Tom McFadyen on How AI is Reshaping B2B Commerce

speaker-0: AI is rapidly reshaping how B2B buyers discover, evaluate, and purchase products. And for distributors, that shift is happening faster than most are prepared for. From AI-powered search and answer engines to the rise of agentic commerce, the rules of digital engagement are being rewritten in real time. To break it all down, I'm joined on the MDM Amplified podcast by Tom McFadden, founder and CEO of McFadden Digital.

which is a firm that has been helping distributors and manufacturers navigate digital commerce for nearly 40 years. In this episode, Tom shares his perspective on how we got here and where things are headed, including why answer engine optimization and generative AI search are emerging as critical channels, how distributors should be prioritizing AI use cases based on ROI, and where many companies still fall short, especially when it comes to data and connecting digital investments to real business impact.

We also look ahead to what's next from AI driven product discovery and automated ordering to the growth of agent to commerce and what B2B buying could look like by 2030. It's a wide ranging practical discussion on how distributors can better position themselves for an AI first commerce landscape. Enjoy. Tom, thanks for joining the MDM Amplified podcast.

speaker-1: My pleasure. Thanks for having me on, It's great to be on your podcast and ⁓ active member of the NAW organization as well. speaker-0: Absolutely. McFadden Digital has long been servicing wholesale distributors for almost 40 years now.

So I think that any long time MDM reader has at least some familiarity with the company, but I'm sure that there are plenty that might not know what you guys are all about. So Tom, can we start off with an intro to first who you are in your role at the company and then who McFadden is in a nutshell and its role in B2B distribution? speaker-1: Certainly, as the company name implies, I'm pretty tightly intertwined with the business. started it 39 years ago and really servicing mostly B2B customers, manufacturing and distribution.

the early or late 80s, we were doing document management, early form of content management, document and technologies like that for those that recall those in the days of the 80s and 90s. That morphed into content management, web content management in the early 90s. And then, In 1997, we started building out e-commerce sites. So was Content Plus Commerce with one of the early players in that space, ATG, our technology group, which got bought by Oracle.

And we've continued doing e-commerce for 29 years. It's been our core bread and butter. We added marketplaces as a solution offering or a different type, a more advanced type of commerce about 10 years ago as one of our go-to-market offerings. And we've built dozens of marketplaces as well.

with companies like Miracle and MarketPlace. Now in the last few years, we've really expanded into the AI offerings around commerce. It's interesting how it's somewhat of a full circle because AI relies on data or content, and that's what we started with 39 years ago. I'll talk a little bit more about that, but the importance of data towards AI and answer engine optimization, GEO, and running your whole business around AI.

speaker-0: Well, 29 years in e-commerce, it's almost hard to believe that we're almost three decades into McFadden being part of the digital commerce sphere here. When we bring in AI into this discussion, depending on who you ask, some people say AI has been around for 40 years. But by some other definitions, it's only been around as far as agentic AI for the last, let's call it three to four years. So I'm curious to ask from your definition or your perspective, how long would you say that AI has been part of digital commerce?

speaker-1: 20 years at least. some of the early technologies like automated recommendations, you might also like YAML or collaborative filtering. Amazon started this. There were other technologies that had that.

ATG even had that about 20 years ago where you look at one product and it automatically based on machine learning will recommend cross cells, up cells, substitute cells based on that product. We've also seen a lot of AI in ⁓ demand forecasting, inventory management, pricing optimization, dynamic pricing. These types of solutions have been around for at least 10 years in some areas, becoming more more popular. Chatbots, of course, becoming smarter and smarter.

But AI has had some element in those solutions for five, 10 years. But going back 20 years for some of the automated recommendations, I'd say. Delivery route optimization has certainly been optimized with large ⁓ AI models in the fulfillment space. speaker-0: That's just a great clarification there.

And I want to bring up that McFadden is sponsoring an NAW webinar on May 7th that, as you mentioned, is all on the topic of answer engine optimization and generative engine optimization. So, AEO and GEO and how distributors can really win in this age of AI search. I'm moderating this webinar and this is a topic that really didn't exist just three to four or so years ago. And now it's suddenly a critical topic for any distributor engaged in digital commerce.

