, models are trained in Amazon SageMaker, tuned with Automatic Model Tuning, and deployed through real-time and multi-model endpoints. For production, SageMaker Edge Manager runs predictions on AWS IoT Greengrass edge devices, and operators consume near-real-time dashboards.

Implementation partners1

Deloitte

Sources & evidence1

Evidence: Low25/100Evidence strength