AI in supply chAIns: navigating the future
September 5, 2024
By John Soderberg | Senior EVP Information Technology
By John Soderberg | Senior EVP Information Technology
|
In April, I had the pleasure of co-presenting the keynote address at the Fastenal Expo in Nashville. The topic was “Navigating the Future of Artificial Intelligence in Supply Chains,” and my co-presenter was Vic Miles, the Americas region business leader for Microsoft. (Microsoft is
a key technology partner in Fastenal’s cloud computing and AI journey.) Vic also hosts the popular podcast Beyond the Tech and is an all-round keen observer of business technology trends. If you had a chance to catch our presentation at the Expo, this article will be a refresher. For everyone else, we’re happy to share what we’re working on and thinking about in the AI space. Note: The primer on Generative AI in the following section is heavily based on thoughts Vic shared at the Expo. He is definitely the expert here. First, how did we get here?It may feel like AI dropped out of the sky fully formed about a year ago, but a long road has led up to this moment. Some of the earliest work was done in the 1950s and 1960s when researchers first began to explore the possibility of creating machines that could "think" like humans. They called them expert systems. The challenge: You had to be an expert to use them, which at the time probably meant seven or eight people in the entire country.
In the 1990s, we moved on to machine learning, which was good at finding and identifying patterns but not so good at drawing conclusions. The next step was deep learning (circa the 2010s), a machine learning technique that uses neural networks to process data and make decisions. Fast forward to Generative AI, which we can define as having two essential elements.
With today’s AI technology, we’re approaching what’s known as “general intelligence,” which refers to a single machine being smarter than any given human on any given topic. Think of something like a spelling bee. It’s no surprise that a machine can spell any word in any language faster and more accurately than any human on the planet. The same machine can just as easily beat a grandmaster at chess. That’s general intelligence. But Generative AI isn’t just about processing information; it’s about creating new content based on the user’s prompts. To produce the new work, it can either reference all available information from the Internet or limit the scope to a repository of curated information. (More on this important distinction in a minute.) Perhaps the most amazing thing about Generative AI is how fast it has taken off. As a point of comparison, it took nine months for there to be 100 million active monthly TikTok users on the planet, an unprecedented rate of adoption. ChatGPT flew past that milestone in a mere two months. The world is still grappling with the risks, protocols, opportunities, and applications. But this much we know: AI is a disruptor. It is a productivity multiplier. And as business leaders, it’s our job to learn how it can assist us and our teams. The key word is “assist.” As always, people make the difference. The vision is to make AI an enabling technology that keeps us focused on service and decision-making, not the minutiae of data input and daily procedures. It’s apt that Microsoft has named its Generative AI solution “Copilot” – because it’s there to help, not to replace. 
We are at the beginning of a fourth industrial revolutionThis may be the most significant technological inflection point of our lifetimes – one that will help define not just the tech industry but how the world works for decades to come.
What we've learned and where we're going nextFastenal was an early adopter of Microsoft M365 Copilot. Our employees (and likely many of your own employees) are using it to digest reports, summarize meetings, create or enhance
content, produce high-quality language translations, assist with coding – the list goes on and on. Along the way, we’re learning the importance of creating effective prompts to obtain our desired results (and discovering it’s more of an art than a science). Importantly, we’re also learning that you can toggle between two modes in Copilot: “web” and “work.”
Okay, we’ve learned about the evolution of Generative AI and touched on some early lessons learned. This brings us to the heart of the matter: How is Fastenal using AI to help you? I’ll answer that in two parts: first by discussing one of the use cases we’ve already put into operation, then by looking at the long game – how we’re capturing the high-quality data needed to power future AI innovations. Harnessing our collective knowledgeOver the last four years or so, we’ve been building an internal knowledge base where content is created, stored, and managed. It began as a repository of pre-written responses (penned by various subject matter experts within our company) to potential questions from Fastenal team members. Employees would interact with this content through an internal chatbot we affectionately refer to as “BLUE.” With 23,000-plus employees (70% of whom work directly with, and field questions from, customers), the content quickly ballooned to 125,000 question-and-answer pairs.
