Magic or Mundane? Exploring Generative AI in Procurement

Exploring the cutting-edge world of generative AI in procurement offers both opportunities and challenges for today's businesses. In our latest episode, we sit down with procurement strategy consultant Paul Malott, who shares his poignant insights into how artificial intelligence is redefining traditional procurement roles. 

Generative AI enables tasks that range from automating invoicing to proactively managing supplier risks, transforming mundane processes into streamlined systems that significantly improve operational efficiency.

Paul breaks down the complexity of implementing AI, highlighting misconceptions that often cloud its perception. 

Rather than being a magical solution, AI should be seen as an incredibly powerful tool that can only be effective when used with a clear strategy. Understanding the technological landscape, coupled with the ethics of AI use, is vital for professionals as they navigate this new terrain. 

Through a candid discussion, we also reflect on how generative AI could pave the way for a future where supply chains become increasingly autonomous, relying less on human intervention. This prospect not only raises crucial questions about the implications for employment and decision-making but also suggests exciting possibilities for enhancing our work-life balance.

If you’re keen to gain a comprehensive understanding of how generative AI can be harnessed in procurement and what it means for the future, this episode is not to be missed. 

Equipped with practical advice and thought-provoking insights, Paul’s expertise can help you make informed decisions about adopting AI technologies in your own organisation. Don't forget to subscribe, share, and leave a review to help us reach more listeners curious about the impact of AI on business!

Links to Paul:

https://beacons.ai/automations24

https://www.linkedin.com/in/paul-malott/

www.sourceflo.ai

Chapters

• 0:18 Introduction to AI Unscripted
• 26:05 Generative AI's Impact on Procurement
• 1:05:00 Future of AI in Supply Chain Management

Transcript

Speaker 1: 0:00

Welcome to AI Unscripted, where we explore the intersection of technology business in the future. Artificial intelligence is everywhere from friendly chatbots writing quirky tweets to generative AI art tools conjuring mind-bending images, ai can look downright magical. Today we're diving into a topic that's transforming industries at an unprecedented pace Artificial intelligence, or, more specifically, generative AI, in procurement. Ai has already changed the way companies make decisions, optimize supply chains and drive efficiency, but with the rise of generative AI, we're seeing a whole new level of automation, predictive analytics and even creative problem solving in procurement. Imagine AI not just as analyzing contracts, but drafting them, not just predicting supplier risks, but proactively offering solutions.

Speaker 1: 0:49

So what does this mean for procurement professionals? Is AI an enabler, a disruptor or both, and how can businesses leverage this technology to gain competitive advantage? In this episode, we'll talk to procurement strategy consultant Paul Mullat and we'll break it all down what's real, what's hype and what's next in the AI driven procurement revolution. But let's get started. So, paul, what are the biggest misconceptions businesses have when implementing AI in procurement and how do we navigate those challenges?

Speaker 2: 1:20

Oh, karen, thanks for having me First of all, I appreciate it and great question, so I think you. Thanks for having me. First of all, I appreciate it and great question. So I think you know you mentioned it.

Speaker 2: 1:27

Ai can seem like magic, right, there's a lot of really cool and interesting things that artificial intelligence in general can do, right, and you mentioned it all predictive analytics, ugly, terrible tweets, anything right, can it can, really it has the potential, if used in the right way, to solve a lot of, um, mundane issues. Right, and everybody's talked about this, the tasks that we can complete using ai and different types of ai. But, um, it's not a magic wand, right, it's a tool, more or less, and it's's a tool that needs to be used smart. Think of getting a hammer and believing that you're a carpenter, right, it's the same kind of thing If I watch a YouTube video about how to change an oil in my car and then try to change the oil, I'm not going to be able to do it. But in some cases, people believe that AI is that kind of end-all, be-all solution, and I find that a lot of what I do in the beginning is educate. Right, that it's sure, it can definitely solve problems, but when applied, smart.

Speaker 1: 2:35

Do you think people want to hear that, paul, at the beginning? Because they're all so giddily excited by AI that they believe it can do anything and therefore realists like you or I are come along and talk about rather boring things like what's your business strategy, why are you spending the money, and so on. It can put them right off.

Speaker 2: 2:55

Yeah, I mean, you nailed it. There's something interesting that I study in my doctoral studies at the University of Michigan, flint, and it's information asymmetry between a buyer and a seller and what the gap is between of knowledge on the topic or product they're selling. Right, so take that and apply that to AI. When a buyer believes that AI is that magic wand and the seller's coming in and explaining well, here's the reality of the situation. Finding that middle ground is very tough and businesses that have already paid the money for it are not very receptive to the idea. But those that are looking to still get better understand that there's still a lot of value to be had. Right, Even though it's not going to be the end-all, be-all, implementing AI in the right way is still going to get the results that they're looking for. Maybe just not the you know, the nuclear bomb approach, right?

