What challenges did Merck face before AI implementation?

Reduce false rejects in pharmaceutical manufacturing and inspection. Ingest and contextualize near real-time manufacturing and inspection data. Handle limited defect training data for complex defects.

Source

Evidence: Low25/100

Delivering Innovative Health with Generative AI Solutions at Merck

Use case typeQuality inspectionUpdated 3 days ago

Merck uses AWS to solve false rejects in pharmaceutical manufacturing by ingesting and contextualizing near real-time inspection and process data. It also uses generative AI methods to create synthetic defect image data for complex defects where training data is limited.

Organization: Merck
Industry: Pharma
Location: United States
Published: July 2026

Reported outcomes

Strategic outcomes

Other strategic outcomeReduced false rejects in manufacturing inspectionCost efficiencyImproved product availability and yieldBetter decisions & insightEnabled faster investigations and corrective action

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

Customer: Merck
Provider: AWS
Maturity: Unknown
Linked source: AWS customer video

No explicit deployment-stage evidence found.

Customer identity supportedSource describes one deploymentMaturity evidence evaluated

Primary read

Use case focus

Showing 3 of 4

1Quality inspection
2Real-time analytics
3AI model training

Reduce false rejects in pharmaceutical manufacturing and inspection.
Ingest and contextualize near real-time manufacturing and inspection data.
Handle limited defect training data for complex defects.

Merck uses AWS Glue and Amazon Kinesis to ingest, transform, and contextualize near real-time data.
The harmonized data is loaded into Amazon Redshift and used by analytics dashboards including Amazon QuickSight.
Merck's AI/ML platform is built on Amazon SageMaker and uses AWS DataSync to ingest defect image data across sites.
For complex defects, Merck uses generative AI approaches such as GANs and Variational Autoencoders to create synthetic defect image data.

Technologies

Amazon SageMaker AWS Glue Amazon Kinesis Amazon Redshift AWS DataSync Amazon QuickSight

Through the use of AWS services, Merck has improved product availability, increased product yield, enabled rapid response to investigations, and enabled rapid Root Cause Analysis and corrective actions all while providing significant time and cost savings.

Architecture

Real-time manufacturing and inspection data is ingested with AWS Glue and Amazon Kinesis, harmonized and loaded into Amazon Redshift, visualized in Amazon QuickSight, and paired with a SageMaker-based AI/ML platform that uses AWS DataSync to move defect image data across sites; synthetic defect images are generated with GANs and Variational Autoencoders.

Sources & evidence1

Evidence: Low25/100Evidence strength