United StatesHealthcare and life sciences2021 – ongoingMerck & Co. (MSD)
HawkAVI analysis of rejected drug-product inspection images
Merck reported a 50% reduction in false rejects across several product lines using HawkAVI
Problem
Automated visual inspection rejected usable medicines because process variations could resemble defects. Specialists had to assemble information from separate systems to investigate the causes.
Approach
Merck built HawkAVI on AWS to combine inspection images with manufacturing data. AI and machine learning helped identify rejection patterns and recommend corrective actions. Manufacturing teams used the findings to adjust inspection rules and investigate problems.
Results
Takeaway
The reported reduction came from analyzing inspection failures and correcting their causes. HawkAVI supported manufacturing teams' decisions about inspection settings.
Categories
- Size
- Large enterprise
- Industry
- Healthcare and life sciences
- Business function
- Production and quality
- Innovation type
- Process innovation
- AI technology
- Computer visionPredictive analytics and ML
- Adoption
- Built in-house
- Evidence
- Company's own claim
Sources
Source summaries
Merck employees described HawkAVI in a guest contribution published by AWS. They reported a 50% reduction in false rejects across various product lines and explained how image analysis supported corrective manufacturing actions.
Full details are in the original sources.