# Who Decides When AI Recommends? — full webinar transcript October 9, 2026 · Zinit webinar · Approximately 60 minutes This is the complete automated Zoom transcript, with original wording and speaker labels. It has not been editorially corrected and may contain transcription or attribution errors. Use the timestamps to check important statements against the video. Replay: https://zinit.com/events/procurement-ai-operating-model/replay ## Transcript **[00:00:03]** James Meads: Okay, good morning, good afternoon, everyone. A very warm welcome to this October 9th webinar with Zinnit. I'm just gonna wait another minute or two to let anyone trickle in that's maybe a couple of minutes late, and then we will get started. So, while I'm doing that. **[00:00:22]** James Meads: Just a quick introduction from my side before I do a couple of polls, first of all, to kick things off. **[00:00:29]** James Meads: And to introduce you to our panel today. My name is James Meads. I'm going to be your host today. A little bit of my history. I'm a former procurement professional. I worked 16 years in the automotive and consumer goods industries across the UK and Germany before I ultimately left. **[00:00:51]** James Meads: to do my own thing, and I focus much more now in the procurement technology space. I run a directory website with over 400 procurement tech solutions and some data points around those, and I provide advisory services and consulting to procurement technology startups, mainly in the go-to-market area, also help them with a bit of content too, as I'm doing with **[00:01:15]** James Meads: The guys from Zenit today. **[00:01:18]** James Meads: We've got a great panel lined up for you to discuss the question of who decides when AI recommends, or should we say, when should a human be in the loop, and why should they be in the loop? We're going to dig into that in quite a bit of detail with our panel. **[00:01:37]** James Meads: Just a little bit of housekeeping before I launch the first poll. So, at the bottom of your screen, you will see a button, an icon with Q&A on there. **[00:01:47]** James Meads: That's where during the webinar, if you have a question that you want us to tackle, just pop the questions in there. We are all able to see the questions, so we'll be making a note of those as they come in. **[00:02:01]** James Meads: We're going to do Q&A at the end. So at about 1:45 p.m. eastern time, we'll finish off the content, the romp of the webinar, and then we'll go into Q&A and tackle all your questions. So I would encourage you if something sort of flashes in your head when you're listening to some of the answers and you have a question, pop it in there and we'll tackle them all at the end. **[00:02:23]** James Meads: So, with that said, I'm going to kick off our first poll. **[00:02:28]** James Meads: So, how this works… **[00:02:33]** James Meads: is you just need to click on one choice here. So, I just want to get a flavor of how mature your organization is. Where is AI in your procurement workflow already today? Is it not in use yet? **[00:02:47]** James Meads: Are you using it maybe for analytics and reporting? Is it already making recommendations for you? Or even are there any of you that have it executing some tasks on its own? **[00:02:58]** James Meads: Really interesting actually watching these answers come in. **[00:03:04]** James Meads: Yeah, okay. **[00:03:07]** James Meads: That's pretty surprising. I thought there would be more of you **[00:03:11]** James Meads: So there's only 33 answers, okay, but yeah, let's give it another… **[00:03:16]** James Meads: Let's maybe give it another 30 seconds. **[00:03:19]** James Meads: Just for you to… To answer that, but yeah, it looks like almost half of you… **[00:03:25]** James Meads: And not using it yet. Wow, that does surprise me. **[00:03:30]** James Meads: Okay. **[00:03:32]** James Meads: Interesting. **[00:03:35]** James Meads: All right, I'm just going to give you a few more seconds. It looks like 70% of you have answered. If you haven't, then just click on whichever one is relevant. Really love to get a few more answers here before I close this out. **[00:03:50]** James Meads: All right. **[00:03:54]** James Meads: So, 45% of you, AI is not in use yet. Interesting. I didn't think it would be that high. 21% using it for analytics and reporting, 29% **[00:04:06]** James Meads: percent already using it for making recommendations, and just 5% of you are already executing some tasks. **[00:04:13]** James Meads: With AI on its own. **[00:04:16]** James Meads: Hmm, okay, cool. That's really insightful as we go into this, and… **[00:04:22]** James Meads: Yeah. Panel, what do you think of that? So I'm going to introduce the panel now, ladies and gentlemen. So Samir, if I can kick off with you, sir, just give us a quick intro and some of your background and what do you think of that poll as well? **[00:04:37]** Samir Khushalani: No, thanks. Thanks, James. Good morning, good afternoon, perhaps even good evening to all of you. I'm Samir Khushalani. I've been working in the procurement space for over 3 decades, and **[00:04:47]** Samir Khushalani: And I've had, I guess, the privilege of helping companies transform their procurement function, serving in roles, leadership roles at EY, KPMG, and most recently, McKinsey, where I served as the digital procurement and operating model practice leader. **[00:05:04]** Samir Khushalani: Earlier this year, I stepped away from full-time consulting to become a senior advisor, which has been super exciting. **[00:05:11]** Samir Khushalani: I advise procurement executives on how they can unlock impact using AI, and what that means for their operating models. **[00:05:18]** Samir Khushalani: I'm also working with a few tech startups, including Zenit, on their product roadmap and go-to-market strategy. And then finally, I mentor students who are interested in careers in supply chain and procurement at Rice and North Carolina State. **[00:05:33]** Samir Khushalani: In terms of the poll, James, I thought… I was actually very surprised, because generally what we see is **[00:05:40]** Samir Khushalani: The vast majority of people are dipping their toes into the proverbial AI water, but most of them have yet to scale. So to see, or at least in this particular audience, that nearly 45% of the participants were not using it at all did surprise me, admittedly. **[00:06:01]** James Meads: Yeah, me too. Thanks, Samir. So, someone else that advises procurement tech startups, a competitor in the house. You're gonna have to be careful. Okay, awesome, thank you, Samir. So, Tom, if I can go over to you now, sir. **[00:06:13]** Tom Lynch: Yeah. Thanks, James. My name is Tom Lynch. I've spent the last twenty six years leading global businesses for Corning Incorporated, a specialty glass and ceramics company. **[00:06:25]** Tom Lynch: I've done that across a variety of industries from telecom, life sciences, automotive, consumer electronics. **[00:06:33]** Tom Lynch: Sometimes headquartered in Asia, sometimes headquartered out of the U.S. **[00:06:38]** Tom Lynch: I've spent most of my career in operations, leading **[00:06:44]** Tom Lynch: Operations or general management working very closely with **[00:06:49]** Tom Lynch: The supply chain and procurement professionals in our business. **[00:06:53]** Tom Lynch: the technologies and the processes that they use. So, as I think the poll indicates, we're just embarking **[00:07:02]** Tom Lynch: on the use of AI, so I think today's discussion will be… **[00:07:07]** Tom Lynch: quite helpful and quite useful, and I'm very happy to be participating in that part of the discussion today. **[00:07:13]** James Meads: Thank you, Terry. Thank you, Tom. And then finally, last but not least, Terry, over to you. **[00:07:19]** Terry Nadeau: Thanks, James. Terry Nadeau, like