So whether listeners are getting this podcast before or after that live event on the seventh, what makes this topic so vital for distributors right now? speaker-1: Well, one way of looking at it is it's yet another channel, ⁓ but a very important and extremely ⁓ critical ⁓ channel, especially as we see more agentic commerce coming on board. And I'd add to AEO, GEO, ACO, or agentic commerce optimization as yet another type of that, which build upon the first generation of SEO, ⁓ which started 20 plus years ago with the beginning of the web searches.

So as a new channel, ⁓ distributors have evolved from print paper catalogs to ⁓ or branch counter sales to e-commerce or maybe EDI before that. And then other channels came on Punch Out or RFQ, procurement systems, vendor managed inventory, then other marketing channels, social and sometimes even ordering through social channels. AEO, GEO are other channels in which buyers can find information about your products. And that can be in short little snippets like intro energy optimization.

So when Google Gemini, for example, you type in something into the search bar and you have the option from Gemini to give you a quick little answer which is ⁓ referencing you. Or maybe it's in a ⁓ page long or multi-page deeper research being performed by one of the AI LLMs pulling back. You want your business, your products, your distribution channels to be cited in there as solutions. Moving forward with that same, again, all of this relies on having the right data in the system upon which content and data, again, this technology has been around for decades around managing content.

You need to make sure your content is properly structured, properly accessible, properly available so that you are cited by the answer engines, the LLMs. As this evolves to become more more popular, the agent at commerce. The ACO or agent at commerce optimization is going to make sure that you also have product availability, shipment costs, tax calculation, fulfillment, delivery, timeframes, etc. All that information also available, which is more than just the content.

It's the commerce enablement of that same product, which is obviously very important for distributors making the transaction. speaker-0: Well, it is a lot to manage right now for any distributor involved with e-commerce. It's just in this age of AI, data is the name of the game right now. And there's such a heightened focus on it.

And it's just great that there's partners like McFadden that can really help distributors navigate all this. And part of that is that we are thrilled to have McFadden return as a sponsor of our Shift Conference that is almost here. It's coming at May 12th through the 14th in Denver. You guys are sponsoring it.

our digital cohort breakout. In those breakouts, they're really designed to foster small group discussion with similar non-competing peers. From McFadden, there's going to be Jeff Mikos and Trey Oliver there on site to engage with the attendees and really help spur discussion about their digital transformation and commerce readiness in this AI era. Are there any particular nuggets that you're priming Jeff and Trey to keep an ear out for?

speaker-1: A lot of it's really around ROI. There are hundreds, literally hundreds, and we've actually documented hundreds of use cases for AI in a business. The challenge for a distributor or any organization is to prioritize which initiative you're going to work on first. it getting more traffic to your site?

Is it the AEOGEO? Is it the conversion rate? Is it the pricing optimization? Is it fulfillment optimization?

Is it maintaining inventory levels appropriately? And part of that process of determining which AI solution you want to implement first is what's your problem. So identifying where you have, where you'll get the best return on an investment. It's a lot of companies have somewhat, you know, had their day of let's do a little pilot and try this and try that.

Whether or not it's really solving a business problem. There have been a lot of technology for the sake of technology. The real important element is identifying where you can receive the most return or optimization from investing in an AI initiative. That's what often take something from just a pilot into production, where you're gonna start seeing returns and actual, the finances of it make sense.

speaker-0: That makes sense. Well, you and the team, as you alluded to, just published your fourth book titled AI Best Practices for E-Commerce, which delivers more than 125 practical use cases of how B2B organizations are applying AI. And there's so much AI research being done right now, including plenty by MDM and NAW. But can you touch on what the core purpose of this book is?

Maybe what sets it apart? And then perhaps teasing one or two of the key findings that help illustrate why listeners should go out and get it. speaker-1: Yeah, sure. Thanks for thanks for mentioning that.

It's our fourth best practices book starting with 20 years ago. We wrote ecommerce best practices, then we had a marketplace best practices and a follow on marketplace book. So this we just finished this year. We actually wrote over 300 use cases, but we had to trim it down to get into the biggest book we could print 520 pages, 8 1⁄2 by 11.

