- Top Distributors Lists
- Market Research
- Free Reports
MDM webinar highlights the growing role of AI in distributors’ business operations and why it’s imperative to develop a strategy that considers competitors’ capabilities and customer expectations.
The world of artificial intelligence is evolving, as previously siloed sectors integrate to create powerful capabilities for efficiently managing and operating your business. Still, confusion about the broad concepts of AI and how they interrelate is holding many distribution leaders back from taking advantage of the technology’s potential.
In a recent webinar, “How Artificial Intelligence Disrupts Distribution,” Ian Heller, COO and president of MDM, and Tony Corley, senior product marketing manager at Epicor, spoke with Tom Gale, CEO of MDM, about what AI really means and practical implications of its use in wholesale distribution.
Heller began by outlining various concepts developed by Michael Wu, Ph.D., chief AI strategist at PROS, including his definition of four types of AI:
1) Perceptual AI. Where a machine perceives an outside input, such as an image or voice command, and uses that information to make decisions and act. Always human and computer interactions, examples include ordering by voice (such as Amazon’s Alexa) or by camera and interacting with a chatbot on a website.
2) Internet AI. Recommender systems on websites track our interests and preferences and make product suggestions based off of them. This is why we each see a unique homepage when we log onto Amazon or other sites. A well-run business website, Heller said, will show visitors items specific to what they’ve purchased and browsed in the past.
3) Business AI. Intelligence that makes business decisions, such as automatically tracking inventory and replenishment needs. It can also provide fraud detection and make fast decisions on processes such as stock trading and even medical diagnoses.
4) Autonomous AI. The process by which machines act like human beings, doing tasks such as robotic warehousing and autonomous truck driving. Operating by themselves, they react to outside stimuli, making decisions without direct human control.
All four forms of AI are fueled by big data, which is defined by four characteristics outlined by authors Viktor Mayer-Schönberger and Kenneth Cukier, Heller explained, summarizing the description in their 2014 book, “Big Data: A Revolution That Will Transform How We Live, Work, and Think.”
- First, analysts can now take in entire, enormous datasets because today’s computer processes and databases can handle them. They no longer have to settle for representative samples of larger datasets.
- Second, the size of these datasets allows for some “messiness” within the data, as Heller put it, so that it does not need to be pristine to still be usable. The sheer size of the dataset tends to offset any errors present in it.
- Third, it’s often the case that unintended uses of the data prove more valuable than its original intention. For example, Heller said, traffic reporting on apps such as Waze or Google Maps evolved from the ability to track the rate at which mobile phones inside moving vehicles switch between cell phone towers.
- Fourth, the data is useful to identify correlations — without assuming causality. “Sometimes the correlations can be compelling, and yet, at the same time, spurious,” Heller said. It takes judgement and evaluation to determine cause and effect.
Another concept associated with AI is machine learning. Although many algorithms use big data analytics to drive actions, they don’t necessarily get smarter over time, Heller said. AI-based models incorporate machine learning, which enables them to improve on their own with time and experience, leading to better analytics, better recommendations, smarter actions and improved outcomes, he added.
Applied to the distribution industry, machine learning can improve inventory performance by monitoring not only sales history, but also other factors such as weather information, economic