AI, data companies likely to benefit as manufacturers integrate AI tech: survey

Companies developing artificial intelligence products and support appear poised to benefit as a wide range of manufacturers begin to incorporate AI into their daily operations.

A joint study by Microsoft (NASDAQ:MSFT) and MIT Technology Review Insights found 35% of the manufacturers surveyed have already put AI use cases into production. While most, 64%, are currently researching or experimenting with AI.

“Many executives that responded to the survey indicate they intend to boost AI spending significantly during the next two years,” the study reads. “Those who haven’t started AI in production are moving gradually.”

The survey sample consisted of 300 senior executives from around the globe working in organizations earnings $100M or more in annual revenue.

Nearly 60% of executives surveyed expect to increase AI spending by 10% or more in engineering and design. Another 43% plan to spend the same on factory operations.

Larger companies move faster to integrate AI

While most major manufacturers plan to integrate AI into their operations at some point, the largest have made the most progress.

Aerospace, automotive and electronics manufacturers are the most likely to already be deploying use cases into production.

Nearly 80% of companies surveyed earning revenue of $10B or more annually are already deploying AI use cases. That figure drops to 38% for companies with revenue ranging from $1B to $10B. It nearly disappears after that, with 2% to 4% of companies earning $100M to $999M deploying use cases.

However, even the majority of the smaller companies surveyed are still researching or experimenting with AI in some capacity.

“Everyone in manufacturing is excited about AI,” said Philippe Rambach, chief AI officer of Schneider Electric. “But relatively few are using AI at scale to transform the way they work.”

Smaller companies point to talent and skill shortages impeding progress with AI. The costs of maintaining and improving AI models also become problematic for manufacturers with tighter budgets.

“While we see limited-impact uses of AI among some producers, there is little evidence of AI-led transformation,” said Ben Armstrong, executive director of MIT’s Industrial Performance Center. “We’ve seen few manufacturers extend the use of AI techniques beyond the front office to production operations.”

Product design tops AI use cases

So far, the predominant AI use cases by manufacturers are in product design, content creation and chatbots.

“Design is increasingly happening in simulated environments, which can greatly reduce cycle time,” said Indranil Sircar, Microsoft’s chief technology officer for manufacturing solutions. “Design engineering is becoming a lot more data-centric, and AI is enabling it through simulation.”

Manufacturers have also spent resources developing AI techniques to improve productivity and efficiency.

How to handle colossal data?

The study finds the toughest challenge for most manufactures in scaling AI is the data, as the industry creates greater amounts than other industries. And much of that data is currently unsuitable for AI models.

To address this, 57% of respondents said they are increasing machine connectivity. The same percentage of respondents also said data quality is the toughest challenge related to integrating AI into operations.

“AI requires a level of data maturity,” the study reads. “Determine how well the organization collects, stores, and processes data, and take concrete steps to redress weaknesses before taking AI use cases into production.”