Order management goes first in the agentic era | Ep. 10

Episode 10 – Order Management Goes First in the Agentic Era – Transcript

Matt: Welcome to Data Versus Commerce, where we explore the messy middle between database and doorstep. I’m Matt Johnson.

Floyd: And I’m Floyd Blaikie. Let’s dig in. Hello, Matt.

Matt: Floyd. It’s good to be back with you.

Floyd: It is, it is good to be back. Believe it or not, I don’t just podcast — although I think it would likely do the world a favor if that’s all I did. But in my day job, I lead the marketing department at Pivotree, and we’ve been working on a couple of reports. As part of what we do, we like to keep a pulse on the industries we serve, see what people are up to, keep tabs on the movers and shakers. That’s not the topic I want to cover today — we recently released our State of the Industry retail report, and I’m sure we can go into that with one of our retail experts on a future episode. But the one we’re about to release is the State of the Industry in Industrial Manufacturing and Distribution through Q2. I think it’s a really cool report, and I’ll tell everyone how to get it at the end of the episode. But I wanted to tell you the wackiest, coolest thing I think we found in this report: Swagelok has made a huge leap forward using AI for order management. They’re now handling 97% of order-to-cash through automated multi-agent AI — it’s lifting purchase orders, putting them straight into SAP. 97%! Have you ever in your life imagined that amount of order-taking could be lifted and shifted into an AI system?

Matt: It’s really incredible. And if anybody from Swagelok is listening, I’d love to talk to you — not to sell you anything, I’m just genuinely curious how you did this. Because if those numbers are true, the amount of change management that had to happen, the amount of governance that had to happen before you could realize those numbers, is just staggering — and really a case study for everybody across industrial. So good on you guys, that’s really impressive stuff. And hey, I think that’s where we’re heading.

Floyd: Absolutely. And maybe it’s a little scary for the people who’ve traditionally been at the counter, or on the phones, or receiving the faxes, or whatever we used to do back in the olden days. In fact, I think most manufacturers and distributors are still there — from what I’m seeing in this report, only about a third of manufacturers and distributors are really on the leading, bleeding edge of this kind of order management workflow. It’s interesting to see how that differs from what’s going on in retail. The agentic era has come for retail too, but it’s coming in the front door — the shopping, finding the products, even adding to cart. You can do the entire purchase process through AI, although I don’t think a lot of people are doing that much right now. But the agentic era in retail is about finding and buying the product from the front end of the store. In industrial, though, there often isn’t a storefront to put a robot in front of, right? That’s what you’ve seen, I’m guessing, in IMD.

Matt: Yeah, there’s so much to say about this, and I love that the headline of our report is order management goes first — it’s leading this AI push in organizations, and for very good reason, because you said “back in the olden days” — actually, it’s not so old. This is happening today. You’re going to have a department of inside sales reps literally printing papers out on a printer, picking those papers up, and moving them across the organization physically. That’s just reality for most manufacturers and distributors today. That’s not uncommon, because the process of actually taking an order from quote to cash is extremely manual. They’re limited by legacy processes, and turning a ship like the Titanic isn’t easy. Changing the way you fundamentally handle an order is extremely risky — to your customer experience as well as to your internal teams. A factory can’t just shut down while you figure out how to go digital.

Floyd: No, and I think there’s still a pretty big cohort of B2B businesses trying to figure out how to get started. One of the stats we uncovered in the report was that 71% of B2B businesses do offer e-commerce to some extent, but that doesn’t necessarily mean it’s a digital experience from start to finish. You might be able to log into a portal and place an order, but it might still be somebody at a desk getting the printout and typing it into another system — it’s not necessarily a connected ecosystem. So you’ve got that reality: not all B2B has adopted e-commerce. The ones that have — only about a third are actually doing the full thing start to finish, and certainly not to the Swagelok extent. On the other hand, you’ve got analysts like the Gartners of the world saying that by 2028, AI agents are going to mediate more than $15 trillion in B2B spending, and 90% of B2B purchases will be AI-assisted. Does that sound real to you, or is this just the number-crunchers making things up for slide decks? What do you think?

Matt: I watched that same — I watched the video of that analyst, I think his name’s Daryl Plummer, if I’m wrong I apologize. When I first heard him say that, my jaw dropped and I thought, “What is this guy on?” I didn’t really believe it at first, but I’ll say this: I’m starting to become a believer, because I’m watching companies advance so quickly. Even in our own business, what we’ve done in six months would’ve taken six years — the innovation, how quickly you can experiment and adopt new processes and technology, it’s just crazy. Look at what Swagelok’s doing — you’d think that took years and years, but it probably took months and months, and that’s the difference today. With the right leadership, the right processes, you can very quickly change the way you do business. Is it believable? It’s believable. Is it going to be a stretch for a lot of companies in industrial? Absolutely, it’s going to be very difficult. One of the things our report points out is the gulf between those who are leading — on the bleeding edge — versus those who are sitting still. That gulf has always been there — there’ve always been digital leaders, like an industrial MRO distributor doing 60% of its revenue online compared to a not-small business doing maybe 1% online. That gulf’s always been there, but it’s getting wider faster than ever before. So yeah, I do believe it, because the big guys who own most of that market share are going to advance very quickly, and you’ll see billions in revenue through industrial being mediated by AI.