Speaker 1: 3:54

Well, I hope so.

Speaker 2: 3:54

I don't know.

Speaker 1: 3:56

Many a bomb underneath things, metaphorically speaking of course, but again it is as you say. It's a tool that can be used. It's one tool in the toolkit, but a very important one. So where have you seen it? Working, Paul? What are the types of things that companies are doing that are really impressing you?

Speaker 2: 4:11

Well, honestly, three-way match for invoicing right and purchase order. It's very simple automations and coupled with AI, right to read PDF documentation, and it eliminates human error in the process of taking a PDF, for example, and doing something with it. Now we can use AI to do that three-way match match the PO to the goods receipt and we're essentially replacing SaaS ER by by simple, logical automations, and to me, that's the. The mind blowing thing of it all is that simple has a bigger impact, right, and when you do it um, again, another thing that's prevalent and that is in the studies is when you're doing these simple tasks incrementally and building on it. That's where you get the biggest value, right, so you know, to tie it all together, it's not take a magic wand and everything's you save your a hundred million dollars. It's you gradually build to that point where you created the structure that's going to get you there.

Speaker 1: 5:18

And people talk a lot about AI and then they talk about generative AI as if these are massively different things, two different planets that are never aligned. Where are you seeing generative AI coming into things and how do you see that distinction between AI and gen AI in businesses?

Speaker 2: 5:35

Great question. So the way I like to lay it out is thinking of AI like a family right, and I'll add in predictive AI in there as well. Right, there's the guy in the basement, how it all started, the nerdy dude in the basement who could tell you about last week's news Right and everything about it and all the statistics, or if in America it favorite football players, and you know that's how it started. Right, where we could learn, we could use our human intuition and our expertise to derive conclusions from that predictive stuff. With, with, with generative AI, we're now able to do a lot of mundane tasks that we previously couldn't do and it's literally generating content for us and with the right kind of human involvement, human in the loop that content's really good and it takes I'll use my own example in research and writing papers.

Speaker 2: 6:35

I'm sorry professors who, you know academia is also battling. Ai, just like business, is right, but I extensively use AI to flush out my concepts, write my concepts more clear. I use it for writing grants, grant proposals for the government. There's all kinds of practical implications for just everyday tasks right in business as well, implications for just everyday tasks right in business as well. And now you're coupling that kind of generative AI into a system right, we're starting to see agentic systems today that use multiple AI modules right, and you could have chat GPT talking to Perplexi, talking to Claude to format research and publish to a certain platform the data you're looking for. Now we're starting to talk about using the three types or the two types of AI predictive and generative in an overall system to essentially take over the task and learn from itself. It's really fantastic, actually.

Speaker 1: 7:39

So we can do a lot, obviously, and companies are doing a lot. Nogentic is an exciting space to be in. Do we run the risk, though, that businesses fall into the trap of over-relying on automated decisions, potentially leading to irrational outcomes?

Speaker 2: 7:56

Yeah, 100%. So the Dunning-Kruger effect is something that's really interesting when applied in the context of AI. So, again, it's the idea that you get a hammer and you become a carpenter just because you have a hammer. Ai, because it does things so well, can give us the misconception that we are all of a sudden very good at whatever it is we were using AI to do, and that leads to overconfidence and incompetence. Right, so that's something that's my version of the AI bias discussion. Right, because it's something we all need to be aware of as we use these tools. Are we using them right? Are we using them appropriately?

Speaker 1: 8:40

And so on. You, in your writing, once suggested that deep seek. By the way, an open source release isn't just a technological move, but a strategic one as well. How do you see it influencing global AI governance? What are the risks there for businesses adopting that model?

Speaker 2: 8:57

Sure, sure. So from the ethics and bias standpoint, I mean, there are plenty of examples out there on tweets and you know LinkedIn posts that show propaganda for lack of a better term right Within it. You know saying one thing, not saying something else within deep seek, but for me, the article that I wrote in what I'm really interested in actually deep diving further with regards to deep seek, is how countries and companies can use general information strategically right. So something like deep seek, where you're able to essentially obtain and search a lot of very interesting things very quickly. How can governments and businesses then use that very strategically? It's the opposite of having very targeted data, right, but using the general data as a basic strategy. And where do I see this coming into play globally? It's really, I think you know a country, china making an attempt to strategize that information.

Speaker 1: 10:05

Although arguably we all do that.

Speaker 2: 10:07

I find those see.

