Hamilton AI Strategy Advisors helps businesses build and implement artificial intelligence strategies while navigating a rapidly evolving AI landscape. Founder Brooks Hamilton believes that while there are risks associated with any new technology, AI will transform business.
Ian Heller and Jonathan Bein, Ph.D., talked with Brooks Hamilton about how distributors can achieve greater efficiency and effectiveness through artificial intelligence and where they can begin. According to Hamilton, the key is to “understand where humans act like machines. That’s where the opportunities lie.”
Jonathan Bein: Let’s talk about different levels of AI. For instance, AI as a feature, task-based AI or using AI to manage rebates or cross-sell, generative AI or general-purpose AI. How do people get started?
Brooks Hamilton: I like that framework, Jonathan, because it helps differentiate the applications we see and how much we will need to introduce into our organizations as well as associated fears. For feature-level AI, there are great tools to summarize meetings or make speech-to-text. For task-specific technologies, I’ve seen optimization used for pricing or more general-purpose algorithms such as clustering used to identify similar types of customers and purchasing patterns. These are quite different from the techniques we have seen rolled out with ChatGPT. We ask which one is real AI. The answer is all of them. It is data automation, where we ask a computer to make a decision and help our cognitive process.
Ian Heller: What are the three most successful applications of AI in distribution?
Hamilton: When I think about distribution, it’s a fascinating area because it is not just about how well you can execute the task but how well you can organize the data around these tasks. As you become larger and larger as an organization, the opportunity for more organization and more efficiency takes place. But similarly, the amount of complexity becomes a lot higher. Some of the more impressive areas where I’ve seen general AI is in route planning to ensure you are using one of your most expensive combined resources, including trucking and human capital. The second is being able to pull the pricing lever, and then the third is warehouse optimization. I’m going to throw a fourth in there, which is demand forecasting, to understand how to use your working capital best.
Heller: What about customer-facing tools like CRM and marketing automation? Are you seeing some innovations that are making a difference?
Hamilton: I am just beginning to see these come out on the generative side of AI. One is in some of the improvements that Salesforce has made as they put together their auto-logging for opportunities. The other is opportunity identification I’ve seen from a few vendors. It’s great to combine with other marketing automation to drive some of those sales improvements.
Bein: I like that you mentioned route optimization. It’s one of the things we talked about at the conference. The literature on optimization is vast and spans from AI to what’s known as operations research. One of the things that is hard is the number of choices you have. If you have a driver that has to go to 10 different stops, there are about 3.6 million different possibilities.
Hamilton: Agreed. I think about some of the work that’s been done on route planning or how to solve problems with a large solve space that can range from four answers to 4 billion answers. In the cases when you have high complexity, you likely have a lot of choices. If you remember the AlphaGo tournament with Lee Sedol years ago, one of the approaches they used was instead of trying to work through the large solve space, they had the application develop intuition. They got the application to play many games and develop intuition, kind of a Monte Carlo-type approach for figuring out the best paths. Then, they let it rip. There are a lot of business applications that have so many choices. I don’t make choices based on data 100% of the time. Instead, I use my intuition as a way of navigating to the right choice, even if it’s not a perfect choice.
Heller: Let’s bring this back to distributors. How do you explain the difference between AI and advanced technology? Why are we using the term artificial intelligence and not next-level technology?
Hamilton: Historically, what AI has meant is the thing that’s not here yet. Whenever we master a specific technology, we just call it technology. We don’t call it Roomba’s House AI; it’s just a Roomba. We’re calling it AI instead of fancy chatbot this time around because the advances have been so shocking to have an object that can interact with us just as a person would. At least in that it can definitively answer the touring test of proving that it can behave much like a person does in terms of speech and interaction.
Heller: What is the role of your technology vendors to figure out where you can apply AI?
Hamilton: We’re all choosing the parts of the economy we want to specialize in. As a distributor, we want to be excellent in all of the tasks that distributors are great at and should be performing. Similarly, both software and hardware vendors should spend as much of their hard-earned cash as possible to become great software companies. They should be looking out into the future and figuring out how they can either solve their customer’s existing problems or solve problems that their customers have wanted to solve but couldn’t because the tools didn’t exist or were too expensive. For industries working on thinner margins, this is a time to look to your vendors to step up and provide strong thought leadership on how AI can be used to solve problems that either you see today or are coming.
One thing I would note is that McKinsey did a study looking into AI leaders versus those not prioritizing AI. They found that those who were not leading were looking at efficiency plays. Those who were early adopters or more innovative adopters were trying to figure out how they can build new business or solve a problem. How can I be more valuable to my end customer or how can I create a new line of business to grow my top line?
Bein: So, they are looking at effectiveness, not just efficiency.
Hamilton: That’s right. How can I improve the overall value proposition of my organization?
Bein: In distribution, what might that look like? Is it going to mean better customer experience?
Hamilton: Some of this reminds me of another data automation approach. There have been a few ways technology has come through the general and industrial economies, including ecommerce. I would look at AI as ecommerce in terms of impact, but it will touch more portions of your organization such as sales, HR, operations and administrative. There was a moment when ecommerce solutions came out, and retail shoppers were like, “Hey, when I check out where are the cross-sell and upsell recommendations, how do I make sure that this works just as seamlessly as a retail shop or a retail ecommerce experience?”
That’s where the bar is set. We’re going to see something similar play out in the AI realm as we have in the ecommerce realm. As end-users and consumers, we’ll become accustomed to more and more AI experiences that make it easier for us to shop. Organizations are familiar with what we want to buy when we want to buy it.
I should be able to have any number of questions answered immediately. I should know how to use a product and have very clear ideas about how I’ll be serviced throughout the lifecycle of that product. As consumers, they will expect the same criteria will be applied to the business realm. What we see coming forth on the consumer side, like ChatGPT is available to anybody on the planet, the whole set of capabilities will roll out and set expectations for how distributors must use it.