Quality control is one of the most critical steps to provide guarantees of patient safety to both regulators, internal QA managers and, of course, the patients themselves. In most pharmaceutical companies, quality control is a highly manual process, which not only introduces the risk of human error but also drives up costs and, ultimately, the prices that patients or their insurers have to pay.

Quality control covers a plethora of different tasks, including environmental control, visual inspection of products, leak testing for liquid products, disintegration time and dissolution testing for tablets, and many, many more. The common denominator of all these tasks is that there is an input (an image, a graph, or a time series) and a desired output with some tolerances. As it turns out, models based on Artificial Intelligence excel in solving such tasks and, therefore, there is enormous potential in automating manual quality control tasks with AI.

Example: Visual inspection of products

One of the tasks that AI can be applied to right away is visual inspection of products. As seen in the video below, inspection is carried out manually in most pharmaceutical companies. Some companies have explored the use of automated visual inspection systems but even these can be significantly improved, as they are often based on traditional rules-based computer vision algorithms that are far inferior to modern models based on AI.

Criterion AI’s model zoo hosts multiple AI models for automatically inspecting both liquid and tablet products. By using images taken of the products under the same lighting conditions as the current manual inspectors or the existing automated visual inspection equipment, AI models can be trained to drastically improve detection rates of defects (usually to over 99.9%) while reducing false rejection rates to the order of 0.1%.

The same story can be told for many of the other processes in quality control and, therefore, we urge you to create a free account, upload some images (or other types of data) from your quality control process, and train one of the suitable models in our model zoo to see what sort of performance you can expect. If your case is solvable by AI, there multiple benefits to be had.

With Criterion AI you can

  • Train AI models to automate quality control processes
  • Significantly shorten the time to conduct QC tasks
  • Drastically improve detection rates of:
    • Environmental issues,
    • Nonconformities, and
    • Product defects
  • Lower false rejection rates and boost gross margins
  • Simplify manual processes and SOPs around QC

If you would like to learn more about how to leverage the power of AI in quality processes, please reach out to us. We would love to talk to you about how AI can help you out.

This website collects usage data to help us improve your experience of the site. Click here to accept or here to opt out.