The AI Transformation: How AI and Big Data are Shaping the Future

Welcome to another exciting episode of the Digital Supply Chain podcast! I'm your host, Tom Raftery, and today, I'm thrilled to bring you an insightful discussion I had with Keith Hartley, the CEO of LevaData.

In this episode, Keith delves deep into how LevaData is revolutionising the way sourcing professionals in the manufacturing industry understand, manage, and optimise their supply chains. With an innovative approach that harnesses the power of artificial intelligence and large language models, LevaData is making significant strides in the industry.

Keith shares the fascinating journey of LevaData and its mission to provide unparalleled insights into parts, ingredients, and metals data. He underscores the critical importance of data in the current digital era and how it can be leveraged to drive strategic decisions, optimise operations, and gain a competitive edge.

But that's not all. Keith also outlines the future of LevaData and the broader supply chain management landscape. He envisions a world where procurement and supply chain planning converge, with a focus on part-level and ingredient-level intelligence. This vision, as he explains, will undoubtedly revolutionize the way we approach supply chain planning and management.

Finally, Keith invites listeners to experience LevaData for themselves. He extends a personal challenge to all, to engage with the software and see firsthand the transformative potential it holds.

So, if you're a professional in the manufacturing industry, a big data geek, or simply someone interested in the intersection of AI and supply chain management, this episode is a must-listen. Don't miss out on these valuable insights! As always, I look forward to hearing your thoughts and comments.

Stay tuned for more enlightening conversations on the Digital Supply Chain podcast. Until then, keep innovating and transforming your supply chains!

And if you want to check out the video version of this podcast, you'll find it at

Elevate your brand with the ‘Sustainable Supply Chain’ podcast, the voice of supply chain sustainability.

Last year, this podcast's episodes were downloaded over 113,000 times by senior supply chain executives around the world.

Become a sponsor. Lead the conversation.

Contact me for sponsorship opportunities and turn downloads into dialogues.

Act today. Influence the future.

Transcript

Keith Hartley: 0:03

If you look at enterprise software, you know, no one would say that a CFO, or a controller doesn't need an E R P system to do their job. It's just part of doing their job. Yeah. No one would tell a sales organization, you don't need a C R M system to do your job. Of course you need a CRM system, but you turn to direct materials, you turn to sourcing on the direct side. And there is no broad acknowledgement that they need a tool to do their job.

Tom Raftery: 0:34

Good morning, good afternoon, or good evening, wherever you are in the world. This is the Digital Supply Chain podcast, the number one podcast focusing on the digitization of supply chain, and I'm your host, Tom Raftery. Hi everyone, and welcome to episode 318 of the Digital Supply Chain Podcast. My name is Tom Raftery and I'm excited to be here with you today sharing the latest insights and trends in supply chain. Before we kick off today's show, I want to take a quick moment to express my gratitude to all of our amazing supporters. Your support has been instrumental in keeping the podcast going, and I'm really grateful for each and every one of you. If you're not already a supporter, I'd like to encourage you to consider joining our community of like-minded individuals who are passionate about supply chain. Supporting the podcast is easy and affordable with options starting as low as just three euros or dollars a month, which is less than the cost of a cup of coffee, and your support will make a huge difference in keeping this show going strong. To become a supporter, simply click on the support link in the show notes of this or any episode, or visit tiny url.com/ d S C Pod. Now. Without further ado, I'd like to introduce my special guest today, Keith. Keith, welcome to the podcast. Would you like to introduce yourself?

Keith Hartley: 1:54

Yeah. Hi Tom. Thanks for having me. my name's Keith Hartley. I'm the CEO of LevaData. LevaData is, a, a company in the intelligent sourcing space. So we're helping sourcing agents for direct materials, global commodity managers, and, and sourcers do their job, with data better. So giving them contextualized insights around all things, parts, ingredients, and metals. So it's something we we love to talk about, and I'm just thrilled to be here with you today.