So we ended up actually open sourcing the content, so it's now available on a wiki ⁓ available for free to everybody. So AI Dash best. bestpractices.com, AI best practices with a dash between the three words, .

com. We got up to 500 use cases in there. But the point I want to make is again back to prioritizing or understanding ⁓ which of those use cases is going to make the most sense. We have some ROI graphs which talk about different phases of different value streams.

Again, putting some structure into all of these different potential applications of AI. We have over 500 use cases. There are over 50,000 AI tools out there. ⁓ How do you pick a tool?

How do you pick a use case? How do you pick a part of the business that's going to deliver the most ROI for the investment? And we broke that down into different value streams within the organization. So the starting value stream is commerce.

So market, sell, fulfill, service. Three pretty discreet, understandable phases of the commerce value stream or life cycle. And we also have different value streams for the product development lifecycle, meaning the products that a company makes, software development lifecycle, human resources and recruiting life cycles, et cetera. So different value streams.

And we also structure it into a pretty interesting org chart, which we have a poster of this. Happy to send that out to anybody who's interested. Just shoot a note to info at mcfadden.com.

We'll send you a free poster, which structures this into, it's a 36 inch by 24 inch poster. ⁓ hundreds of use cases by the org chart. So across levels of CXO to VP to director to manager and within each of those roles within an organization, what are the top half dozen or so use cases for each of those dozens of different roles within an organization? So I'd say that the main theme that we're trying to get across here is how to put some structure around many use cases and many tools.

and pick those that are going to make the most sense for you. can either be along the value streams upon which a business operates or across the organization structure of a company. speaker-0: That's great, because it's one thing to identify hundreds of use cases in AI, but when you can segment like them based on role, that makes it much easier for someone to navigate specifically to the applications that are pertained to them, rather than having the page through and sift through everything to find the ones that are a little more applicable to what they're doing.

Well, switching gears here a bit, we at MDM usually try to avoid negative sounding language when it comes to the questions that we pose to the industry and to experts that we interview. But once in a while, I think it does provide a strong discussion point. So I'm going to do that here with you with something that I think fits very well with your expertise in asking, what are distributors doing wrong when it comes to digital commerce? Or maybe not painting with such a broad brush.

What at least are then what are distributors doing? wrong most frequently on that front. speaker-1: I'd say, as we mentioned earlier, data or content underlying it. one of the good things is that AI can now help enrich that data or better structure that data.

So the foundation upon which commerce and AI works is catalog, obviously, is a big element of that. It's usually, I think of it as three major categories. You have your products, you have your customers, and you have your orders. And really, it's a subset of those ⁓ three major sets of data, which you an underlying database may have a thousand different tables or different structures for managing all that.

But ⁓ PIM or the other MDM master data management, master data management for the product catalog, for customer information, really structuring that appropriately. you can get the best example I mentioned earlier, cross-sell, up-sell, substitute-sell of products between your items, between customer segments, between industries that you go after. making all that available. Fortunately, as I mentioned, AI can now enrich that.

So you can use AI to go through product PDFs from a manufacturer to enrich the distributor's online catalog database. You may get a 40 character, very terse description of a product in the name, but on your commerce site or for your counter sales reps or your call center reps, you may want to have a much lengthier description with some more. Narrative text that's probably SEO optimized for your site, AO optimized as well. And enriching that can come from the source product, the manufacturer of the product, or it can come from some engines, which again, read PDFs.

They look at an image and say, this is blue, this is red, this is purple, et cetera. So AI can certainly help a lot in that enrichment area. So I'd say data quality is an important foundation for both commerce and for AI enabled commerce. speaker-0: Yeah.

And when we ask distributors, what's their biggest hurdle or ⁓ hesitancy to diving in with AI, to no surprise, their immediate response is it's data. Our data just isn't good enough. They don't feel comfortable or confident to layer AI on top of their current data situation. So yeah, great point.

speaker-1: And that's part of the services we provide, but there are also lots of commercial tools available that help with that. And the AI tools can help with that. I'd say a second ⁓ element in which that could be improved is change management and really the process of deploying commerce across an organization. So again, going back to that example of the cross-sell upsell, substitute sell, not only is that helpful to have that in a commerce system, but once that that functionality, that data, that knowledge is available.