James said. I've been in procurement and supply chain for over 36 years. I've done pretty much every role in **[00:07:33]** Terry Nadeau: Cut across, similar to Tom, multi-different industries, anything from automotive to industrial. **[00:07:41]** Terry Nadeau: My last six years at Johnson Controls before retiring, I was the chief procurement officer and vice president of global materials. **[00:07:52]** Terry Nadeau: you know, I'd say over my career, I kind of got, a reputation of driving transformation and major change, right? Really trying to elevate the procurement function to the C-suite. **[00:08:09]** Terry Nadeau: The last couple of years after retiring from Johnson Controls, I've been working with a small automotive company, and I've been doing quite a bit of supply chain consulting, helping procurement organizations reach their full potential. **[00:08:24]** Terry Nadeau: you know, also, super surprised on the results, but also excited, right? I mean, you think about the value that, you know, your organizations are going to be able to drive as you embark upon AI and Agentic **[00:08:39]** Terry Nadeau: models and stuff like that, so… so I think this will be a rich discussion, and I… I think there's going to be a lot of opportunities for… for all of you guys to drive shareholder value. **[00:08:52]** James Meads: Fantastic, thanks, Terry. Okay, so that rounds off the intro, so before we go into our first question, I've got another poll for you, that will… **[00:09:03]** James Meads: That will get the gears turning. **[00:09:05]** James Meads: Before we run into our first question. So if you can again, please fingers on the buzzers. What's the biggest barrier to letting AI take on more procurement decisions in your organization now? **[00:09:18]** James Meads: Probably a little bit of all of them, right? But you can only pick one. So pick the first one that pops into your head. And I'm again very, very curious to see what comes back. **[00:09:37]** James Meads: Okay, interesting. This one's… A little bit more even, but yeah, there's still quite a few coming in. Alrighty. **[00:09:47]** James Meads: I'll give it another 20-30 seconds, then I'll… Share the results. **[00:09:56]** James Meads: Yeah, there's still a few of you that haven't. If you've **[00:10:01]** James Meads: If you've got thoughts on this, just click which the first one is that comes into your mind, comes into your head. **[00:10:09]** James Meads: All right, it looks like about two-thirds of you have answered, so I'm gonna… **[00:10:16]** James Meads: I'm gonna end this one now. **[00:10:20]** James Meads: Okay, let's have a look. Alrighty, so… **[00:10:24]** James Meads: This one's perhaps a little bit less surprising, so 44% of you said trust in the recommendations, and then number two were issues around tech and integration, and the only thing that maybe surprised me a little bit here is data quality. You know, as procurement professionals, we always complain **[00:10:43]** James Meads: that we have crappy data, and that we've not had the necessary investment in procurement data. You know, when we speak to our sales colleagues, they don't typically have issues with data, but boy oh boy does procurement in terms of how unstructured a lot of our data is. So that one kind of surprises me a little bit. **[00:11:01]** James Meads: Alrighty, I'm gonna stop sharing that, and I'm gonna jump into the first question, which I'm gonna put, Samir, to you. So, first, and just for the audience, we're not going to do any slides, this is really just a panel chat, so we're not going to bore you with PowerPoint. **[00:11:17]** James Meads: So Samir, beyond efficiency, what new sources of value can AI unlock in procurement? And are there any examples of tangible results that you've seen specifically from, you know, your very long career in consulting? **[00:11:32]** Samir Khushalani: This is a nice way of saying I'm old, but okay, I'll ignore that for a moment. But no, honestly, I think that's a great question, and I especially like the premise of the question, because **[00:11:42]** Samir Khushalani: Yes, clearly there is an efficiency impact, which is all about doing things better, and that comes from reducing cycle time, improving productivity, and so on. **[00:11:52]** Samir Khushalani: But that's just one component. With AI, you don't just… you don't only do things better, you also get to do better things. In other words, things you could not do before. **[00:12:04]** Samir Khushalani: And that's specifically around value creation and value preservation, and I'll double-click into each of these. **[00:12:10]** Samir Khushalani: So, if you take value creation for a moment, with AI, an example of how it can create value is by allowing procurement to increase the spend influenced without adding headcount. So, a good example of this would be tail spend, which is often a category that's overlooked, right? **[00:12:30]** Samir Khushalani: And Walmart actually had a public case study about this. I think they had over 100,000 suppliers, about 20% of them were the proverbial tale where they had standard terms that were often not negotiated. **[00:12:43]** Samir Khushalani: They used AI to autonomously negotiate with all of these suppliers based on some guardrails that they had defined. **[00:12:51]** Samir Khushalani: AI did the text-based negotiations. They made offers, they received counter-offers, traded across variables, and so on. And at the end of it, there were about 68% of the participating suppliers with whom they were able to get around 3% of savings. **[00:13:07]** Samir Khushalani: and extend the payment terms by 35 days. So that's pretty, you know, measurable, tangible value, all done by AI without really any human intervention beyond the setting of guardrails. **[00:13:21]** Samir Khushalani: And there are lots of other examples, but just in the interest of time, I'm just giving you one. In the area of value preservation, which, by the way, means preserving the value, protecting the value that you just created. **[00:13:32]** Samir Khushalani: by reducing leakage, etc. And one way to do that is by ensuring that the invoices that suppliers submit **[00:13:40]** Samir Khushalani: are compliant with the contracts. In other words, they reflect the contractual terms that have been negotiated. **[00:13:46]** Samir Khushalani: And I can actually give you a real-life example of this, at a global pharma client that I, that I served last year. **[00:13:55]** Samir Khushalani: So this client has an R&D spend of over $4 billion, so just their R&D spend, right? And we looked at their contracts in clinical trials outsourcing, and these are global, complex, multi-year contracts, several hundred pages in length. **[00:14:12]** Samir Khushalani: With very, if you will, onerous rebate structures. So there were these contract extension rebates, a volume rebate, a non-compete rebate, a new award rebate, and so on, and different rebates applied to different scopes of work. **[00:14:27]** Samir Khushalani: And then the invoices you would get were also quite complex, several, you know, more than 100 line items each, right? **[00:14:35]** Samir Khushalani: And by the way, all of these contracts and invoices invariably were PDF files, so to reconcile the two would literally have taken an army of humans. **[00:14:44]** Samir Khushalani: But with AI, what we were able to do was extract the appropriate terms from both documents, and we uncovered a leakage of nearly 4%. **[00:14:55]** Samir Khushalani: And this leakage came from things like labor rate discrepancies, wrong foreign exchange calculations, inaccurate inflation adjustments, missing rebates. And on top of all of that, almost as a bonus, we found Recurring spend. **[00:15:12]** Samir Khushalani: That was not covered by the contract, in other words, spend that had not been