Procurement, AI, and Two Critical Questions

I just sat in on the Spend Matters webinar Navigating the Future: Meet AI in Procurement.

Right off the bat, it was a very well-run 60-minute webinar that provided succinct insights on various elements of AI and its impact on Procurement. Having attended two different webinars in the past 48 hours, I can clearly see a difference between how a provider presents AI and the analyst approach.

While both were informative and insightful, they were also somewhat complementary in that they had different focuses and views of how AI is changing the procurement world.

At a high level, providers’ general focus is on proving how and where AI can be used today to benefit procurement.

With an emphasis that we are still very early in the procurement-AI relationship, the analysts’ breakdown is on the specific areas where AI can have an impact today and in the future.

Old AI Versus New AI?

Abigail Ommen – Spend Matters

Thanks, Pierre. Hi, everyone. I’m just going to take a few minutes to go over some of the overall trends we found in our AI and procurement survey. So this is not necessarily going to be GenAI specific or anything like that, but kind of a step back, bigger view than that. So, this survey covered more than just GenAI. As I said, it covered traditional types of AI as well, and that is used across the entire procurement technology, like vendor marketplace. So we found that nearly all SDP technology vendors have some sort of native machine learning capability, over 90%, in fact. So, when you look at more traditional AI, it is definitely out there. It’s definitely prevalent, right? GenAI was a little bit different.

As you read the above transcript, what is your notable takeaway? Of course, I have highlighted what part of the above statement had the biggest impact on me.

It stood out for me because it reminded me of the self-learning, nascent AI platform developed and implemented to support the Department of National Defence’s acquisition of MRO parts for its national infrastructure in the late 1990s and early 2000s. Here is the link that will provide a high-level overview of that procurement technology: When It Comes To AI in 2023, We’re Going To Party Like It’s 1998.

What is worth noting about traditional AI is that the questions regarding the importance of data and self-learning algorithms have remained unchanged in the past twenty years—which is the focus of today’s post.

Two Critical Questions

Here are screenshots of the two “crucial” questions I asked the panel of analysts regarding AI, data, and implementation.

Question 1

Question 2

To digress, Bertrand’s statement that “discovery is making a comeback after years of reducing the supply base” brings a smile to my face because I have been writing about how vendor rationalization was a bad idea going back to 2005. Here is the link to the March 2024 post Supplier discovery versus vendor rationalization: A case of an irresistible force meeting an immovable object explaining why.

Based on the answers to the two questions above, I still support Rob Handfield’s position that clean data is a must. His analogy about implementing technology without clean data, like packing your suitcase while you are running to catch a departing train, continues to resonate.

That said, we can’t sit around and wait for the perfect data set to happen before starting to take advantage of AI, or else we will never get started. In very simplistic terms, we need to build a bridge between traditional AI and Generative AI, which is something I have been thinking about doing over the past year or so.

Stay tuned.

30