Artificial Intelligence takes real work
March 6, 2026
By John Soderberg | Senior EVP, Information Technology
By John Soderberg | Senior EVP, Information Technology
|
Artificial Intelligence is marketed as being effortless. With a few clicks or conversational prompts, we’re able to get instant answers, unlock insights, and even automate decisions, or at least that’s the vision.
But this article isn’t about how easy AI can be from a user perspective. It’s about how hard it is to lay the groundwork for effective AI systems, particularly in the context of supply chain management. As I’ll explain, the word “effortless” does not apply here. It takes hard work to make AI work. Building strong data foundationsWhen we engage with customers to understand and modernize their supply processes, we often discover environments plagued by messy systems, inconsistent sources, and decades of technical debt (the long-term cost of short-term fixes).
In most cases, the parts in their ERP system are presented in a multitude of ways depending on the consuming facility, reference supplier(s), and the employee(s) who entered the reference data in the system. Yes, that same root part may have five or six different numbers. This is assuming the enterprise is working in a single ERP with an actively managed Material Master. It becomes even more challenging if you’re not. Unfortunately, you can’t hit a reset button on years of redundancies, inconsistencies, omissions, and errors. Teams of people have to “get in the trenches” to collect, clean, structure, validate, and govern the data. If it's done right, it includes:
None of this is easy, but it's all essential. It's how we lay the foundations for a data-driven supply chain and the AI future. The only question is, who is going to do it? Garbage in, garbage outA while back, I had a contractor come to my house to talk about resurfacing our garage and adding some new shelving. I was excited to get started ... then came the caveats. "We don't remove your stuff from the garage, and we don't clean your floor. You have to clear and prep everything before we start." (Let's just say it wasn't a fun weekend for the Soderberg family.)
I had garage-cleaning flashbacks when meeting with a new customer a few weeks ago. They had recently paid a consultant to come in and work with their part-level data, which led to a greater need to identify a bunch of their parts and assets. Here's the thing though: The consultant didn't do the work. They hired temp workers to come in and do it, then charged the customer for the labor. Now, I'm not judging the temp agency - they never claimed to be supply chain experts. But I'll say this about the relationship between data quality and AI effectiveness: garbage in, garbage out. Having come in after this project was completed, our team had to make numerous data corrections related to the items the customer needed managed. This represented a fraction of the parts the temp agency reviewed and confirmed, so we can only imagine how many discrepancies are still in their system. To put a harsh light on it, they basically paid someone a hefty fee to create bad data and ensure AI won't work. The worst part? It's not a unique story; we see it all the time. Doing what others won't (or can't)It's easy to come in and say, “You should digitally catalog all your spare parts and tools.” But here’s that question again: Who will actually do it? Your employees? A temp agency? Some other “third party to your third party”?
At Fastenal, we do this work every day for our customers. And when I say “we,” I’m talking about our fastener engineers, our lean specialists, our professional project managers, our technology consultants, and our local teams. In a word, experts. Their work includes painstaking projects like checking OEM prints, creating an exhaustive “plan for every part,” building planograms to digitize inventory (with a description, location, and min-max for each part), and not only diagnosing process bottlenecks but also designing, implementing, and managing solutions to bring forward the grounded data required to make AI effective. Which begs the question: Why isn’t this the norm in our industry? Doesn’t supply chain management by definition involve actual work? Yes, there are product distributors that offer some of these services, but they tend to be smaller companies with narrow product expertise, tech portfolios, and geographic reach. For national distributors, the trend has been to disinvest in local resources in favor of centralized clearinghouse operations, automated call centers, and third-party delivery services. (The downside: When it comes to inventory management and other high-touch services, there’s only so much you can do from afar.) Consultants and software providers offer analytics platforms and recommendations but leave you on your own to capture, interpret, and act on the data. Often the findings are constrained by the quality of the data you provide and its correlation to their data rule set. If you opt for a cloud-based solution, the “cookie-cutter” approach becomes even more rigid, with even less support for change management. This brings us back to our central question (but with some added context): Who will do the ground-level work to implement, execute, and continuously improve your data program – not just one time at one site, but on a consistent basis everywhere you operate? Keeping a human in the loopIn the tech world, "human in the loop" means that a person is involved within a defined process or outcome - it's not entirely autonomous. Sometimes AI is assisting the human. Sometimes the human is assisting AI. It's an ever-evolving cycle of improvement.
In the context of supply chain, the "loop" starts by gathering grounded and trusted data (reflecting physical reality in your business). It ends by acting on what the analytics indicate (improving physical reality in your business). Fastenal can help throughout this journey. With a combination of high-touch service and high-tech solutions, we capture trusted data to feed analytics tools. We also provide analytics tools to help you (and us) understand the trends, cost drivers, and opportunities in your business. Most importantly, we're there to take physical action on those insights, whether it's improving processes on the floor, substituting products, adjusting our local service or the upstream supply chain, or adapting the technology itself. It's not just about automating and optimizing the current state. It's about reimagining how things should work vs. how they do work - this is where the real opportunity is. With trusted data, smart analytics, and a great ground game, we'll help you break out the "legacy trap" into a more cost-effective future. As part of our tech strategy, we're empowering your Fastenal service team with AI tools to improve their knowledge, response times, and problem-solving capabilities in the field. Think of it as a virtuous cycle:
... Rinse and repeat. How do you bridge the gap between digital insights and real-world operations?This is the fundamental challenge every company faces in the digital age. We believe the winning paradigm isn't "AI versus people." It's "AI plus people" working in a cycle of continuous improvement. This is why we invest in what truly moves the needle for your business: great people, close to your business, empowered by technology.
When I meet with customers, they often refer to Fastenal as their de facto Material Master. By establishing and managing trusted, grounded, and globally consistent data around industrial supplies, we're helping them gain insights today while preparing for AI tomorrow. But we're more than a data partner. We're the people on the ground managing change, executing daily tasks, and driving results - the "bridge" between data insights and real-world operations. So, if you're ready to move forward in your AI journey, we're ready to roll up our sleeves and get to work. You may also like:Vertical Divider
|
Like what you're seeing here? Subscribe to the Blue Print for FREE and get the magazine sent right to your address.
|