influenced by procurement. **[00:15:18]** Samir Khushalani: So another example of how, with AI, we were able to tackle something that clearly was not being tackled before. **[00:15:27]** James Meads: Yeah, I like that. That's a really interesting example, especially around the contract leakage, because it's a big one, isn't it? In that central procurement or corporate procurement will do a global frame agreement, but then usually there's a plan somewhere in the world. It's probably usually in a couple of European countries that just doesn't want to implement it or, or has their own supplier that they don't want to get rid of. And and you get that contract leakage, but also just the invoices against. **[00:15:52]** James Meads: the contract, you've got one set of data in one system. **[00:15:56]** James Meads: Typically the ERP or resource to pay suite. And then you've got your contract data in a, in a CLM and they don't talk to each other. And and yeah, I can and orchestration now can, can, can enable that much, much more. It's interesting that the example that you gave of Walmart, I good friend of mine's head of marketing for that company that came in and did that autonomous negotiation. So it's a it's a small world. Tom, Terry, do you have any thoughts? **[00:16:21]** James Meads: What's on this one that you want to jump in before we, before we move on? **[00:16:26]** Terry Nadeau: Sure, sure, I mean I'll. You know, Samir gave a great example right of tailspin in the lower end of the categories. Right, if I take the flip side of this and talk about, you know, category strategies, right, you know, I think most of you got your. **[00:16:43]** Terry Nadeau: category management teams or whatever working on, you know, the strategically important categories. And for me, that's where you're driving a ton of value, right? You're differentiating your product offerings, **[00:16:58]** Terry Nadeau: And so what I really get excited about is, you know, generally in the past, we would create a strategy, it'd take us forever. **[00:17:07]** Terry Nadeau: And then generally it sat around for about a year, right? Before you had the resources or time to update that and freshen it. You know, what we're seeing going forward is really what we'd call like a living category strategy, right? And so you're gonna be able to just constantly update category strategies almost in real time, right? Understanding market dynamics and really taking advantage **[00:17:31]** Terry Nadeau: Advantage of changes in the market, right? You know, things might be tightening from a market perspective, so you're reducing supply, so you're not overpaying, and then you may flip it right on the other end where, you know, the market is down from a price perspective and you really want to start building inventories, right, because you do have the demand, but the pricing is where you want it to be. And, and I just. **[00:17:55]** Terry Nadeau: you know, for me, category strategies is really about deep domain knowledge, right? And so, what you're gonna see with AI is you're gonna be able to get these agents to get **[00:18:07]** Terry Nadeau: an unbelievable depth and understanding of that, right? And because they're agents within your company, you're not losing, you know, that intellectual property and knowledge, right? And so it's going to be a super powerful tool. It's going to allow us to focus our resources really on relationship management. **[00:18:29]** Terry Nadeau: Right? And really driving the change based on the recommendations and and the strategy in and of itself. **[00:18:37]** James Meads: I love that example of category strategies, and this is one of the things that I've spoken about on LinkedIn in the past as well, that a procurement category manager, like you said, is a deep subject matter expertise on that specific category. Although I do think **[00:18:53]** James Meads: The worm is turning a little bit with all the market intel that's out there now that someone can come into a new category and get up to speed much faster. But I always say that the value in the strategy is actually implementing it. **[00:19:05]** James Meads: You don't need someone on a six figure salary fiddling around with boxes in PowerPoint and, and trying to get slides to fit. You know, that's that's work that a good admin assistant should be doing. But inevitably, procurement professionals end up and end up doing that work. So yeah, I think that's a great example. Just before we wrap this up. **[00:19:25]** Terry Nadeau: the. **[00:19:26]** James Meads: Yeah, yeah, go ahead. **[00:19:26]** Terry Nadeau: James, you know, the interesting thing, Samir's comment, you know, about the quality of work, right, what people spend time on. **[00:19:33]** Terry Nadeau: To me, what that's gonna do organizationally is really create your department and your your business to be very, very sticky, right? Because buyers are gonna be able to do things that they like and that they enjoy doing, right? And how in creating value. So that's the other big thing from my perspective. Sorry about that. **[00:19:52]** James Meads: Yeah, no, you're spot on. Retention, especially for the best people. **[00:19:57]** Terry Nadeau: Yeah. **[00:19:57]** James Meads: is going to become more and more of a factor now, especially, you know, with things like with remote work and just the temptation for procurement professionals to go and work for consultancies or go and work for tech startups. It's. You're absolutely right. Yeah. Tom, over to you. Before we round this off, is there anything, anything that you want to add to Terry? Terry's comments? **[00:20:17]** Tom Lynch: You know, that's okay. I think in the interest of the time, I'll comment on the next couple. **[00:20:24]** James Meads: Alrighty, okay, so the next one that we had then was more around what an Agentic Procurement workflow looks like in practice. So, Samir, I'll stick with you on this one. Can you share some examples of how it actually changes the way the work gets done? **[00:20:41]** Samir Khushalani: Yeah, no, happy to. And I think this is one where, you know, we tend to say Agentic workflows as if it's this one monolithic thing, and the reality is it isn't. **[00:20:52]** Samir Khushalani: it's a lot more nuanced, if you will. At least the way I look at it, I think there are really 3 archetypes of Agentic workflows, with varying levels of agent autonomy. **[00:21:03]** Samir Khushalani: But humans are always in the control… are always in control, and I think that's really important, especially today, because in the news, we've been hearing so much about AI agents going rogue, but **[00:21:15]** Samir Khushalani: So, yes, humans are in control, but they're, you know, agents can get different levels of autonomy based on these three archetypes, and I'll touch upon each of these, right now, right? The first one is what I would call human in the loop, and this is one where **[00:21:31]** Samir Khushalani: Agents perform a lot of the work, but they require explicit human intervention. **[00:21:37]** Samir Khushalani: to do certain… at various control points, like approvals, decisions, etc. But humans are needed to execute every single transaction. **[00:21:46]** Samir Khushalani: That's human in the loop, and an example of that could be, an example of a workflow could be a supplier award for strategic categories. So here you've got AI that evaluates the proposals, it scores the suppliers, it models scenarios, and then it recommends an award. **[00:22:03]** Samir Khushalani: The category manager reviews that recommendation, and then approves the award. So, in other words, a category manager is always needed. **[00:22:12]** Samir Khushalani: Especially for strategic categories. **[00:22:14]** Samir Khushalani: The other, the next workflow or next archetype would be human on the loop. **[00:22:19]** Samir