You can also make that available to your call center folks. If somebody calls in and says, hey, I ordered this, it didn't come, what can you do for me? And they have a substitute sale offering. Same with your ⁓ branch counter sales reps.

If they can look on their system and see, here's a better version of that product that you might want to consider, or this will go well with that product, want fries with that, or you want to supersize your Big Mac, make that capability available. also in chat in your automated chat bot, why not enable the chat bot with that knowledge to cross sell, upsell, substitute sell, for example. speaker-0: McFadden as a company, like we touched on early on, has been serving this industry for as long as digital commerce has essentially existed.

So you and your team have really seen the full history of e-commerce and marketplace evolution here in B2B. And I figure that must also give you a pretty good sense of where this is all headed. So what are you most interested to watch over these next, say, three to five years or further out? in terms of what B2B digital commerce might look like in 2030 or even further out.

speaker-1: Agenta Commerce, think, is one of the biggest change and agents in general, know, having an AI ⁓ agent do something for you on your behalf, much like a travel agent would do, would book tickets for you or travel for you. An agent that does activities for you, like ordering products for you. I think it's actually, even though lot of the buzz around that is in the retail or B2C space, I think there's a better use case and applications of it in the B2B world, especially distribution.

⁓ be surprised if the smartest AI could predict what my wife's going to come out of a department store with in a retail setting. But it's much easier to predict ⁓ when a manufacturing or distribution site ⁓ needs MRO replacement based on MTBF failure points of machines or in a discrete manufacturing process engineering. When you run out of supplies, what your inventory needs to be, predicting the additional resources or products that you need for your factory line. A lot of that is much easier to predict with AI in a B2B setting than on a consumer trends.

So I think there's a lot of opportunity for replenishment and other types of ordering through agents B2B happening. And we started seeing some of that evolution with procurement systems, the Arriba, Coupas of the world, but really AI enabling that with a Jentic to make it much more seamless and interoperable between ⁓ organizations. And we're starting to see a lot of standards come about. ⁓ an agreement on that despite some initial competing standards for ATTENTIQ ordering.

⁓ So I think that's going to be a great opportunity for a lot of B2B distributors and others in the B2B space. speaker-0: It's such an interesting time to be alive right now in B2B commerce. So we touched on the new book, but besides that, what has McFadden really been up to recently? I know that you guys are very well represented at numerous industry events, both in-person and virtual.

So what else do you guys have going on? speaker-1: really focusing on the AI enablement and enhancements of commerce. And that can be around the AEO-GEO, it can be around conversational commerce, the chatbots, it can be around ⁓ automated order entry. you know, taking in emails or texts or images and creating orders out of that.

Really, there are so many different ways in which AI can enrich ⁓ commerce capabilities. So, I'd say that's our latest wave of trends from going from document management to content management to commerce to marketplace to ⁓ AI and really seeing more of the agentic as the future. still think marketplace will be a big ⁓ part of that as well. ⁓ Not just ⁓ the Amazons of the world, but we've seen from Granger and Zorro, a first party ⁓ commerce company like Granger decides to start up a ⁓ marketplace Zorro and scales it to a billion dollars and 10 million products in a decade.

That ability of the marketplace model to scale, I think will become even more powerful now that agents can do the ordering and the marketplace business model really places that those processes and communication and business rules and relationships of the ecosystem of buyers, the middle person, the marketplace operator and the sellers. That whole ecosystem, think the marketplace model is very similar to what a JanTech commerce is enabling. It's just a different technology. speaker-0: Well, Tom, that's really everything I have for you, but do you have any final thoughts or parting words of wisdom on this general topic of digital commerce and B2B before we sign off here?

speaker-1: The future is very bright and AI, think, is making it even brighter. And McFadden Digital would love to help anybody on that journey wherever possible. speaker-0: Perfect. Tom, thanks again for joining the MDM Amplified Podcast.

speaker-1: Thanks, Mike. It was great to be on your MDM podcast and also to be a member of MDM for so many years. I really enjoy all the educational content that MDM publishes. Thanks so much for that.

speaker-0: Thanks for listening to the MDM Amplified Podcast, which is all about elevating voices of distributor service providers. You can find our entire episode library at mdm.com slash podcasts or wherever you listen to podcasts.