Speaker 1: 10:09

As a European I think I am anyway I find some of the debate fascinating. You know, in the nicest sense it's okay, america, let's spend 500 billion. We're going to lead the global race in AI. Next minute deep sea comes out, and whether it's 6 million, 60 million, who knows? It wasn't you know, 500 billion it's. Oh my God, look at those Chinese. They're attempting to dominate the world. There's something wrong, you know. It's just sort of going like as ever. I'm sure the truth is somewhere in the in the middle there, but it's fascinating just to watch this debate. But whoever wins the ai race to a degree and I wish it was humanity, not necessarily one government or another, you know will certainly competitive advantage, wisewise, militarily, do really really well. So where do you see this all? Moving to Paul, where's AI going? Where's Gen AI going? Where are we going to be? Dare we even try and predict, better than a large language model or a small language model, where things are going to end up in the next two or three years?

Speaker 2: 11:02

Oh, two to three years. I see ecosystems that are self-managing right. So thinking of it from a supply chain standpoint, and when you talk about agentic AI and what the basic fundamental value proposition of agentic AI is, which is learning and prioritizing right. I see a fully AI managed supply chain ecosystems Right, where we only need human interjection at the absolute most chaotic moments. Where we absolutely need to interject our experiences in the context Right. But data and simple ERP data things that when I started my career, I would have to log into a company's thing to let them know that I shipped a part to them, for example. You know that went away with EDI, but now we're going to take EDI a step further and we're going to basically it'll just be self-managing itself Right, and only when we need to select a dual source or a new supplier. Do the humans interject Right, only when there's a real problem.

Speaker 2: 12:12

I'm most interested in taking what you said, you know, and building on it Humankind Right. Ai should fall in the humankind realm and I really want to put the proposition out there, while it's still able to be put out there, that AI sure as hell makes it more productive out there, while it's still able to be put out there that AI, sure as hell, makes it more productive, your activity is more productive, but it also allows us for the opportunity to get some leisure right. And when you think about I grew up in Detroit, so I'm partial to automotive industry when you think about Henry Ford and his automation of an assembly line and what came after that was a five-day work week, right, less overtime although some people really strive off of overtime payments Perhaps we as humanity can go for a four-day work week.

Speaker 1: 13:01

I would love it.

Speaker 2: 13:02

That's all I'm saying with three children. Give me more leisure time to spend with them, right? So we tend to look at AI like, hey, it's productive, it'll drive value to the bottom line, it'll save you money All true, right, but let's not forget who we are and what we're in this whole thing for. And it's maybe not leisure for all of us, but it's to enjoy what we're given, right, what the opportunity we have. So maybe we should use this tool to give us more chance to enjoy that right well, yeah, I kind of hope we would.

Speaker 1: 13:32

You know what I mean. I sort of see two extremes here, one is the, what you might call boomer.

Speaker 1: 13:37

You know ai is for humanity. It will allow us to achieve amazing things, you know. And if you take supply chain, then if we could optimize supply chains, of course, then we can remove all the inefficiencies and we can actually get product or service to places that have never had the benefit or advantage of it. Then you've got the doomers or gloomers, depends which end of the spectrum you want, which is, giving a gentic ai or ai free reign will result in the decimation of humanity.

Speaker 1: 14:04

You know our yeah yeah, our ability to work, our ability to decide, and we'll all be out of jobs.

Speaker 1: 14:10

It's funny, I never quite see the bit in between, which is you know again, probably, where the truth lies, which is it can't do all the things that all of the great media is describing. It won't have all of the great dangers that the doomers and gloomers are going out with. I suggest humanity still has a role to play in society, but if we have the best of everything and negate the worst of everything, then, my goodness, we would be in a better place. So it'll be an interesting space to find out and to see what happens over the next couple of years. But, paul, I want to find out a little bit more about you, what you do and your excellent writing.

Speaker 2: 14:45

How did it go about doing that my primary mode of conversation with my community is LinkedIn, and it's LinkedIn and then my name, paul-malott M-A-L-O-T-T. I also run right now three brands. I have a software that's Sourceflowai. It's focused on using AI to build supplier relationship right Supplier relationship management. And then A24, automations 24, which uses automation, makecom, low-code, no-code automations and develops systems for companies, and both of those you can find online wwwautomations24.com and wwwsourceflowai.

Speaker 1: 15:32

Fantastic.

Speaker 1: 15:33

What I'll do is in the notes of the podcast webinar and every other medium that I can get AI to turn this into, I'll include. Paul, thank you so much indeed today. I love your view on the world of AI and generative AI. I find it gorgeously balanced, and I don't often say that, but I'm normally dealing with people who are at the extremes Folks. Please watch, listen in, jump on to Paul's content. It's excellent. Until next time, everyone, I wish you every success in the AI business and generative AI world, Paul, thank you so much indeed, sir.

Speaker 2: 16:06

Thank you, appreciate you.