Is AI A Supply Chain Savior Or Silicon Valley Snake Oil?

I recently sat on a panel at Manifest 2024 with colleagues from Sony, USC, Mattel MAT , and Forum Ventures. It was interesting to hear what supply chain leaders are talking about in regards to artificial intelligence, but even more fascinating to hear their thoughts on what the future holds for supply chains and manufacturing.

The reality is that when I talk to leadership at manufacturing companies, AI isn’t top of mind. In fact, according to panelist Nick Vyas from USC, he found other priorities (like forthcoming regulations around sustainability) mattered more than AI to those he spoke with in India and Thailand. I’ve found the same.

Don’t get me wrong, AI is key to building the connected manufacturing ecosystem with predictive insights and streamlined processes that we need. But it’s one piece of the puzzle, and without humans in the loop and real-life applications, it’s just noise.

What is top of mind right now is tight budgets and supply chain instability. The expectation of business expansion has never been higher, so supply chain leaders need to do more with less. And it’s clear these pressures are ongoing.

Diversify and De-Risk Supply Chains

Depending on who you ask, the solution to supply chain instability is onshoring, offshoring, or nearshoring—you name it. They’re all reasonable responses to supply chain volatility, but the case for regional diversification has never been stronger.

I’d argue that every supply chain is at risk in today’s world, but globalization isn’t necessarily the problem, so it’s not clear that moving everything to the US is the best solution for US companies. At Fictiv, we’ve moved into four global regions as protection against some of these risks, and I know other companies (like Tesla TSLA and Apple AAPL ) are doing the same.

Building strong relationships with suppliers in regionally diverse networks pays dividends. When one region is impacted by instability, volatility, or extreme weather, we have more opportunities in our back pocket to get through it and offset some, but not all, of the issues in an unstable world economy.

According to fellow panelist Maia Benson from Forum Ventures, diversifying sourcing makes sense. When manufacturers spread their risk, they can easily pivot if things are breaking down in their network. This helps build resiliency.

Increase Operational Efficiency

Another aspect of building smarter, more resilient supply chains is understanding workflows and automating where you can, which brings me back to supply chain leaders having to do more with less. My approach as a technologist is all about how to apply the best type of automation, and when we define automation, we think about humans plus machines. I think it’s really important to call out the difference here. Everyone's talking about AI, AI, AI, but the real magic in manufacturing is in the interplay between AI and human action.

Here’s where AI can really help: There is a ton of manual work involved in sourcing new suppliers, onboarding suppliers, requesting quotes, reviewing design for manufacturability feedback, requesting order status updates, troubleshooting order issues, and triaging disruptive supply chain events — all of these tasks combined add up to a significant amount of friction and drag in the sourcing and manufacturing process that has a real impact on the bottom line.

AI technology today can automate many of these manual tasks to dramatically accelerate and streamline the procurement process. For example, AI algorithms can instantly price a part for manufacturing, saving days or weeks on quoting today. Similarly, AI can analyze the manufacturability of a part, streamlining the communication between a supplier and an engineer. AI-powered predictive analytics can also help with demand forecasting, inventory management and supplier risk management as well. But again, AI is only as good as the data and the humans in the loop.

Buy Versus Build

When it comes to AI, a question I often hear is, “Should we buy an AI solution or build one ourselves?” The advice I always give is if AI is going to differentiate you in the market and be a core competency of your business, go ahead and build. If not, if it’s not something that is going to be a core competency, a differentiator, or generate revenue, it’s probably wise to look to other, existing solutions.

As an analogy, if you look at all of the parts that are on the bill of materials for the product you're building, which of them are critical to your company differentiation? In these cases of buy versus build, ask yourself this: "Do I need to own this and its core strategy? Or is this something that is part of another company's core strategy?"

Greg Javor from Mattel agrees. On a recent trip to Asia, they looked at off-the-shelf solutions to automate shop floor scheduling, some powered by AI. These are plants with 5,000+ people, so scheduling is a good place to automate. Mattel looked at these ready-made solutions but found them head-scratching in their complexity. Ultimately they weren’t going to work for Mattel, so they decided to build their own.

The Global Supply Chain of the Future

Looking ahead, I see a future where everything in the supply chain is interconnected. And more focused on traceability, sustainability, and how companies like ours are accountable—not just to regulatory bodies but to ourselves. Our core belief is that we want to harness technology to unlock the creative potential of humans. Technology and AI are brilliant accelerators, but they work in concert with that human potential.

Maia Benson predicts a real-time supply chain with optional human decision-making at core inflection points. For Greg Javor, it’s an integrated end-to-end supply chain that is increasingly direct-to-consumer. Mark Hanson believes the future supply chain is a continual data flow between the start of the supply chain and the end.

Cut through the noise, and the core truths for suppliers and manufacturers remain: Run leaner, be more agile, and improve profitability to thrive in this economic climate.