ChatGPT Swears It Can Optimize Your Supply Chain and Inventory Systems
Since the AI-bot ChatGPT busted onto the scene just a few weeks ago, it has elicited some sharp opinions. Some folks seem to be certain it will revolutionize the world, coming for many of our jobs along the way. Others point out its responses are often difficult to verify or outright inaccurate. And when Microsoft’sMSFT +2% Bing put the technology underlying ChatGPT to use in an invite-only trial period—Microsoft poured $10 billion into ChatGPT-maker OpenAI—the waters got even murkier. Journalists and scholars pointed out the AI engine’s unnerving threats and weird emotional manipulation. One New York TimesNYT +0.7% reporter said Bing fell in love with him.
These are the flashy headlines, but I fear our 21st century culture is doing something it’s inclined to do: Obsessing over something, finding reason for backlash, then backlashing against the backlash. Meanwhile, we forget to have real conversations about the thing itself.
Personally, I’ve found success using ChatGPT as an all-knowing pseudo-assistant for whatever task I happen to be working. The other day, I didn’t know how to do something in Excel. I asked ChatGPT, and within seconds it spit out the function, how to use it, an example, and two other similar functions to consider.
I’ve also asked it to help get me started writing an email, when I’m struggling to formulate my own words. I would never copy-paste verbatim, but sometimes it only takes a few key phrases to get the creative juices flowing.
These are the ways I’ve used ChatGPT to enhance my day-to-day workflow. But they’re largely industry-agnostic. I had not heard of manufacturers using the tool within their operations, but I was curious of any industry-specific applications.
So, I asked it. How can ChatGPT be used by manufacturers to help their business? It spit out explanations about customer service, lead generation, and product recommendations. But here’s the piece that really caught my eye:
Supply Chain Optimization: ChatGPT can be used to automate and optimize manufacturing processes, reducing waste and increasing efficiency. For example, it can help to schedule production runs, monitor inventory levels, and coordinate with suppliers.
Whoa! That sounds like very real business value. Problem is, I have no idea how it would be used in these ways. So, again, I asked. How can ChatGPT help manufacturers schedule production runs, monitor inventory levels, and coordinate with suppliers? And, again, the response was both promising and entirely unactionabScheduling Production Runs: ChatGPT can analyze sales data and customer demand to predict which products will be in high demand. It can then use this information to schedule production runs in a way that minimizes waste and maximizes efficiency.
Monitoring Inventory Levels: ChatGPT can be integrated with a manufacturer's inventory management system to monitor stock levels in real-time. It can send notifications when inventory levels are low and suggest when to place new orders with suppliers.
Coordinating with Suppliers: ChatGPT can be used to communicate with suppliers and coordinate deliveries. For example, it can send automated messages to suppliers to confirm delivery dates and quantities, and provide real-time updates on production schedules and inventory levels.
I think I stand for all manufacturers when I say we would sign up for these services in a heartbeat. But, of course, hurdles remain. Without being fed real-time data and creating an incredibly comprehensive view of a manufacturer’s suppliers and vendors, it will of course be impossible to provide accurate information about things like inventory levels and demand. It may be able to automate messages to partners along the supply chain, but if those messages contain incorrect information, they’re doing more harm than good.
If and when we reach an era of more robust information sharing across the supply chain, and if and when ChatGPT becomes a more reliable source of accurate information, there may well be something here. Until then, proceed with caution. At MIT Technology Review, Melissa Heikkila writes that AI language models are “notorious bullshitters, often presenting falsehoods as facts.” She continues: “They are excellent at predicting the next word in a sentence, but they have no knowledge of what the sentence actually means.”
Of course, in the meantime, you should also proceed with the baseline work I’ve been advocating for some time. Digitally connect as much as possible in your production line and supply chain—so you’ll be ready when AI truly does offer a solution to help you manage all these connected pieces and parts.
Bottom line: For now, ChatGPT is merely a tool in the toolbox. It will not replace any key functions nor do my work for me. Manufacturers succeed when they’re able to think creatively and institute solutions that distinguish themselves from competitors—those innovative ideas ultimately come from the human beings who are most engaged in the business.
But as a tool, it’s a powerful one. On an individual level, ChatGPT can make us more efficient by cutting down on research time and providing some key points to incorporate into whatever we may be working on. Just be sure to verify.
On an organizational level, the use cases are blurrier. ChatGPT thinks highly of itself and its ability to transform manufacturing. I’m less convinced, although its suggestions are certainly not inconceivable.
Manufacturers should begin playing around with ChatGPT and understanding how language models like this one work—the upside, the limitations, and all the grey areas in between. This sort of AI will undoubtedly one day help us create more optimized and efficient operations. And the sooner we learn more about it, the better. For now, it can help you finish day-to-day tasks faster—and that’s not nothing.