Tom Raftery: 2:25

Nice. Nice. And just before we get into the meat of the podcast, if I am someone who is sourcing one of those kinds of things, metals or whatever, what is it you do for me?

Keith Hartley: 2:38

So at leather data we've built, an engine or an aperture that allows us to take a lot of data that does not make sense and is out there in disparate pockets around the world and captive inside of organizations. We contextualize all of that data using AI and using our, advanced algorithms, and we're able to associate those attributes down to the manufacturer's part number, the MPN level. We're the only vendor that benchmarks at the specific MPN level and what a company can do with that, is they can make informed buying decisions. Are they getting the best cost? Mm-hmm. Not just inside the company, across division, but relative to others in the world. And how do they de-risk their supply chain at the part level? Part level risk is critical for having a robust supply chain and something, the LevaData that we spend a lot of time talking to our clients and our partners about. Which is, make sure you're getting the best price and make sure you have resiliency down to the MPN level. It's critical for today's modern supply chain to have that, otherwise they're exposed.

Tom Raftery: 3:57

Okay, nice, nice. And you're just back from a Gartner event, Keith. tell me a little bit about that. What, what were the main themes that you saw there? I'm, I'm sure, I'm sure there was nothing around AI, for example, cuz that's not topical at all.

Keith Hartley: 4:11

Yeah. You know, I, I was at the Gartner Supply Chain Symposium for a few days and it was a wonderful event. I've been going for years and, you know, it's kinda like a homecoming. You see a lot of familiar faces. Some at same companies, some at different companies, but it's always good to catch up and, learn and hear more about the topics of the day and all things supply chain. So I saw a few themes this year that really, I would say popped out at me. One is AI is everywhere. Mm-hmm. Uh, I think on every vendor's booth and every vendor message, had to do with ai and you know, AI is very topical and the way I like to frame up AI is, it's just the next extension of, real decision science. So, The notion of, you know, for I've been in software development for a long time, and we've always updated algorithms based on what we've learned. And it used to be you'd update the algorithm twice a year in the on-premise world and then you move to SaaS and you'd update that algorithm every six weeks. And even with Agile development cycles and two week spins, The algorithm would get updated every two weeks. And now with AI, you're getting into continuous algorithm updating. So it's an exciting time and it's just speeding up some of the work that many of us have already been doing for a long time, which is just learning from contextualized data and applying better, more informed algorithms to it, and the real explosion of AI. Is exposing the fact that a lot of companies have not viewed updating those algorithms as learning from data. It's just a natural extension of what we've been doing. So, you know, now that the AI engines are in place, we have LevaData have been using, AI for years mm-hmm. To update our algorithms and to get better part matching and make it easier to get large, unstructured data into our platform. AI really helps us because every bit of data that we are able to see and learn from, we, we learn from the next bit of data that comes into the system. So it's a really exciting time where we apply AI to things that have historically held customers back from enterprise software. So I'll give you an example. In the old days, you used to give a bunch of data files to a vendor and the vendor would go off and spend months and months ,or some vendors still spend quarters and quarters getting the data into their system so that you would have a usable prompt. Fast forward to today at LevaData, we take billions and billions of spend data, hundreds of spreadsheets, and we contextualize that and can play that back with a high degree of accuracy in days to our customers. So where we've made big investments is on the AI engine to make sure we can contextualize faster and cleaner any data that's out there related to direct material sourcing. So large language models, AI, it's something that we at LevaData have, spent a lot of time doing. That's one of the dominant themes from the Gartner conference. You know, if I were to say another that I saw, it's this notion of supply chain and procurement playing more of a collaborative role. So historically, procurement has been walled off from all things supply chain, particularly supply chain planning, and that's never really made sense to me. As a supply chain planner is doing their job and using robust data systems to make accurate forecast and to iterate on what their supply chain plan should be. Ultimately, action has to be taken and that action is taken by the direct material sourcing agents, people who go out and buy the material that we're gonna then build products for. And what I see is this. It's, it's building the collaboration between all things sourcing and all things supply chain planning. And that was evident at this show. As I talked to more and more practitioners, I saw more and more procurement people, senior procurement people at this supply chain conference. And so my hope and when I started to see, you know, some bits the of a harbinger for. Is that the supply chain ecosystem begins to cross over and, collaborate more with direct material sourcing. To me, that is a huge value add for today's modern organization to understand, you know, what they can buy. Not just where they're gonna go, but how they're gonna get there.