Khushalani: This is where humans are not involved in every transaction. They intervene, but only on an exception basis, when certain predefined conditions are met. **[00:22:29]** Samir Khushalani: So an example of that is, invoice exception management. Now, this is one where, let's say you have AI that's reviewing the blocked invoices queue, and God knows there tend to be a lot of blocked invoices. **[00:22:42]** Samir Khushalani: AI is now diagnosing Why the invoice was blocked? **[00:22:46]** Samir Khushalani: and then resolving the discrepancies. And those discrepancies that AI cannot resolve get escalated to a human. **[00:22:54]** Samir Khushalani: Contrast that to today. It's the job of AP, human AP analysts who are, you know, managing the workflow. **[00:23:02]** Samir Khushalani: Or managing the SKU of, exceptions, of blocked invoices. And if I were to just take it down to the next level of detail, if you will. **[00:23:10]** Samir Khushalani: When invoices get blocked, it could be for a variety of reasons. A common reason, that I have often seen is when the goods receipt doesn't get posted. This is in a typical three-way match situation, right? **[00:23:23]** James Meads: Yeah. **[00:23:23]** Samir Khushalani: So in a scenario like this, today, AP would be finding, oh, the goods receipt hasn't been posted, they call up the requisitioner, and so on and so forth. Instead, AI diagnoses this, they identify the requisitioner, follow up with them, and then wait for the goods receipt to be posted. **[00:23:39]** Samir Khushalani: And if the three… if… when the goods receipt is posted, and the goods receipt PO and invoice match. **[00:23:45]** Samir Khushalani: The invoice gets approved for payment, and if they don't match. **[00:23:49]** Samir Khushalani: Then it gets escalated to a human, who then determines if the invoice should be rejected, or partially approved for payment, and so on. **[00:23:55]** Samir Khushalani: But here's an example of a number of invoices that will not necessarily require human intervention, but the AI agent takes care of it. **[00:24:06]** Samir Khushalani: And then the third, archetype, if you will, would be human above the loop. And this is where humans define the objectives, the policies, the guardrails, and so on. **[00:24:17]** Samir Khushalani: And agents execute the work within those boundaries. There's full auditability and dashboards, etc, to review the work of the agents, but they're… the agents are operating within the guardrails established by humans. **[00:24:29]** Samir Khushalani: And then the worldwide example that I just gave around autonomous negotiations, in my mind, as an example of human above the loop. **[00:24:38]** Samir Khushalani: Because there, or at least in this situation, the humans are defining the guardrails of acceptable payment terms, discounts permitted, concessions allowed, termination provisions, and so on, but the agents are the ones negotiating. **[00:24:52]** Samir Khushalani: And humans are not… don't need to be involved. **[00:24:55]** Samir Khushalani: So, these are the three, if you will, archetypes, as in, depending upon the workflow, and by the way, that's the design decision most companies need to make, which workflow should be in, you know, subject to human in the loop, human on the loop, and finally, human about the loop. **[00:25:11]** Samir Khushalani: But there is one, one additional point I would like to make about this, which is, yes, there are archetypes. **[00:25:18]** Samir Khushalani: But they can also be described as maturity horizons. So think of it as graduated autonomy, because even if the right archetype is human above the loop. **[00:25:30]** Samir Khushalani: Because of the lack of trust, as we found in the poll that was just conducted, people may not be willing to jump to human above the loop on day one. They may want to start with human in the loop. **[00:25:41]** Samir Khushalani: And then as trust is developed, you go to human on the loop and so on and so forth. **[00:25:46]** Samir Khushalani: So, again, it all goes back to what… which industry you're in, what regulations you're subject to, what level of trust and capabilities you've got, but that is kind of the, you know, the way I… at least I see things progressing, in this particular space. **[00:26:03]** James Meads: Thanks, Samir. And that really just accentuates the point, don't try to run before you can walk, doesn't it? With the three different steps that you've got there. Tom, have you got any experience of that from your career or have you seen anything since then? **[00:26:17]** Tom Lynch: Yeah, so I… well, first of all, I love, the archetypes. I think that's a great way of thinking about it, and I… it does sound like there needs to be an evolution. And even, you know, even prior to AI, **[00:26:29]** Tom Lynch: When I've seen businesses and procurement organizations trusting each other, building **[00:26:35]** Tom Lynch: a collaborative environment, that's when there's success. And when there's mistrust or lack of alignment, you know, we all have experienced, that's usually when things, you know, struggle. And so now we've got another agent. **[00:26:50]** Tom Lynch: that we have to build trust with, so I think it's a great… **[00:26:55]** Tom Lynch: it got pointed out by the poll, that's what a lot of us are struggling with, is how do we trust this new agent? And now you have a three-party. **[00:27:06]** Tom Lynch: effort to gain trust, and I think what we'll see and discuss further is there are ways to do that and build that trust with Agentic capabilities. **[00:27:20]** James Meads: Thanks, Tom. Terry, any parting thoughts before we move on to the third question? **[00:27:25]** Terry Nadeau: You know, just, you know, working off of Samir's comment in regards to AP, because I think that's a great example, I think most of… of you guys would. **[00:27:37]** Terry Nadeau: and have experienced just significant issues, right, from a payables perspective. And you think about **[00:27:44]** Terry Nadeau: The amount of time… I could talk about Johnson Control. We used to spend a significant amount of time resolving and trying to pay, right, our suppliers, which, A, is not value-add to the buyers, nor is it good for the relationships between your supplier partners, right? And so… So this whole notion of, you know, on the loop and really **[00:28:09]** Terry Nadeau: Starting from that, you know it doesn't seem tactical or it doesn't seem that strategic from the perspective of what you're going after, but what I think you guys will find out is that it's going to free up time right. That then allows you to do the more strategic stuff in your business. And again, I'm big into people and stickiness and creating a world-class organization. **[00:28:33]** Terry Nadeau: It just really reinforces, right, that, **[00:28:37]** Terry Nadeau: you know, that this is the place you want to be, right? As you deploy this, I think you're gonna find it, you're gonna be an employer of choice for procurement professionals. **[00:28:49]** Samir Khushalani: And Terry, when you did this at JCI, did you find a very tangible improvement in, if you will, the morale and the turnover? Was that kind of reflected in that as well? **[00:29:00]** Terry Nadeau: We lost it. Yeah. Yes, Samir. We lost a significant number of people, right? I mean, we, you know, from my retirement, I mean, we were just getting into that, but I would, I would say on the flip side of it, we lost a lot of good people because of, you know, relationship issues, just time you had to spend keeping supplier shipping product, because of issues related to. **[00:29:25]** Terry Nadeau: Exactly the example you have right. Ap has all these invoices that come in. They get you know if there's not a match. They sit in a pile and they don't get touched until it becomes a big issue for the