Floyd: Yeah, absolutely. I think electrical distribution is what we’re seeing on the leading edge — I don’t want to generalize too much, but we’ve got AI ordering and quote assembly from Graybar, that was another thing we covered in the report. If I’m a leader in the manufacturing or distribution space, it’s easy to look at those stories from Swagelok, Graybar, and other organizations that have rolled out really sophisticated AI-enabled buying experiences, and say, “That’s so far ahead of me, I can’t even think about that right now.” But it wouldn’t be a very good report if I was just repeating press releases — there’s other stuff we look at too, and one of those things is job postings. The strongest hiring signal of all of Q2 — that’s the time range we’re looking at in this report — was product data. One in seven companies we tracked, across over 700 companies in North America, are putting out product data roles. These are channel infrastructure roles — not web roles, we’re not just talking about writing product content for the e-commerce storefront. These are people getting hired to feed other companies’ systems — PIM roles, engineers to build AI agents. I think that’s the signal that the middle of the pack is working to catch up. Are you seeing that too, with new hires and job postings at the organizations you’re talking to?

Matt: I am seeing that, and you’re right, there’s a mad dash right now to catch up. Everybody in leadership understands that the place where things are going to break down in back-office AI automation initiatives starts at the product data level. There’s customer data too, which is a big concern, but yeah — product data is essential to making the order management, the quote-to-order automation work, because that’s where the human intelligence has always lived in these companies. You have to understand the catalog to handle a customer’s quote request or enter an order correctly. If it relies on the human tribal knowledge that lives inside these inside-sales departments, you’ll never be able to fully realize that kind of automation that AI promises — and not just promises, but is delivering for some of the market leaders.

Floyd: All right, let’s put some heat on a guest who’s not even here to defend themselves, which is very unfair. Recently we had the founder and CEO of Tromml on the podcast, Lauren, and she was talking about how her organization is using AI to help field sales do relationship-based selling in a better way. I feel like these findings are at odds with that in some way, because if order management is the first place to go, and the person at the desk is going to be replaced by AI, does that mean relationship selling is no longer important? I feel like that’s something B2B does not want to let go of very easily.

Matt: No, it’s not. What I was thinking about just a few minutes ago, when we were talking about these roles coming up — there’s a demand right now for product experts, a demand for people who can manage and own that data. Really, the thought that came to mind when I hear “hiring” is this changing demographic. I know there are executives sitting in boardrooms right now saying something like, “That’s okay for Graybar, that’s okay for Sonepar or Swagelok — we’re going to win with relationship, that’s going to be our competitive advantage, they can automate all they want, our people are our value proposition.” That’s great, but it’s also a real danger, because of the changing buying demographic — that next-generation purchasing manager, purchasing director, is adapting faster than you’re adapting. What I mean is, they’re no longer as constrained about who they can purchase from — they’re simply looking for the fastest, cheapest, easiest way to manage their supply chain. If that’s the case, relationships start to break down — not in terms of importance, but in terms of the value you’ve always relied on. Managing the order is no longer where the value is in face-to-face sales. The value is being available, being the human face to the business — but what happens behind that relationship is really what purchasing managers care about today, and will increasingly care about. They don’t have time to be calling in, emailing, following up, asking questions — that’s just not what this next generation expects.

Floyd: Yeah, I think you’re right — I think the bar to earning the relationship just gets a bit higher. You don’t have to force someone into a relationship by making a human the gatekeeper — you have to earn it in a different way, which I think is interesting. I’ve been in manufacturing for 12 or so years, and I remember in those earlier days hearing, “Oh, Millennials are coming and taking over, Millennials are going to be the biggest purchasing group,” and now it’s Gen Z. Us Millennials are getting old — I’m pretty sure I need bifocals, and I’m solidly a Millennial. The research supports that too: 67% of B2B buyers now want a rep-free experience, which should be a little scary for people whose relationship skills are covering some sins in the product data department. And 45% of B2B buyers have actually used AI in a recent purchase. This isn’t a theory, this is already happening — those dang kids are taking over our buying processes, get off my lawn. So even if the order counter becomes an API, you still need the relationship — you just have to earn the right to start that relationship by having a good experience up front.