Tom Raftery: 9:18

Okay. And, In the likes of sourcing, and I mean, you said yourself that you guys are helping with sourcing. What about, you know, getting access to information like things, things like the carbon footprint, implications of buying a particular product or, there's a big push in Europe now as well, and it's, it's going further beyond Europe, but you know, do you have. child labor or slave labor in your supply chain? Are those kind of things starting to bubble up as well?

Keith Hartley: 9:50

Yeah, it is. There's increasing interest around all things E S G which is, it's a really exciting time. It's still a very wide open field in E S G. When regulations are in place and set, it allows small software companies and large software companies to write code and attribute those E S G characteristics to your workflow. So at LevaData, adding E S G data is really simple for us. It's another data set and what we've done is built the aperture that lets us take all of these large disparate data and bring it into our, our system and synthesize it. So ESG is just one financial health and ratings. We bring in a lot of data already, and I do see an increased interest in E S G, and there are kind of two flavors that I see emerging right now around ESG. One is kind of that supplier level - is a supplier green, are they carbon reducing? Are they hitting some benchmark? Are they, you know, are they buying good practices, et cetera. But where I play at LevaData is a click deeper. I know more about a part about plastic resins and materials and metals than anybody else. I have more attributal data at those parts, those commodity parts. So I can trace that part and it's life; who the CMODM is, who's, uh, building it, where it's built, which of the 11 factories that build that part inside of the contract manufacturer, where that part is from, what their labor laws are, what the regulations are. So I don't just do a supplier risk fly by of of that type of data. We go down to the specific mpn, which is really critical. You see a lot of companies who have finished good shortages. That's not by accident. It's because some part in their supply chain they were not able to source and they were missing something from the final bill of materials. That's because companies are not going to the end state, down to the, the part or the ingredient level and building robustness there. They're trying to build robustness, at a higher level. And in times, like we're living in today with product price fluctuation, with scarcity of material. With some goods not flowing as easily from one part of the world to the other. If you don't have part level risk and good benchmarking data on what you should be paying, you're likely not optimizing your spend in, in every category. So I, you know, that, that's kind of Yeah. It's, it's an exciting time because now we have the data, now we have the data to take action on that. And so ESG to me, it's just one more data source and we are really well positioned to contextualize that and deliver whatever the regulation by country, by geography, by governing board that is required.

Tom Raftery: 13:11

Okay, cool. I mean, you mentioned risk there is, was risk another theme at the at the Gartner event?

Keith Hartley: 13:18

Yeah. Risk was, I think We do, part risk better than anyone else and, and so that's knowing what the lead times are, what the years to end of life, what the age is of parts, what distributor inventory is, we know more at that MPN level. So when I talk about risk, there are a lot of vendors who do you know supplier risk. Are you red, yellow, green? Are you, do you pay on time? And do we have a healthy relationship or a neutral relationship? That's not, when I think about risk, I think about can I get every part inside of a BOM or to do a finished good? Can I get it and build it? And you know, something as simple as a computer mouse has 200 parts in it. So just to build a simple computer mouse, which we don't really think about that much. Requires 200 different parts, and when you go to real sophisticated products, the amount of parts, the amount of components, the amount of ingredients and plastic resins and plastic parts that go into that, it's numerous. And just missing one parts means you can't have a finished good. It means you have a work in progress. And that work in progress waits until you're able to source that one remaining part. So when I think about risk and when I talk to people, at all shows about risk, many of them are just now waking up to part level, ingredient level, metal level risk. And we call that, you know, supply part risk, part risk and risk intelligence. To me, this is the natural evolution of where supply chains are going. It's not enough just to know if the vendor is good or bad. I need to know if the specific part that I order is at risk, how do I find an alternate vendor? How do I find an alternate price, and how do I potentially not have a shortage of producing finished goods?