business. So supply. **[00:29:40]** Samir Khushalani: Complaining, and stuff like that. **[00:29:42]** Terry Nadeau: Yeah, but the team like talking to you know the team there, which I still do. I mean it's just it is creating a huge tailwind for these guys in terms of morale and just you know. **[00:29:57]** Terry Nadeau: The results that they're starting to deliver because where they get to spend their time, you know, it's it's a big deal. **[00:30:06]** James Meads: Yeah, I saw that too as, I mean, my bank, I worked for Kimberly-Clark for a number of years and we had similar problems with payables and it wasn't because the company had. **[00:30:16]** James Meads: a bad morale in terms of paying suppliers on time, it was just that the whole vendor master data and accounts payable process was completely broken. It goes back to your point, who owns vendor master data? Because in theory, it's finance and accounts payable, but when there's a mess, and accounts payable don't pick up the phone. **[00:30:35]** Terry Nadeau: Thanks. **[00:30:36]** James Meads: Who the suppliers call, right? It's not the accounts person. Yeah. Okay, so we've talked a lot about the technology already. We're going to move now to the second half of the discussion to focus more on the people side, because technology and people have to work together. So, as AI agents take on more work. **[00:30:58]** James Meads: Let's talk now a little bit more about operating model. Tom, how should procurement structure roles, skills and how you measure success? You know, what does good look like? How does that change? **[00:31:10]** Tom Lynch: Yeah. **[00:31:10]** Tom Lynch: Well, I think, you know, I guess you're incorporating some of the previous answers, so… **[00:31:16]** Tom Lynch: If an organization is working on the tail. **[00:31:20]** Tom Lynch: Using Agentic capabilities, to focus on that and move… and shift the people's focus, as Terry mentioned, to more strategic **[00:31:29]** Tom Lynch: partnership with business, you know, negotiating terms, supply security, specifications that are going to enable your product or service to improve and drive improved ROIC. So I think, you know, that's obviously the dream, is… **[00:31:46]** Tom Lynch: Shift people to the more strategic partnering roles that both… both business and procurement and the supply chain. **[00:31:55]** Tom Lynch: want to have happen. So, that's if you have people working on the tail. Then there's also the case where you don't have enough resources or focus to work on the tail, and so this **[00:32:08]** Tom Lynch: Capability now coming into play. **[00:32:11]** Tom Lynch: Gives you some savings you never had before. But I think the key thing is, I would call it shifting to the right, is moving your team's focus and capabilities to partnering with the business. **[00:32:25]** Tom Lynch: and driving business value is now enabled. And I think the other thing that AI offers, I think Terry mentioned it. **[00:32:35]** Tom Lynch: New people coming into the organization. **[00:32:38]** Tom Lynch: can get up to speed much quicker with much better information. You have exposure to the world of suppliers that… and a lot more detail about them, and the feedback on those suppliers. And so, if you come in as a new category manager, because, you know, things happen, succession happens. **[00:32:58]** Tom Lynch: The business that may… **[00:33:01]** Tom Lynch: You know, business partner who's been maybe in the business for 10, 15 years can now trust that the category manager can come up to speed much faster. And I think that's a great value that AI brings to the table. **[00:33:15]** James Meads: Yeah, I love that. I'm actually doing some work with a procurement tech startup or scale-up now that deals with chemicals and raw materials sourcing, and that's actually one of their sort of marketing messages, that if you've got a category manager with 20 years of experience that retires. **[00:33:32]** James Meads: Then then how do you replace all of that in-house knowledge? And you're absolutely right. It it it it allows a new person that has existing category management skills to step into that new category and and be productive much more quickly than would have previously been been possible. Yeah, 100%. **[00:33:52]** James Meads: Terry, Samir, any thoughts on this one around operating model? **[00:33:57]** Terry Nadeau: Well, you know, I… I'm building off the maturity model thought process that Tom started. You know, I think depending… if you're… **[00:34:05]** Terry Nadeau: You know, in a less mature organization or department, right? I I think you're gonna do exactly what Tom said. You're gonna start getting after things you never were able to get after because you didn't have the resource. **[00:34:19]** Terry Nadeau: Right? I mean, I think we can all I don't think any of us ever get the resources we think we need to get the job done. Right? So you tend to you tend not to focus in areas that are less important just because of the resources. **[00:34:32]** Terry Nadeau: But I think as we move across that spectrum, and as you become more and more mature, I actually see procurement organizations being smaller. **[00:34:43]** Terry Nadeau: in much more strategic, right? I… I see the organization spending a significantly greater amount of their team working with their key internal customers, be it engineering, product management, the C-suite, right? And you're actually gonna… **[00:35:03]** Terry Nadeau: You know, you think about… **[00:35:05]** Terry Nadeau: skill sets you need. Relationship management's important today, but to me, it's gonna become more and more important as we go forward, right? And… and what… what I think really… **[00:35:18]** Terry Nadeau: Helps, you know, the organization in general is that the communication and the understanding of where the business wants to go becomes much greater because of the time and effort you're able to put into, you know, meeting with customers and or key supplier partners. **[00:35:37]** Samir Khushalani: I couldn't… **[00:35:38]** James Meads: Yeah. **[00:35:39]** Samir Khushalani: I couldn't agree with you more, Terry. I think, you know, the… as the organization becomes more strategic, clearly some groups will, and maybe this is the elephant in the room, but some groups will become smaller, you know, especially the operational buyers, the accounts payable analysts. **[00:35:54]** Samir Khushalani: those folks. And, you know, they could be repurposed to doing other things, but if agents are doing most of the work. **[00:36:01]** Samir Khushalani: The role for those folks would be more directing agents, rather than actually executing the tasks, right? **[00:36:08]** Samir Khushalani: And to your point about the category managers, I would see them really transitioning to relationship managers, focusing on their suppliers, the strategic suppliers, as well as their internal customer, just to foster closer relationships with them. **[00:36:23]** Samir Khushalani: You know, we've done a survey at McKinsey, and we found the skills that are needed **[00:36:27]** Samir Khushalani: Of all the different skills, the ones that become most sought after in the future will be judgment and critical thinking. Because you just can't accept what an agent's, I mean, what an LLM tells you. You've got to have the judgment to discern, interpret, and, you know, interpret the information in context. **[00:36:46]** Samir Khushalani: There was one other part of your question, James, that I think at least is just worth touching upon. I won't go into it in too much detail, but it's this whole notion of measures of success, right? Because **[00:36:57]** Samir Khushalani: If the roles are changing, the skills are changing, agents are doing the work, then clearly we need to establish different measures of success. **[00:37:06]** Samir Khushalani: And