Celonis conducted a survey of business leaders, all from large companies, some from manufacturing, and found that 72% of respondents say that process issues are getting in the way of implementing AI initiatives. 84% of the respondents were actively working to bring AI into their companies, but 72% of those surveyed were having trouble. What kind of trouble? Process trouble.
Now, Celonis is a process mining company. They produce software that tracks and analyzes processes so those processes can be optimized. It might seem as though Celonis is a hammer that sees everything as a nail, as the saying goes. Still, since a business process is any structured set of tasks leading to a chosen outcome, manufacturing businesses are going to have a lot of nails.
Data mangement
Data quality and quantity is likely to be an issue. AI models rely heavily on high-quality, well-structured data to learn and function effectively. Many manufacturers struggle with fragmented data across different systems, inconsistencies, and inaccuracies, hindering effective AI training and deployment.
Data silos among departments can add to the confusion. Capturing data directly from machinery seems like a great idea, but it’s easy to end up with multiple firehoses of information without any obvious way to synthesize and analyze the data for actual use.
The Skill Gap
Implementing and maintaining AI in manufacturing requires expertise in data science, AI engineering, and domain knowledge. Many manufacturers lack the in-house talent or struggle to attract and retain skilled professionals in these areas.
At the same time, workers may resist introductions of AI, whether from ignorance and suspicion or from realistic worry about losing their jobs to machinery that they’ve trained.
IT challenges
Integrating AI solutions with existing manufacturing systems and workflows can be complex and require significant effort and resources. Incompatibility within IT systems can slow down adoption of new technology.
Strategic limitations
Without a defined vision and specific goals for AI implementation, manufacturers might end up with solutions that don’t address their core challenges or deliver tangible value. Often, leaders may react to the idea of AI with, “We’ve got to get that!” without a clear idea of what AI can do for their facilities or how to use it.
Focusing solely on short-term benefits or cost savings can lead to missed opportunities for long-term strategic transformation through AI. On the other hand, AI in manufacturing is still in its infancy. It’s hard to envision future uses.
Ethical quandaries
Bias and fairness are already visible issues for AI: AI models trained on biased data can perpetuate existing biases in decision-making and product development, raising ethical concerns. Questions of copyright, intellectual property, and fair compensation have already tripped up many industries in their use of AI. Manufacturing hasn’t yet hit the headlines in these areas, but it’s only a matter of time before we see how these issues will affect our factories and warehouses.
Transparency and explainability come up in questions of data privacy, hiring, and worries about data misuse. Understanding how AI models arrive at certain decisions is crucial for building trust and ensuring responsible use within the manufacturing process. Again, many companies may not have the required knowledge and experience in-houses.
In manufacturing as well as in other fields, AI comes with promises and challenges. By overcoming these challenges, manufacturers can unlock the true potential of AI and transform their operations for a competitive edge in the digital age.