Tom Raftery: 15:32

Cool. Cool. Yeah. Uh, and if you, I mean, if you do see that a particular vendor, uh, is likely to have a shortage of a particular critical component, do you then through LevaData, offer alternate vendors?

Keith Hartley: 15:49

Absolutely. Inside of their supplier ecosystem, we even, show alternate vendors. There is RFX capability as well, so you can actually go to those vendors and request, uh, request a quote, request a price, that type of capability to go one step further. We don't do just alternate vendors. We know more about alternate pricing vendors, so part of what makes us unique is at the manufacturing part number level. We know more about that part across vendor and how to realize savings opportunities. So, you know, we have customers in a lot of different situations. You know, one of which we have a customer who has multiple divisions, and each division is buying the same part from the same vendor, all at different prices. So visibility, understanding where you're buying and what you're buying internal, but eventually the conversation turns to. Are we getting a good price relative to the market? And I know more about that part and what the market is paying from a lot of different data sources. And we serve that up to our customers. So we become this benchmarking engine for their commodity parts, for plastic resins and for metals. And that's, that's the lane that we have, we've owned, we we're pretty proud of being in that lane at the very bottom, if you want to call it, of the supply chain, because this is where so much of value is lost and has been lost, is companies don't know if they're getting a good price and don't understand how to de-risk at that individual m p n level. It's really hard and, and LevaData spent years developing our data aperture. That lets us serve up these insights. And one thing I didn't mention, Tom, is the work that we do is really, really hard and complex. And that's why many vendors have stayed away from going all the way down to the MPN level. And now we we're in the market with an easy to use point and click software that people can use and take all of this really complex data modeling and AI algorithms. And actually use it to perform their job better. So we're really, really bullish on the future of sourcing and sourcing's role to play, a collaborative role with supply chain.

Tom Raftery: 18:21

Okay, cool. Interesting. We started off talking a bit about AI and it's been in the news a lot the last few months with the launch of ChatGPT and just yesterday at Google's io event, all the news that came out about Bard and PaLM2 and all these other announcements that they made. And one of the things that these large language models feed on is big data sets. And it seems to me that's one of the things that you have in spades. So are you gonna be taking advantage of some of these large language models, generative ai? And if so, what will you use it for in LevaData? What kind of use cases do you see for it within your own organization?

Keith Hartley: 19:05

Yeah, I mean, LLMs large language models have been around for a while. I think they're, um, You know, AI itself is, you know, I define it as just the algorithm is being updated in real time. And generative AI is, the algorithm is updating itself. Mm-hmm. So that's very exciting. Now part of what makes manufacturers and the data problem exist is because of how disparate the data is. It's very complex. Yeah. So assuming a manufacturer tracks all their data and has all their data, none of it makes sense together unless you have a data aperture that can interpret and, assign attribution to individual elements across things that don't seem to make sense. So we use AI and we have used AI for years. To make those connections, right? Every time we ingest information on a part, our AI engine learns about that part, unique code strings of that part, so that the next time we see that code string, we're able to identify the part sooner than we were previously, or our system gets smarter around parts, ingredients, and metals because of AI. At the end of the day, the AI engine is only as good as the underlying contextualized data that we have started with and brought to this point, which is the most robust complete data set around blended. And so what a large language model is dependent the large, the first L is dependent on a very, very large data set, and so we've got a lot, you know, years of experience at blending that large data set. Understanding what didn't work, tweaking our algorithms. And now that AI has come to fruition, we've written our AI algorithm so that it can continue to progress and learn. It's why we're bringing to market the ability for customers to drag and drop their files into our software and be able to quickly see the data match. We're able to do large level imports and data aggregation on the backend and contextualization, and these aren't just big words for us. This is kind of how we deliver value, the contextualization word. I, I, I love that word, but what it means is I can take a data set and understand most, if not all of that data and where it should be attributed to in the ecosystem of parts, ingredients, and metals. So, AI holds tremendous promise in the data blending in the back end of our system. That's where we use it and the front end when customers bring us data. So I expect the algorithms to get even better because, and, and even faster. Hmm. Because of the underlying data model that we have has been constructed appropriately. We have the large, we have the l, in the large language model sense. So we've got the data set and it's good contextualized data, and now AI can pick up on the intelligence that we've coded and take it from there and, uh, be even more intuitive and even more informed than we were years ago.