that's important, because right now, there's a lot of… **[00:37:10]** Samir Khushalani: debate and discussion about AI not delivering the impact. Well, are we measuring… are we creating the right incentives to be able to maximize the impact that AI can provide? And this is where I like to talk about both adoption and absorption. **[00:37:27]** Samir Khushalani: So, adoption is… it measures behavior change. Are people using agents? **[00:37:33]** Samir Khushalani: Absorption measures work transfer. Has the work actually shifted to AI? **[00:37:38]** Samir Khushalani: Because, candidly. **[00:37:40]** Samir Khushalani: Just because you have high adoption, but if you have low absorption. That means people are using the tools, apps, and agents, etc, but the underlying work stays manual. The impact will be subscale. **[00:37:53]** Samir Khushalani: So as we define the performance measures, we need to make sure that it is actually impacting the work by allowing agents to do it in an end-to-end manner, versus us just, you know, kind of dabbling with it, but still relying on ourselves to do most of the work. **[00:38:11]** Samir Khushalani: And this also then will come with the right level of governance models, decision rights, oversight mechanisms, all of that has to be revisited, right? And that… **[00:38:21]** Samir Khushalani: In fact, there was a Harvard Business Review article that just came out, I think it was on… in August of this year, which said a critical success factor is even when you're doing pilots with some of these agents. **[00:38:34]** Samir Khushalani: you should start thinking about the governance models during the pilot, not later on. So, you know, start thinking about it from the get-go. And I think that's… that's really sage advice. **[00:38:46]** James Meads: I love all three of those answers, so I'm gonna… I'm gonna ask you a simple yes-no question, and I'm just conscious of the time, so, I mean, if you've got a couple of points to elaborate on it, that's fine, but let's not dwell too long on it. **[00:39:00]** James Meads: Based on what all of you have said, if I summarize. **[00:39:04]** James Meads: We're gonna see fewer operational roles. Procurement teams are gonna get smaller, but probably not drastically smaller. **[00:39:12]** James Meads: And there's going to be a lot more need for critical thinking, people that can interpret results and that can communicate effectively and eloquently to the wider business. So, you know, marketing and internal comm skills, if you will. **[00:39:26]** James Meads: So… On that basis, do you think that the category generalist That is… **[00:39:34]** James Meads: very proficient in soft skills, and that perhaps can do a little bit of building and tinkering with AI agents. Are they going to be the new hot property in procurement? **[00:39:48]** Samir Khushalani: I would say yes. **[00:39:52]** Samir Khushalani: You're muted, Terry. **[00:39:58]** Terry Nadeau: Sorry about that. Yeah, I would agree. I absolutely think, **[00:40:03]** Terry Nadeau: Those skills are going to be very, very important in the future. **[00:40:08]** Tom Lynch: Yeah, I think as you pointed out, Samir earlier, the soft skills with external and internal customers. **[00:40:16]** Tom Lynch: The people that can do that well, they'll have the tool of AI to help support them in the other areas. **[00:40:22]** James Meads: Yeah, that's really interesting. It validates what I think as well. And I think historically, procurement leaders had a tendency to employ technocrats, didn't they? And I think that's really gonna have to change if we look at the way this is going. **[00:40:38]** James Meads: So I think on that point, it kind of neatly segues into the final question that we had. Terry, I'll put this one to you first. **[00:40:47]** James Meads: Do you have any specific advice that you would give your fellow CPOs as they navigate this rapidly changing landscape with AI at the forefront? **[00:40:59]** Terry Nadeau: Well, I mean, I was kind of chuckling a little bit, James, at the poll when you mentioned quality of data, and surprised that that isn't a bigger issue. I mean, to me, my advice would be, you know. **[00:41:14]** Terry Nadeau: Start, make sure that the quality of your data is super, super good, right? I mean, I I think getting your data. **[00:41:23]** Terry Nadeau: Quality where you want it, thinking through where you want to take Agentic AI and and and how you want to use agents and creating that vision and getting buy-in across the organization. I think those two areas are going to be super important for people that are just embarking on this and and going after it. **[00:41:46]** Tom Lynch: Yeah, and I might add to that, one bit of suggestion is… **[00:41:51]** Tom Lynch: Samir used the word pilot, is to… **[00:41:54]** Tom Lynch: Pick, because trust, and we saw it in the poll. **[00:41:58]** Tom Lynch: Is such a concern. I think maybe the roles… **[00:42:02]** Tom Lynch: Got very little answer, but maybe that's a blind spot and underestimated. But I would pilot it in a an area that's target rich. **[00:42:10]** Tom Lynch: So you can get some results early, and you can learn… **[00:42:14]** Tom Lynch: What you don't know, and build that… build that credibility within the organization. **[00:42:21]** Tom Lynch: And then… and then spread out that capability. So maybe that's a general business practice for something new and innovation, but I would apply it here to… **[00:42:32]** Tom Lynch: To learn as you go, and then expand. **[00:42:37]** Samir Khushalani: I think from my standpoint, you know, when I talk to CPOs, one of the questions I'm often asked is, should I be a first mover or a fast follower? And my advice would be lean into AI. The biggest risk today isn't making the wrong AI decision. **[00:42:55]** Samir Khushalani: It's making no decision what everyone else is learning. **[00:42:59]** Samir Khushalani: And I would go so far as to say that leaders of tomorrow **[00:43:03]** Samir Khushalani: will be the 1st movers of today. **[00:43:05]** Samir Khushalani: And again, referencing this HBR article. **[00:43:09]** Samir Khushalani: they talk about how procurement is, frankly, the ideal function to lead the Agentic shift. So as CPOs, you're in a prime position, because you're leading a function that is, really, it has work that's structured, the results are measurable. **[00:43:26]** Samir Khushalani: they're tangible, and we have a lot of judgment-intensive tasks. So there are enough number of use cases that you can use AI for. Just pick the ones where you can deliver the highest impact, like Tom was saying, but then just jump into the, jump into the AI pool. You know, you don't learn swimming by just gazing at other swimmers, right? **[00:43:47]** James Meads: Yeah, I love that. The classic consultants two by two, right, Samir? It's which one lands in the top right hand corner that's high impact but low disruption or low cost. Yeah. **[00:43:58]** Samir Khushalani: By the way, the answer to that can be different for different companies. Again, another question… the next question, then CPOs will say, so which use case do I start with? And it's like, that's not for me to say. Let's look at what problems are you trying to solve, which ones are the highest priority for you, and then we, you know. **[00:44:13]** Samir Khushalani: We then, determine which are the use cases we want to be focused on. **[00:44:19]** James Meads: Yeah, 100%. It's only when you completely map someone's source to pay or source to contract process that you can… that you can confidently answer that question. Yeah, okay. Awesome. Alrighty, so before we move on to the Q&A. **[00:44:35]** James Meads: And there's about 5 questions that have come in here. **[00:44:40]** James Meads: So, we'll tackle those one by one. I've got a couple more polls that I want to go