We recently connected the Microsoft AI model to this knowledge base, and the difference has been night and day. The traditional chatbot used a rule-based keyword matching process – it was very literal in matching the (pre-written) responses to the questions asked. The downside: If there wasn’t a direct match, the user didn’t get the help they were asking for. Contrast that with our new AI-powered tool, which is trained on vast troves of content to create the most statistically probable expert response to any given question. (It also provides links to the source content used to generate the response.) Fastenal experts still manage the content repository to ensure trustworthiness. The difference: The new tool is generating richer and more relevant responses, with far fewer non-responses. Ask BLUE: what's in it for you? I mentioned above that around 70% of our 23,000 employees are in front of customers like you every day. A big part of their job is to answer your questions and quickly find solutions. The value of our new BLUE copilot is that it empowers each Fastenal representative to respond faster and more reliably than ever before. In a sense, the collective knowledge of the Blue Team is at their fingertips. Building a strong data foundationOkay, now for the long game: using AI to help identify risks and opportunities in the supply chain. AI has the capability to unearth those kinds of insights, but it can’t do it in a vacuum. To produce useful results, it must consume relevant, comprehensive, and trusted data. I came across a headline in Forbes with an interesting spin on this idea: “AI needs data more than data needs AI.” In other words, while quality data is valuable with or without AI, AI cannot be effective without quality data. Data is the fuel that makes the engine run.
With that as a backdrop, we’re working to build a strong “data foundation” as part of your Fastenal program. This groundwork is made possible by some unique aspects of our business model. First, unlike many large full-line distributors, our growth has been almost entirely organic, not gained through mergers and acquisitions. Today we operate more than 3,400 locations across 25 countries, and all are connected by common systems and shared data. Second, we’re not just a fulfillment company; we’re a supply chain company with a heavy investment in global sourcing, logistics, and local service. As a result, we’re able to capture data and optimize efficiency across the supply chain – from the source, to the last mile, to the point of use. Finally, we’ve invested aggressively in our Fastenal Managed Inventory (FMI) technology program. The combination of mobile apps (used by our local service teams) and embedded point-of-use devices allows us to capture and share a complete data story: what’s coming into your facilities, what you currently have on hand, where each product is stocked, and how it’s being used in specific areas of your business. When you put it all together – global systems, end-to-end supply chain, and point-of-use data capture – you can see why many organizations think of Fastenal as their de facto material master. They rely on us to generate high-quality data around product categories that aren’t well supported by ERP or procurement systems. Again, AI is only as good as the data it consumes. If you have good data around your procurement activities, you can imagine using AI to help rationalize your spend, consolidate redundant SKUs, switch to more cost-effective product alternatives, set scheduled deliveries (based on buying patterns), and identify opportunities to switch certain repetitive needs to a managed inventory program. But let’s take it a step further. If you also have good data around the inventory and consumption habits within your business, AI could help us answer some big-picture questions:
Today, finding those answers for thousands of MRO and OEM products involves a lot of complexity and work. The vision is that AI can simplify the pathway by consuming quality data around not only what you’re buying but also how the products are being managed and used within your business. This floor-level insight, combined with an ability to effect change across the supply chain (i.e., to act on information), is where we believe Fastenal is uniquely positioned to add value. 
"We always overestimate the change that will occur in the next two years and underestimate the change that will occur in the next ten." - Bill Gates
Keep learning and moving forwardAs with any new and disruptive technology, there can be a risk of moving too fast (without the proper processes and structures in place), but there’s also a risk of losing focus on the long-term opportunity. In his 1999 book Business at the Speed of Thought, Bill Gates famously wrote, “We always overestimate the change that will occur in the next two years and underestimate the change that will occur in the next ten.” … In other words, don’t get too caught up in the current hype, but as Mr. Gates phrased it, “Don’t let yourself be lulled into inaction.”
The good newsYou have a committed partner in Fastenal. Today, we’re using AI to augment the knowledge and service we bring to your business. We’re also building a solid data foundation to power tomorrow’s innovations, including several AI use cases currently being developed by our IT team. That said, we’ve always found that the best ideas are sparked by you, our customers. Share your vision with us, tell us the kinds of solutions you’d like to see, and challenge us to be a great partner in your AI journey.
Hopefully this article has provided some inspiration for the future we can build together. Vertical Divider
|
Like what you're seeing here? Subscribe to the Blue Print for FREE and get the magazine sent right to your address.
|