Tom Raftery: 23:16

Okay. And what other people who are in the sourcing profession, are they used to using tools like this? I mean, how has that job, how has that job evolved over the last few years and, you know, where is it going from here?

Keith Hartley: 23:30

Yeah, it's a, it's an interesting question, Tom. I, you know, talking to direct material sourcing people, all over the planet, a couple of things ring true. Number one, many of them, they're, they're being underserved, right? They're being massively underserved, and they're still predominantly using 50 year old technology to do their job. They're using spreadsheet and email. Some version of a spreadsheet, some version of an email, and a few relationships with suppliers. And in the new world that we live in, kind of post covid, you know, supply scarcity with, with a lot of macroeconomic shocks continuing to to happen in the world of manufacturing, they're being underserved and they're using really old technology. And it, it's a, it's a personal thorn in my side. Because I feel like the sourcing agent deserves better. If you look at enterprise software, you know, no one would say that a CFO, or a controller doesn't need an E R P system to do their job. It's just part of doing their job. Yeah. No one would tell a sales organization, you don't need a C R M system to do your job. Of course you need a CRM system, but you turn to direct materials, you turn to sourcing on the direct side. And there is no broad acknowledgement that they need a tool to do their job. So we're not a replacement technology at LevaData. We're a do your job the best way that you can tool, and unfortunately, today's direct material sourcing agent is under intense scrutiny. The supply chain people are saying. Make sure you have enough parts to build exactly what our forecast tells us we should build. The finance function at a company is saying, go save two or 3% this year, because that's what we target every year. Yeah. And the suppliers are telling them, Hey, interest rates are a sky high. There's a war in the Ukraine. We're raising out prices. And so the sourcing agent is in this situation where they're under fire. Without a strong, dominant tool set to go fine cost savings and de-risk their supply chain. And so I feel like we owe it to procurement professionals to provide this tool set. They've been massively underserved for years. There's a lot of well-developed supply chain planning and warehousing and transportation technologies, all wonderful markets. This has been an overlooked backend process. That needed digitization and needed a large language model and some AI, and that's what we at LevaData are doing. You know, the time is now to help these sourcing agents and we've got the technology and the ability to help them just, you know, do their job. They're doing an incredible job with email and spreadsheet, an unbelievable job, but they can be doing better with an integrated system that can take more data and deliver them insights to, to allow them to be the hero at their company.

Tom Raftery: 26:53

Sure, sure. How do you tell them that? What I mean by that is, you know, if they're happy, if they're happy with spreadsheets and email, how, how do they, if they're unaware, That there's better options out there. How do you make them aware?

