ChatGPT and other generative AI products have been exploding into prominence over the last six months or so. Now, there are generative AI tools for just about everything. So what about supply chains and ChatGPT?Is the future of supply chains generative AI? And if it is, do supply chain professionals need to be worried?
The truth is, as is often the case, a lot more complex than a yes or no answer. But if you want simple: yes it’s the future, no, don’t be scared.
I’m sure you want a deeper understanding of what that really means though, right? Well… read on.
It’s hard not to give away too much about what we’re working on because, naturally, I feel like those are the best applications. But, essentially, I think there are two really interesting areas in supply chains for ChatGPT and other similar generative AIs. And, for us, those are both around the handling of data.
First, we believe that tools like ChatGPT can help solve the problems that a lot of supply chains face with messy data. Logistics data is relatively well structured, but it uses quite a lot of different language to describe the same things and it's formatted in slightly different ways, but the information that people need is pretty much the same.
So, traditionally, that requires a lot of data engineering to map data to the same format and then clean it and expose it. Because the things that people want to get are fixed, it’s basically like a language problem- perfect for something like ChatGPT.
Second, ChatGPT and the like are well-positioned to handle data queries. Rather than having to manually manipulate data to get the answers you need, why not just ask a language-based model like ChatGPT to find the information for you?
If you're doing supply chain planning, or you're looking at trying to understand complex data, or lots of data at once, you'll need someone to create a bunch of reports. And then find the data to support those reports. And you have to have someone know which reports to go into. They have to filter things, they have to drill down, they have to know their way around, and how to navigate lots of different reports that individually don't give an answer.
Instead, with something like ChatGPT, you could ask human-level questions of data. And it can synthesize a lot of that information for you and give you back a visual chart that illustrates the answer that it's giving you.
This means that the people who are actually using the data have the ability to access it much more directly. And the information loops can be a lot quicker. The VP of your supply chain, your procurement team and your CFO could all see exactly what they need without it having to be filtered through analysts, without giving supply chain managers more admin work and without needing to all be experts in how to use your supply chain software.
There will be an increased onus on data quality. Which brings us back to the idea of generative AI helping us clean data for additional uses. Even so, for most, when you put some dirty data in, you will need to engineer ChatGPT to do what you need to do, plus maybe some traditional engineering.
But data quality becomes really important because that will decide how clearly and quickly you can get information. Then the next thing: is it good information? Is it actually telling me what I need to know?
Those questions are why there will need to be a real focus on data quality and data completeness. We’ll need to think about capturing even more data about what's happening in a supply chain so it can give you more intelligent and more insightful answers.
In that same vein, the focus on clear data and having a natural language way of querying that data means we won’t need to focus on middlemen to do the interpretations for us. Instead, we’ll be able to ask questions of the data much more directly.
I think this idea is mostly based on fear. Instead, I think people who are capable of getting to that strategic sort of level are actually able to do that much sooner than they previously would have.
With the right tools, I think they can get to a position where they can get to that level of seniority quicker. You don’t (and won’t) need to be of a certain age or a certain experience to be able to do those roles. Instead, we'll hopefully just get more impactful professionals quicker.
In particular, we’ll see younger professionals be able to take great advantage of these tools because they’ll use them more natively. Which will help bridge the gap between junior and senior roles.
While you can’t ever replace real-world experience, what this gives you is the ability to do things like create a plan to get the real-world experience quicker, and to start templates for how to build strategies and action plans. And you still end up doing the work, but it shortens the time for you to get the battle scars that the senior team will have gotten.
The day-to-day is broader than just supply chains. It's sort of everything you are doing and everything you're planning for. What I expect is that it will give people ideas on how to tackle problems both in supply chains and in their own lives. In particular, it'll help people do things like the first draft of their work and take away the pain of getting started. So you'll be able to get more done.
For supply chain professionals, it will help you get to grips with both the physical problem and the strategic problem quicker. And that's really the main goal. You’ll still need to get into the details and use human judgement to tailor work to your real-world scenarios, but it gets you to it much quicker.
And you can use it like a mixture of having a permanent intern and a permanent coach at your side. One part for getting your boring busywork done and one part for giving you the backbone of a growth plan.
In the wake of ChatGPT’s meteoric rise to common usage, regulations quickly coming into play. For example, Italy initially banned the AI based on concerns around data protection. Since then, OpenAI has made several changes to the way users can access ChatGPT in Italy.
Meanwhile, several governing bodies in Ireland have banned the usage of it. And both the EU and the UK have had to create AI acts that regulate the ways in which AI can be used and implemented. Which means these questions are much broader than just supply chains.
The main thing to remember is that the cat is out of the bag now. The strong consumer pull that we have seen since its release means that the technology is here and people will find a way to access it. I think, firstly, the analogies will evolve to cater for those kinds of demands. In fact, they already do.
For example, if you use ChatGPT via an API, that data isn’t stored and isn’t used to train the models. It's stored for 30 days for misuse and abuse policy, but then it just is removed. Things like this suggest there will be a (not-so-far-away) future industry around making sure that teams are using these tools in the right way for both efficiency and security purposes. Really this is just a point of maturity on the tech side.
I think regulations are useful to talk about. It makes us ask important questions like ‘What do we want from this technology and how do we want it to work for us?’ And that will help inform the design perspective from the regulatory side. For me, I see this as an important concept in product development around this style of AI.
You probably also need to ask things like ‘What are we losing by implementing this regulation compared to where it might not be in place? And how controllable is that regulation? Is that a rule that someone will be able to get around and is it actually going to be practically enforceable?’ There are lots of crucial questions that you can really only answer through regulation.
But overall, I don’t think that regulation is going to impact AIs, or the supply chains that use them, negatively. I think they can really only help shape it into what people want and need.
ChatGPT and technologies like it are going to change the way we work for sure. And we’ll need to learn how to make them work for us, to make our jobs and lives easier. And we don’t need to be so concerned about this tech putting people out of jobs, especially within supply chains.
Everything is a balance, and there are probably personal preferences that are important to acknowledge, but I think we can't shun this stuff away, not least because it's in our lives now. And it has to be something we learn how to live with and to grow with. The potential of it is too great to ignore and leave on the backburner right now.