Artificial intelligence isn't yet perfect. The technology has led to some major business flubs (like that time Microsoft's AI chatbot "learned" to be racist, sexist and anti-Semetic in 2016). But AI is not going away in the supply chain. In fact, it's prevalence is expected to grow.

In the 2021 MHI Annual Industry Report, 17% of respondents said they use AI already, and another 45% predicted they'll use it in five years. The survey of more than 1,000 supply chain professionals worldwide also found that 25% plan to make investments in AI products in the next three years.

"[AI] is very complex, but what we can use is very, very simple. People don't need to have a deep understanding of the algorithms," said Ben Lynch, director of business data analytics at DHL Supply Chain. "I never thought it would be able to move as quickly as it has, and it's only going to get better."

Growing use cases

In the last five years, AI has shifted the relationship DHL has with its customers.

The company went from giving customers information for something that's already happened to "something that's a bit more predictive," Lynch said. "Through AI, machine learning and data availability, now we can give them insights into not just what's happened but what's going to happen."

That transition has been fueled by sophisticated algorithms that can handle the sheer amount of data collected.

"Every two years, we are generating as much data that has ever been created. By 2023, we'll have twice as much data in the world. Because of this, there's been a big need for technology to help support this data," Lynch said.

AI is also enabling the advancement of other kinds of technologies in the supply chain, like robotics, said Thomas Evans, robotics chief technology officer at Honeywell.

"The complexity and the way-quicker access to AI through third party providers, and also the ability to build AI platforms and deploy them, is a drastic change and asset to supply chain logistics," Evans said. "It's only going to get more advanced as we harness more and more data."

AI isn't a panacea for business problems, though. It has experienced growing pains.

More recently than Microsoft's chatbot ordeal, the online real estate company Zillow shutdown Zillow Offers, an AI-fueled home buying and flipper service, because the company bought houses for higher than they could resell. The company took a $304 million inventory write down in the third quarter of this 2021.

But in the right use cases, AI is sophisticated, effective and already paying off.

A McKinsey report found that, for early adopters, AI-enabled supply chain management improved logistics costs by 15%, inventory levels by 35%, and service levels by 65%, compared to "slower-moving competitors."

The biggest gap Lynch sees to further adoption is simplifying data and "getting data and business to speak the same language," he said.

"As the amount of technology and digitalization increased and opened up more and more data, the struggle we've had is, how do you take this data and transform it into something that makes more sense for the operator?" he posed.

That will help drive decision-making on the operator end.

AI makes data more valuable, for better or for worse

Lynch doesn't expect the flow of data to stop, either. He said he thinks it will grow along with the booming e-commerce landscape, fueled by all the consumer data drawn from that online activity.

"Now that we can understand what a consumer is going to buy next week, how do we set up our warehouses?" he said. "The technology get better and better and we'll get a really good sense at the consumer level as opposed to the aggregate supply chain level."

Wider roll out of AI will also depend on robust cybersecurity, Evans said. If data can't be captured, stored and shared with company partners securely, it becomes a liability instead of a boost.

"That's why we're seeing some of the vulnerabilities and ransomware concerns bigger businesses are having," he said, adding that, right now, he's working on a commercial product release and said that 80% to 90% of his engineers are focused on product security.

As the value of data to run AI systems become more valuable to businesses, they become valuable to bad actors, too. "Over time it becomes more of a target," Evans said.