Keith Hartley: 27:13

You know, it's, it's, it's one-on-one conversations. I mean, we're telling everybody, we're showing everybody, and now we're going to market, letting people try our software, letting them see it. We've, we've done a couple of things. We've told sourcing agents to send us a few commodity parts and we'll quickly show them the level of detail that we can show them on some parts that they care about. We go to market and work with customers and give them access to our system and show them the power. And one thing I've learned, Tom, is, you know, LevaData is so unique in our MPN benchmarking and pricing and, attributal knowledge of a part. You have to really experience what we have. You can't tell people and show them PowerPoint slides. So I'm a big fan of engaging people in the software. As the CEO, I bring up the software and show people the software, and you have to, you have to be in this workflow and understand how underserved it is. And then when we show people our data aperture and how easy intuitive it is and how your job can be easier and you can deliver more value, the conversation quickly turns. Then it comes to, can you possibly take billions of spend data and show me something. And I tell people to challenge us. Give us your files and give us two days. It's amazing what we can show in what our data aperture that we've built. And you know, part of my hope in this podcast is, listener will be listening and check out our website, levadata.com. And challenge us, bring us some parts and engage with us, cuz we will get you in the software quickly and you have to see it to believe it. So that's kind of my, uh, personal challenge to all LevaDatans is to be in the software all the time figuring things out for our users who are just, you know, underserved, they're just not getting what they need to do their job from their companies today.

Tom Raftery: 29:23

All right. And where to next for LevaData? I mean, what do you see coming down the line for the next five, 10 years?

Keith Hartley: 29:31

Yeah. Well, I, I think this convergence of, on the direct side, of procurement, procurement and supply chain planning, emerging. I see there is vast areas for part level intelligence, ingredient level intelligence, metals to play a role in the supply chain planning process. So once you decide to build a hundred widgets, you need to translate that to, do I have enough parts and ingredients to build a hundred widgets? And if I'm short, what action do I. So we are on the, we know how many parts and what action to take and connecting that with the supply chain planning ecosystem, which I come from and which is a well-developed market. I think that's the great next frontier for all things supply chain. That to me is where our future is gonna be. And staying in our lane about knowing more about parts, ingredients, and metals than anyone on the planet, that's the company thesis for me. We're gonna stay where we're at. Our lead in terms of contextual data at the MPN level is huge and we're gonna extend it even more. So I'm looking for more parts and more ingredients that I can track and, uh, put into our, into our model. So I'm really excited that's, that's the future for LevaData.

Tom Raftery: 30:54

Cool. Cool, cool. We're coming towards the end of the podcast now, Keith, is there any question I have not asked that you wish I had or any aspect of this we haven't touched on that you think it's important for people to think about?

Keith Hartley: 31:08

You know, thank you for that. I, the only one that I can think of is you really have to experience LevaData. You have to be in the software to get it. I've attended many podcasts and things sound wonderful and, you know, my commitment to all of the listeners and the viewers of this podcast is we can give you an easy experience to understand the power of LevaData and how we can help you. We're at the ready and we're, you know, direct material sourcing advocates. You know, at, at the end of the day, we're really big data geeks that are trying to help an underserved market. And so let us help you, give us a shot, let us show you what we got, then understand what the system is. I think that's the best way to find innovative technologies and new ways of thinking about your workflow. So, A lot of companies don't allow you to do that with their solution. I'm so confident there's huge value there. We do let customers engage inside the product. So I would welcome any and all, requests for that. So cool. That's, um, really it Tom,

Tom Raftery: 32:25

so if people want to know more about some of the things we talked about in the podcast today or wanna try out, you know, where would you have me direct them?

Keith Hartley: 32:34

Yeah, I'd say go to lev data.com. You can contact us there. We have a free part insight where you can give us a few part numbers and we'll give you, you know, the depth of our information on those. Take us for a shot. You can also request a demo right on the site and learn more about, how we contextualize data. So I'd say go to levadata.com.

Tom Raftery: 32:55

Perfect. Great. Keith, that's been really interesting. Thanks a million for coming on the podcast today.

Keith Hartley: 33:00

Thank you so much. It's been a, my pleasure.

Tom Raftery: 33:04

Okay, we've come to the end of the show. Thanks everyone for listening. If you'd like to know more about digital supply chains, simply drop me an email to TomRaftery@outlook.com If you like the show, please don't forget to click Follow on it in your podcast application of choice to be sure to get new episodes as soon as they're published Also, please don't forget to rate and review the podcast. It really does help new people to find a show. Thanks, catch you all next time.