Ian Heller of Distribution Strategy Group spoke to two industry leaders to find out how these companies are venturing into the new world of AI.
In “How Distributors are Building AI Capabilities,” these experts shared their experiences and the lessons they have learned by applying AI to their businesses:
- Mike Mortensen, President and CEO of ARG Industrial
- Steve Cosgrove, Founder and President, Cosgrove Partners
Did you miss the full webinar? You can watch it for free here.
AI in Distribution: It’s Complicated
Distribution is complicated in a way that retail is not. The products and configurations are more complex, and the volume of data shifting between third parties, your business and customers, is voluminous. Where does AI fit in? Is the hype real — can AI solve those previously unsolvable problems and transform the industry for the better?
Mortensen shared how custom orders are now at the point of sale. Customers approach sales teams without part numbers – just a broken item from the field and a simple request, “Do you have this?”
That’s why it takes so long to get new customer service reps (CSRs) up to speed. Mortensen believes AI is the perfect technology to untangle these complexities. “We see AI taking some of those data points to help a new associate conduct application engineering where they’re saying, okay, we’ve got this problem. We need to get fluid or gas from here to here. What are the appropriate components that all go together to make that magic happen?”
Mortensen’s company has been working in this area since 2019, recently launching better customer self-service tools online. “I think AI will democratize information dissemination within distribution, so we’re able to access the tribal knowledge of the team to provide custom solutions to our customers.”
Cosgrove applies AI for predictive analytics in and around the supply chain. “When we went through COVID, there were massive parts shortages. Some of the supply chain challenges we were facing were almost insurmountable. But you can’t really be short a 90-degree elbow that will prevent an $800,000 house from being built. So, we built a lot of dashboards; we’d built massive data lakes so we could go in and begin to archive information. We built much more rapid cross-referencing capabilities, so we were able to start to think differently about our product information. We came up with new solutions.”
But Cosgrove noted the capital challenges often associated with AI applications. “It’s not an inexpensive venture. With that said, I think you get the biggest ROI on this spend across your business. We’re using archived data from the various businesses we’re in to train so much more efficiently. It’s hard to find a good counter person these days. If you have technical questions coming to your counter person and haven’t figured out a way to bring that information and make it available on a few keystrokes, you’re in trouble. It’s a provocative time for distribution because we are being asked to take some chances (on these new technologies).”
Mortensen said at first they felt “way behind.” “And then we started working with manufacturers and realized we’re all in the same boat trying to solve some of the same problems at the same time. Going forward, the potential for distributors and manufacturers to partner around technology to make information available at the point of sale is huge. I see AI being a part of that.”
Cosgrove pointed out that comparing your AI journey to a behemoth like Amazon can make a distributor feel insecure.
“You go on Amazon, and it already knows the stuff you’re going to buy, it suggests that your shirt size, and somehow it knows that I gained four pounds. You’re just blown away. It’s like how in the world can you know that? If you’re running the average distribution company, you’re like, gosh, how can I do that for my customer? Well, realize that they’ve been investing in this for decades and their data structure, the way they think about it is fundamentally different.”
AI Build vs. Buy: Should You DIY?
Three years ago, Mortensen’s company started identifying external and internal experts to help them customize AI tools and models. Internally, they identified employees who gradually became subject matter experts to build out these tools. “We’re not spending hundreds of thousands of dollars externally,” he said. “We’re investing in these people now building technologies and workflows that are AI-enabled, and the return on investment is huge. It’s not just the product and the process we’re building; we’re also developing internal knowledge on how to do this.”
Cosgrove said most companies need some help on the front end. “You don’t want to make bad investments. There are so many sham software solutions and they’re selling at a level that’s hard to believe. Most people, say 99%, do not really have AI in their business. They talk like they do. They say that they’ve accomplished these things. The fact is there aren’t even enough chips manufactured to power AI at the scale that people say they have.”
Organizations need help separating the hype and applying AI reality.
Part of the challenge is the data behind distribution transactions isn’t consistent. “The manufacturer has a part number, and the distributor has an iteration of that part number, and then the customer has a third iteration of that part number,” Cosgrove said.
But Mortensen views these complexities as perfect for the application of AI. “The part number game is really confusing, but I think that AI will have a real impact on product attribution. It will get the industry away from part numbers and more toward specifications like, ‘Show me a two-inch black iron threaded elbow’ and let the customer decide which one’s available and fits their price and specification needs.”
Where Are Distributors Applying AI Right Now?
Mortensen describes how his company uses AI today:
“We have 13 locations with material hitting our docks every day with packing slips and paperwork attached. I have people checking that the material is correct and then they’re going to a computer and jamming all this stuff in and doing tons of data entry. We’re looking at technology that reads those documents, interfaces that data right into our technology, with queue functionality where it matches with manufacturer confirmations before we even get an invoice.”
Cosgrove’s company is using AI with clients in customer service and the order-to-cash cycle. “We want to augment your decision-making process, so we’ll stand up a dashboard system of cockpit charts to give us different visibility when looking at P&L or a balance sheet. You can look at 100 customers, take a product line, and ask, ‘What’s my penetration of X, Y, Z?’ It’s an extremely useful tool that empowers sales reps to step outside the comfort zone of the products they’re most familiar with to sell more.”
When Sharing is Caring: Distributors on Data Sharing
Machine learning algorithms in AI take human interactions and improve their responses based on these behaviors. The more data you feed it, the better it becomes. However, distributors worry about oversharing because of the value of the proprietary customer data in their systems.
“Well, I would not suggest people do not protect themselves,” Mortensen said. “I would suggest having a legal framework around information that you’re sharing.” He suggests that the propensity to hold data close to the vest begins even before distribution at the manufacturing level.
“We had manufacturers that didn’t want to participate in our PIM initiative and work with a third-party syndication company to provide flat file data on product information so that they could give it to us.” Instead, he suggested, “There’s so much more value than just a product at a price. If you’re not, there’s some internal introspection that needs to happen on where you’re providing value. I think the value of information is in sharing it. That’s where the power happens, at the point that it provides value.”
Cosgrove agreed. “If you’re not willing to share data or do some particular services to create incremental value for your customer, you’re on the endangered species list. You just don’t know it. I get very frustrated with people who hoard information. I sit on the boards of six manufacturing companies. Some of those have enormous direct ship and operations. So, the distributor gets an order, sends it to the manufacturer, and the manufacturer ships the order to the customer. Some distributors will only allow you to ship to them. It seems like they’re the ones that are always yelling at you, right?
“Because they can’t get something fast enough. But you could usurp three or four days of shipping time by allowing the manufacturer to ship directly on your behalf. The mindset is dated. What we’re talking about is how do you use technology to make our people more efficient.”
Ready for the full webinar? You can watch it for free here.