through, just to get your final, final thoughts after listening to our three panelists wax lyrical about AI. So, let's launch this first one. **[00:44:57]** James Meads: Oops. **[00:44:58]** James Meads: Sorry about that, hang on. There we go. So, where do you see the greatest opportunity for AI in your procurement function? I want you to select your top priority. Don't think about it too much, just… **[00:45:13]** James Meads: Select whatever comes into your head first because that's going to be the one that's front and center. **[00:45:19]** James Meads: And there are… there are 4 options, plus the other. **[00:45:31]** James Meads: Interesting. Okay. **[00:45:35]** James Meads: Looks like about half of you have answered, so if you haven't yet, please just **[00:45:44]** James Meads: Grab what comes into your mind and pick one of these, and then I'll give it another few seconds, and then I'll flash up the answers on the screen. **[00:46:03]** James Meads: Okay, I'm going to end this now, just in the… **[00:46:08]** James Meads: Just in the name of time. **[00:46:10]** James Meads: So this one's pretty well distributed, ladies and gents. So we've got spend analytics and category strategy was top. **[00:46:19]** James Meads: Which is perhaps unsurprising, because there's a lot of grunt work that goes into that that is not necessarily core procurement work, I would say. Then contract analysis and compliance was number two, again. **[00:46:31]** James Meads: You know, perhaps not surprising there. Sourcing was close behind it. And number three was sourcing and RFP, RFQ management. And then finally, supplier performance and risk is number four. **[00:46:42]** James Meads: and purchase invoice **[00:46:45]** James Meads: process… purchasing and invoice processing, and number 5. And thankfully, no one picked the other, because I always hate when you've got other in polls, because it's… the risk is that everyone picks that one, so thankfully no one did. **[00:46:55]** James Meads: Alrighty, and then final poll, and then I promise we'll get to Q&A. **[00:47:03]** James Meads: So when does your organization expect to start using AI in its sourcing or RFP RFQ process? Again, just pick one single answer. **[00:47:14]** James Meads: Be an optimist. **[00:47:15]** James Meads: I know it's Friday afternoon, but be an optimist. **[00:47:20]** James Meads: And then we'll move to Q&A. **[00:47:22]** James Meads: We've got just over 10 minutes, so I think we've got plenty of time to tackle the questions. **[00:47:39]** James Meads: Okay, this one's really interesting. It's quite well distributed, I have to say. **[00:47:47]** James Meads: Just give you another 10 or so seconds. If you haven't answered it, then fingers on the buzzers, just, select one of those. **[00:48:00]** James Meads: Awesome. Okay. **[00:48:03]** James Meads: So, what did we get? So, about a quarter of you are already using it. **[00:48:09]** James Meads: Another 21% of you plan to use it within the next three months, then another 12% in four to six months, 21% in seven to 12 months. There's nobody that's saying we're at least a year out or two years out, perhaps maybe unsurprisingly because… **[00:48:28]** James Meads: If you're that if you're that far out, then the reality is you probably haven't yet decided, or you don't have any current plans, which represents about the final. **[00:48:37]** James Meads: 20, 21% of you. Interesting. **[00:48:40]** Samir Khushalani: think. **[00:48:41]** James Meads: Jim. **[00:48:41]** Samir Khushalani: Yeah, it's interesting, if you were to add up some of those, what we're seeing is at least over 50% of the folks **[00:48:48]** Samir Khushalani: planned to do something with AI in sourcing in the next 12 months. **[00:48:52]** Samir Khushalani: Which is, you know, which is an important take away. **[00:48:58]** James Meads: Yeah. **[00:48:59]** James Meads: Indeed, yeah. Yeah, it's good that we're seeing… it's good that people have plans to use it within the foreseeable future. **[00:49:06]** James Meads: Yeah, okay, I'm gonna jump to Q&A. So we had 8 questions, we probably don't have time to go through all of them, so I'm just gonna pick out the ones that are most relevant. So there was a question from William that came in pretty early on, but that was… **[00:49:20]** James Meads: This was more around, sort of, immigration legislation, which is not **[00:49:25]** James Meads: directly impacted to AI, so I think I'm gonna skip that one, William, sorry. So… **[00:49:32]** James Meads: I'm going to take one from Fides Pacheco, first of all. So, is there any annual procurement amount that warrants to shift from AI from traditional process, considering the cost of transition and implementation? So, I think Fides is talking about spend here. So, is the… **[00:49:50]** James Meads: Is the guys a threshold in spend that you think is necessary to justify the investment? I think is his question. **[00:50:00]** Tom Lynch: My initial thought is less on the spend, but the nature of the… Spend, than the amount, meaning. **[00:50:10]** Tom Lynch: How much qualification needs to be done? How rigorous are the specs and integrated? So, you know, on the spectrum of direct strategic materials, indirect. **[00:50:23]** Tom Lynch: commodity. Again, back to piloting and getting to know how to use this capability. **[00:50:30]** Tom Lynch: I think it'll be easier… **[00:50:32]** Tom Lynch: to start with those things, that are closer to indirect commodity, MRO, that type of thing, is kind of my impression. And that, you know, that can be a lot… depending on your organization, that could be a lot of **[00:50:45]** Tom Lynch: A large amount or it could be a small amount, but I think more of the nature of it will give you some early successes. **[00:50:53]** Samir Khushalani: You know, I think there are some inherent assumptions based on this, which I would actually like to tackle, frankly, which is… **[00:50:59]** Samir Khushalani: the reference to cost of transition and implementation, one of the biggest costs of the transition are all our organizational people, capability building, mindset change, etc, as opposed to the technical aspects of it. So, you know, most of us **[00:51:15]** Samir Khushalani: or at least some of us have probably come from the ERP era and the SaaS era, and we've seen technology that takes 12 months, 24 months to implement, and all the money on integrations and so on. Those are things of the past, right? The time to impact with agents, especially when you start out, it's actually **[00:51:33]** Samir Khushalani: much faster. We're talking about, in weeks, you can prove out the value, and if there isn't value, you fail fast, and you move to the next use case. So, I would… honestly. **[00:51:43]** Samir Khushalani: pressure test in, you know, yourself, the assumptions that you might be making around the cost it takes, if you will. But the second thing I think is. **[00:51:53]** Samir Khushalani: There's an inherent assumption here that only some things will shift to AI, and I would also challenge that. **[00:52:00]** Samir Khushalani: Kind of going back to my archetypes of human in the loop, human on the loop, and human above the loop, archetypes. **[00:52:08]** Samir Khushalani: AI is involved in every one of them. They're just different levels of AI autonomy. So I would actually say 100% of your spend should be managed by AI, but in some cases. **[00:52:18]** Samir Khushalani: Humans are doing a lot more work, in other cases, they're doing less work. That should be the aspiration, and that's what you should be thinking about. Now, to get there, of course, it's not one big step, right? You go step by step, but ultimately. **[00:52:32]** Samir Khushalani: all roles will be impacted by AI, and all spend will be impacted by AI. That would be my assertion. **[00:52:40]** Terry Nadeau: