Scaling AI in Manufacturing Operations

A Practitioners’ Perspective

A perfect fit

If the first industrial revolution was set in motion by steam, then Industry 4.0 is being powered by artificial intelligence. And with its ability to automate, digitize, and optimize, AI is the perfect fit for manufacturing operations, from product development to quality control. A computer vision system, for example, allowed GM to detect 72 instances of component failure, preventing massive downtime (a single minute of which can cost a company of that size up to $20,000) while a machine learning system significantly improved Danone’s demand forecast accuracy (reducing forecast error by 20%, lost sales by 30%, product obsolescence by 30%, and demand planner’s workload by 50%).

Enormous potential across the board

The latest report from the Capgemini Research Institute – Scaling AI in manufacturing operations – shows that intelligent maintenance, together with product quality inspection and demand planning constitute a good starting point for manufacturers to focus their efforts in manufacturing operations. That’s because:

  • They offer clear business value/benefits
  • They are relatively easy to implement
  • There is a ready availability of data and know-how
  • There is a possibility to add features that aid visibility and explainability, for ease of adoption by operational teams.

Focus and scale are critical

To tap into the manifold benefits AI can bring to manufacturing operations, organizations need to move beyond the pilot/proof-of-concept stage and deploy at scale. To these ends, we recommend deploying successful AI prototypes in live engineering environments, investing in laying down a foundation of data & AI systems and talent, and scaling the AI solution across the manufacturing network. AI is the rocket fuel behind Industry 4.0, and its natural fit with manufacturing means that organizations that are able to execute these capabilities are the ones that will really take off into the future.

Subscribe to receive an advance copy of new reports from the Capgemini Research Institute

AI in manufacturing ope...

Dimensione del file: 3,38 MB File type: PDF

Infographic – AI in MfG...

Dimensione del file: 3,27 MB File type: PDF

Sound Bites

Neeraj Tiwari, Director manufacturing JV organization at Fiat Chrysler China

We trained an AI system to detect improper assemblies or missing components, such as small screws that are hard to detect for a human eye. The system is extremely fast and efficient, allowing defective parts to be taken off the main conveyor on a separate line to the rework area where they can be corrected. The process not only saves a lot of quality issues at the end-customer but also loss of valuable production time.

Eugene Kusse, Factory director, Upfield (a spin-off of Unilever)

Working in the food industry, we have a responsibility to ensure that the food we produce is both safe for consumption and meets the toughest quality criteria. We have strict policy and procedures in place to ensure that we avoid any of the risks associated with not meeting those criteria.

Siddharth Verma, Global head and VP – IoT Services, Siemens

In the early days, when the accuracy of the system was low, it predicted a few failures which turned out to be false alarms. At these points, it is important to remind everyone that it is a prediction which has a probability of being right or wrong. As accuracy improved, the system was able to predict many failures in advance and saved a lot of cost and downtime, proving its worth.

Key Takeaways


reduction in lost sales achieved by Danone by using machine learning to predict demand variability


share of use cases implemented in maintenance


of automotive OEMs have delivered AI implementations at scale as of January 2019

About the Capgemini Research Institute

Capgemini Research Institute

Capgemini’s #1 ranked in-house think tank on all things digital