AI Makes Slow but Sure Progress in ManufacturingAI Makes Slow but Sure Progress in Manufacturing

Honeywell has published the Industrial AI Insights report that quizzed industrial leaders on AI applications. Overall, the leaders see AI investment and implementation as an essential part of their growth plans. The data also finds that leaders see productivity and efficiency gains as the biggest benefits of industrial AI.

The research finds AI leaders are “sold” on AI for industrial applications. The enthusiasm for a deeper commitment to AI investment is nearly universal, with 94% of those surveyed saying they have plans to expand their utilization of AI.

Leaders in the manufacturing sector are serious about AI with 15% already paying more for AI engineers and two-thirds (63%) saying they are very likely to in the future.

However, only 14% of respondents say they have fully launched their AI rollout plans, with three quarters in the launching (33%) or scaling (42%) phases.

The Industrial Benefits of AI

AI leaders in the manufacturing sector find three key benefits of implementing industrial AI:

The most promising use cases for industrial AI discussed by C-suite leaders include:

Production Tool or Customer Service Tool?

One of the most common uses of AI in manufacturing is in customer service. The use of AI in production applications is less frequent. “It makes good sense that AI is used in the customer service area. There is data to be mined there,” Jason Urso, VP and CTO of Honeywell told Design News. “In the instance where something is going wrong for the customer, AI can search the history and bring up use cases.”

Urso noted that production activities are characteristically different. “AI hasn’t been used much in production because that area is deterministic. What if Google maps said there’s an 80% chance you should take a left here?” said Urso. “The customer-use case is problemistic. In industry production, people have been trying to figure out how to use the problemistic abilities of AI.”

There are areas in production that fit the search-and-retrieve capabilities of AI. “You can use problemistic solutions in production with troubleshooting and maintenance. An expert looks at an issue and sees it’s a bearing problem. What if you had a two-year person who couldn’t tell the problem was a bearing?” said Urso. “What if AI could step in so the two-year person could understand the issue like a 30-year expert? Helping people to perform better where a human decision needs to be made works well with AI.”

All in all, the report suggests the pace of AI adoption in manufacturing will be driven by compelling use cases. AI applications will be measured by improved business performance. New solutions need to demonstrate clear benefits to workforce productivity, safety, and reliability.