Okay. **[00:52:41]** James Meads: interested. **[00:52:41]** Terry Nadeau: I would agree. **[00:52:42]** James Meads: Guys, yeah, Terry, Terry, just really quickly, and then I'll move on. **[00:52:44]** Terry Nadeau: Yeah. No, I would agree with Samir, you guys. I think it's gonna be, a lot… **[00:52:52]** Terry Nadeau: less costly of a move for an organization than launching SAP or some other huge ERP, right? And so, in most of the things that you gotta do, whether it's data quality or, process, right, re-looking at the processes, I think are things you would do anyways, right? I mean, and things you need to do. **[00:53:13]** Terry Nadeau: And so, **[00:53:16]** Terry Nadeau: Yeah, I think it's not as big of a lift as people think, and I think the ability to do this almost in a self-funding way is going to be super doable. **[00:53:31]** James Meads: Okay, thanks guys. So I'm gonna move on to the next question now quickly from Kimberly. What is the best way to get practice using AI in procurement? And she also asks, is there a best AI software platform? **[00:53:44]** James Meads: For the for the sake of time, I'm going to take the second part of that question and just say, Kimberly, that really depends on. **[00:53:50]** James Meads: A lot of different things. I hate to give a consultant's answer, but the right AI platform for Terry back when he was CPO at Johnson Controls is going to be very, very different to if you're a $200 million business with a really immature procurement team of maybe one or two people. So I think the second part, just do a little bit of market research and get to know the market a little bit, and then I would say you'll get a lot more clarity there. **[00:54:14]** James Meads: I have a podcast called the Procurement Software Podcast. If you want to get a bit of feeling on on what's out there. But the first part of the question, what's the best way to get practice using AI in procurement? Anyone want to want to tackle that one? **[00:54:28]** Samir Khushalani: I'd like to actually tackle the second part of the question as well, to kind of maybe **[00:54:32]** Samir Khushalani: It may be similar, but slightly different, which is, I would… I would assert that I don't think you should think about there being just one single platform. **[00:54:42]** Samir Khushalani: I think the likely procurement function of the future is going to be. **[00:54:46]** Samir Khushalani: enabled by an Agentic ecosystem, so it's a multi-vendor landscape where you might be working with multiple platforms, as opposed to the… this monolithic source-to-pay suite that used to, you know, dominate the last decade, or last 15 years, if you will. **[00:55:04]** Samir Khushalani: Because now, with more open protocols like A2A and MCP, etc, you have the ability to have agents from different providers **[00:55:13]** Samir Khushalani: collaborate with each other, and you have the notion of swarms of agents working together. So I do think that's another paradigm that needs to be challenged as to whether it has to be a single platform, or could it be more than one, or multiple? Maybe not 100, obviously, but, like, multiple ones. And then, again, to James's point, do your market research, etc. **[00:55:32]** Samir Khushalani: There's a lot out there that allows you to determine what's right. **[00:55:37]** Samir Khushalani: one way, in my mind, and I'll hand it to Storm and Terry here, but to get practice is, you know, as you're trying to evaluate this, pick a couple of design partners, frankly, from some of these tech startups, and work with them **[00:55:52]** Samir Khushalani: to define these workflows that are tailored for your needs, and then… and that's a really good way, you know, there are lots of courses, lots of universities are offering free courses. Yeah, you can learn the fundamentals, but nothing **[00:56:05]** Samir Khushalani: I can't think of a better way than to kind of, again, jump into the shallow end of that AI pool, work with a couple of design partners, and, you know, apply AI to **[00:56:16]** Samir Khushalani: Reimagine, redesign, reinvented workflows, not just automating what you currently have, of course, but that's a great way to learn, if you will, on the job. **[00:56:27]** Tom Lynch: Yeah, I would just add, you know, first determine where you want to prove success and build credibility within your own organization, and once you've decided that's the right spot, then I would pick **[00:56:40]** Tom Lynch: the software, the applications that would best help you there. That would be my recommendation as a… **[00:56:48]** Tom Lynch: In my experience. **[00:56:49]** James Meads: Okay, okay, thanks Tom. Next question from Mark. Do we seem to be blinded by the hype of AI and are we forgetting the general rules of common sense we've always used before we had new technologies? That's a great one. **[00:57:08]** Samir Khushalani: I think judgment and critical thinking are… that's why they've become that much more important, because you just can't accept everything that AI is telling you. You've got to have the ability to **[00:57:18]** Samir Khushalani: go back to first principles and pressure test, and make sure that you are interpreting information in the right context. So. **[00:57:27]** Samir Khushalani: But I don't think that's mutually exclusive. It doesn't undermine the role of AI. **[00:57:32]** Samir Khushalani: It just means that you just need to make sure that you're leveraging those skills in addition to all the efficiencies that AI provides for you. **[00:57:43]** Terry Nadeau: Yeah. **[00:57:43]** Tom Lynch: Yeah, at least you're. **[00:57:44]** Terry Nadeau: I think we maybe are undervaluing it more than overvaluing it, you guys. I mean, I honestly think that this will drive more change in every aspect of our lives than almost any other technology that's come along. **[00:58:01]** Terry Nadeau: And, and that's why I think the advice of, you know, jumping, you know, how do you learn to swim? You swim, right? Like Samir said. So I think it's just… **[00:58:12]** Terry Nadeau: You know, jump in, and there's a lot of ways to play with it, whether… **[00:58:18]** Terry Nadeau: You know, you partner with somebody like Samir said, or or you know, decide where you can add value, right, based on your organization and partner with your IT team. But spend it, spend an hour, you know, spend an hour a day or an hour every other day really using it. **[00:58:38]** Terry Nadeau: You know, not looking at it, not reading about it, but really trying to use it. **[00:58:45]** James Meads: Okay, gentlemen, thank you so much. We've got about a minute left, so I'm gonna have to wrap things up here. Tom, Terry, Samir, any final comments and thoughts before we, before we wrap up, and wish everyone a nice weekend. **[00:59:03]** Tom Lynch: Nope. **[00:59:03]** Samir Khushalani: It's. **[00:59:04]** Tom Lynch: Not to vermiculate. **[00:59:04]** Samir Khushalani: It's an exciting time, and if we were having this webinar 6 months from now, maybe even 3 months from now, we might be talking about new advancements, and so, you know, just keep watching the space as well. **[00:59:18]** James Meads: Okay, awesome. Gentlemen, thank you so much for joining me today and being great panelists and giving some awesome answers. I hope those of you who have been in the audience, you've taken some value from this and I'm really sorry that we couldn't get to all of the questions that were in there. We did allow 15 minutes for Q&A, but they came thick and fast at the end. So wishing you all a great weekend. Check out Zinnit if you want to learn more about the product and they'd be very happy to hear from you. **[00:59:43]** James Meads: If you're curious what they're doing. Until the next one, take care wherever you are. **[00:59:47]** James Meads: And, bye for now. **[00:59:50]** Tom Lynch: Thank you.