Matt: The tough truth in all of this is you can no longer rely on one or the other — you’ve got to bring both. I think of Grainger — whenever I think about what best practice looks like, I think about Grainger, because they do it better than anyone. They’re everywhere physically — human faces on job sites, in the manufacturing facilities, they show up, they embed themselves. I’m talking field reps — they know their customers, they’re always there adding value. But guess what — the ordering, the catalog, all of it, is best in class. So you get the best of both worlds: you have that relationship, you know your Grainger rep, but you can also go online and order, and make sure you’re never missing the parts and materials your business requires. That’s a very tough value proposition for most mid-market distributors in particular to compete with, but they’ve got to figure out a way. Usually what’s missing is what’s behind the rep, because they’ve relied so long on just the human touch — automating the order management, making sure questions are answered without having to go to their rep, is where most mid-market distributors are really struggling.

Floyd: Yeah. And I don’t think, based on what we’ve seen building out this report, that there’s any way to opt out. You can be the biggest AI hater, you can say, “I don’t care what you say, I’m just going to continue to rely on my relationship-based sales organization” — but you can’t actually make that choice, because machines are ingesting your data whether you want them to or not. You don’t have to sign off and say, “Sure, ChatGPT, you’re allowed to look at my product catalog.” You can certainly block it, but unless you’re really interested in preventing sales, I don’t know why you would. Your products right now are being read, compared, and recommended by AI buying bots — Gartner’s calling them “custobots,” I’m not sure I love that, but I don’t think I get to make that decision. Being behind on digital doesn’t keep you out of agentic commerce, it just makes you lose. Am I wrong on that?

Matt: No, you’re not wrong. There’s just a nuance I’d add: when your end buyer is researching, they’re going to use the industry’s best research tools, because they want to do their job faster. They’re turning more and more to AI to do that research. They’re going to go to the manufacturer first, as the authoritative source of product information — and manufacturers in competitive categories are either showing up, because their data is machine-readable and structured for those LLMs, or they’re not even in the consideration set. They’re either informing the buyer, who then takes that information to whatever approved supplier or distributor they have and makes the purchasing decision — and this is an untapped area for manufacturers. They’ve always known they need good product information online for their customers, but what they’re starting to realize is that the way their products also appear downstream on distributor channels impacts the way these AI models look at their brand. If you’re invisible on distributor websites, AI doesn’t consider you a viable option in terms of supply and the ability to actually get that product.

Floyd: Right — doesn’t matter how good your product photos are, doesn’t matter how good your counter person is, you’re just out of the consideration set entirely. Scary, but I think there’s a lot of opportunity here too.

Matt: Yeah, and it’s not like anything has fundamentally changed, because business is business, and we’ve always done this. The manufacturer has always needed to own the authoritative source of product information and get that information where it needs to go. The challenge today is just that it’s different, and there are more places it needs to go. It used to be about that beautiful studio photo and the nice brochure, the PDF. Unfortunately, a lot of manufacturers are still in that world — it’s all about the marketing brochure, and they haven’t taken the data in those marketing assets and put it into machine-readable formats, or even digital-commerce-friendly formats. That’s step one. Step two is actually getting it downstream — getting it into the hands of your distributor sales reps and into those marketplaces, so you can actually sell.

Floyd: Absolutely. Big PDF’s gonna come for us, I can feel it. But I want to make sure people get to see this report — and I’m actually really excited, because this is something we’re doing now on a quarterly basis, to see what changes or how this evolves between Q2 and Q3, so we’ll be looking at that later in the year. If folks want to see the report — depending on where you’re listening to this — you can go to datavscommerce.com, we’ll have it on our LinkedIn page, you can search us at Data Versus Commerce, and we’ll put it in the show notes and transcript on the site as well. Anything you want to plug, Matt? Any cool stuff?

Matt: Well, thank you for pulling this report together — it’s really exciting to be able to use AI on our own to do massive amounts of research that would’ve taken me weeks and weeks. So Floyd, thank you so much, it’s incredible, this report is really exciting. We’ll be updating it — we’re already looking forward to some events later this year, and we have some exciting new solutions coming out that speak to the very issues this report brings up. So if you want to talk about the report, I’m always game — you can DM me on LinkedIn, or send an email, and we’ll get together.

Floyd: Oh, that’s great news, because I did put your email address in the report — I wasn’t going to tell you and just have it be a surprise, but glad to have your blessing on that. Thanks, Matt.

Matt: Yeah, you got it, Floyd. This was great, and looking forward to the retail report as well — let’s get Dan on and go through that one.

Floyd: All right, let’s do it. See you next time.

Matt: Okay, see ya. Thanks for tuning in to this episode of Data Versus Commerce. New episodes drop weekly. So if you’re responsible for any part of how products get from a database to a doorstep, subscribe now on Apple, Spotify, or wherever you listen.