Procurement technology is developing rapidly. Systems driven by machine learning and artificial intelligence are taking over key decisions about supply strategy and practice that were once the exclusive province of humans. In this conversation with SupplyChainBrain Editor-in-Chief Bob Bowman, Jim Bureau, CEO of Jaggaer, discusses the progress of the technology, and the ultimate role of humans in the process.
SCB: What is your definition of autonomous procurement?
Bureau: We view it as the culmination of cognitive intelligence, the ability to proactively make recommendations based on patterns that are coming from not only from our own systems, but those of third-party systems as well. It brings the level of expertise that a human would have to the forefront, without human intervention. It allows humans to focus on the work that’s most appropriate for them to do.
SCB: How new is the concept?
Bureau: I don't think it’s all that new. Machine learning and artificial intelligence are getting us a whole lot closer to it. There are various steps of autonomous procurement. You’re training RPA [robotic process automation] to see patterns and recognitions. That's been around for 15 years, but it's been augmented by human intervention. What hasn't been around for a while, and is being put into productive use now, is pattern recognition.
We’ve been in business for 25 years. The question then becomes, what do we do with 25 years’ worth of data, so that we can pull it into machine learning and apply that wealth of knowledge to something meaningful? Without good data, you can't do machine learning effectively, or put artificial intelligence against it.
SCB: When we talk about autonomous vehicles, we often use the phrase “self-driving.” Are you proposing self-driving procurement, or is that something of a misnomer?
Bureau: Do I think it means self-driving procurement? No. In places where people are doing fairly rudimentary procurement, and even some that require a little bit more strategy, you will absolutely see some self-driving. That being said, I do think the human facet will continue to exist. We'll be able to focus more on strategy and how we get autonomous components working well in bigger, broader areas. You're going to see an evolution in where you can apply autonomous procurement to different categories, as it unfolds over the course of the next five to six years.
SCB: When you say strategy, could you be more specific about the continuing role of the human? Will they be making the ultimate decision on a supplier, acting as the sign-off person? What will the human be doing versus what the machine is doing?
Bureau: I don't think that the human will be signing off on those things. They’re relatively easy to define and work into an autonomous procurement system. The COVID-19 outbreak is a classic example of what humans will focus on in terms of risk management — the things that people can't get to because they're so bogged down doing manual interventions. I don't need a person to say, "OK, all of these criteria exist; I can sign off on it." A machine can do that. What a machine can't necessarily do is the strategic aspect of my supply chain. How do I apply the capabilities that artificial intelligence and machine learning offer up to me, to be able to hone in on fixing my risk management for things like a pandemic, or global economic unrest? Where you’ll see human intervention is in managing the autonomous procurement framework, as opposed to the details underneath.
SCB: But is the system actually making decisions on suppliers? Is it weighing criteria such as cost, geography, and potential supplier stability? Or is it the human who's doing that?
Bureau: I actually think it's the system that does that. What autonomous procurement brings into the mix is a cognitive ability to take information from multiple areas. The computer “brain,” so to speak, is much more capable than ours to process through volumes and volumes of data.
SCB: So a machine can help you diversify your supplier base, to avoid single-sourcing situations and thereby reduce your risk of supply interruption?
Bureau: Absolutely. Advanced sourcing optimization can take a whole multitude of criteria and characteristics, and identify patterns among various suppliers. Companies are leveraging the capabilities of the tool to manage large volumes of data much more effectively.
SCB: Are we talking about the broad universe of procurement, including manufacturing? Raw materials, components and stuff that comes into a plant?
SCB: And it includes both indirect procurement and strategic sourcing?
Bureau: That’s the beauty of it. You’ve got access to all of that data. Machine learning and artificial intelligence take all of those factors into account. For discrete manufacturing, there's a whole slew of variables that have an impact on the process. Autonomous procurement allows me to pull in more of those than I could ever possibly do without it. It gives me a more refined, quicker and automated way to make decisions.
SCB: Is the technology still developing and maturing?
Bureau: Absolutely. The tools are available today, but we need to continue to work through the use cases. We have customers on all spectrums of autonomous procurement. Some of them are deploying RPA. Others are utilizing the system to do assisted learning and guided recommendations. How do I put that into my use case and workflow in order to get the outcome I'm expecting?
SCB: How important is it to be involved in autonomous procurement today?
Bureau: If you're not doing machine learning and artificial intelligence today, you're going to be left behind in a material way. Those are table stakes that you have to be doing now.
SCB: Is there a challenge in getting people to embrace the technology?
Bureau: There can be an issue of change management. One of the things you have to do with autonomous procurement is let the system drive. Humans have a difficult time with that. If you've ever been in a self-driving car, it's very difficult to let go of the wheel. People don't want to do that. It’s a mind shift.
Timely, incisive articles delivered directly to your inbox.