← 返回

2026.09.30 | TB | Consumer | AI in US Food Distribution Expert

原标题：AI in US Food Distribution Industry – Performance Food Group, US Foods & Sysco – Former C-level Executive at Performance Food Group Co

#consumer #transport

Third Bridge · 2026-09-30

Client-led • 16 Sep 2026

Companies: SYY US, PFGC US, USFD US

Expert Former C-level Executive at Performance Food Group Co

Agenda

Questions

00:05:27] We're trying to understand the US food distribution industry, particularly companies such as PFG [Performance Food Group], US Foods and Sysco. What do you think of PFG at the moment?

Expert [00:06:55]: PFG is still dominating in market growth on the independent restaurant side of the business.

Client [00:16:17]: You once talked about the cost-cutting opportunity. How big is this opportunity? How could it reach its peers' margins? How much of a disruption, or what would it take, to cut out those costs?

Expert [00:16:52]: It's going to be significant. Yes. It's basically eliminating middle management. You're taking some of that nuanced strategic out of the opco where you're taking away some of that independence and you're getting more to a consistent model, but you're really going to be able to cut out a lot of cost.

00:19:14] What other cost-cutting opportunities does PFG have that stand out to you?

Expert [00:19:21]: Yes, but I don't think they're going to do it, at least not for a while. The one I'm thinking of is sales. The reason why they're winning in sales is because they have a different strategy than did US Foods and Sysco. The salespeople, they had lost their motivation to actually grow. We'll put it that way. That created a significant brain drain in both of those organisations as far as their ability to grow sales. It also took away a lot of the motivation and desire from the sales force to continue to grow. PFG has the opposite model. They are not interested in doing that. They keep hiring salespeople left and right. The sales expense is just a burden of growth is the mentality. I'm going to make a comment, it works. It's expensive, but it works. That's why they've been winning in the marketplace for many years now.

There's a lot of expense with that. You're going to make more and more money. You continue to grow there, your bottom line is going to grow faster than your selling expense will go up.

Client [00:21:54]: Are there any other cost-cutting areas that come to mind, or are those the two biggest?

Expert [00:22:08]: Those are the two biggest by far.

00:22:26] Why hasn't PFG cut the middle management yet?

Expert [00:22:43]: Honestly, I think they saw that coupled with what happened, when they made those moves, it created disruption. Not the sales thing. It created disruption. There's no way to do it smoothly. When George Holm was running the ship, the focus was always on top-line growth, top-line growth, top-line growth. Yes, this is in our back pocket, but right now, let's just keep growing and their expectation is to surpass Sysco at some point in time. That's the reason why they haven't done it. I don't want to call it fear. It's avoiding a known problem. The problem is not permanent. The problem is not everlasting, but it is a problem. You will have a disruption in your service level. You will have hiccups that are costly, but those do get worked out and you will have the benefit down the line. It's that transition. You need to have some pretty powerful finance people there to be able to manage all the operating companies across the country.

00:24:24] Although cost-cutting opportunities with AI are a common question when thinking about PFG, US Foods and Sysco, we’re thinking more specifically about US Foods, given it's a tech leader. How do you think about the opportunity for AI to drive efficiencies in this business model? Where are the biggest opportunities?

Expert [00:25:05]: I would tell you the biggest efficiencies can be gained in transportation. If you let AI actually structure your delivery routing every day based on what's the most efficient way, time, mileage, fuel, you win. There's a cost for that. You have a USD 10,000 per week customer that wants their delivery before 6:00, but if you let AI tell you what you're going to do, you got to make sure that you're going to be at that customer before 6:00 because those are the types of disruptions that will cost you to lose customers. US Foods and Sysco might be willing to actually say, "You know what, we'll redo our routing, so I could be there at 5:30 in the morning. We'll take your business back. We'll take your business from PFG." That's a problem. The whole equation on that is customers. Where is the biggest savings from AI? It's going to be in transportation because there's no route that runs the most optimum way it possibly could. AI could do that like that. It could tell you instantly. That's one thing. That's from a cost savings perspective.

Warehouse, you can get some cost savings with efficient slotting, that you slot your warehouse in what will be the most optimum way based on your volume of product that you have, based on the mix of the product you have, you slot it so it can be the most effective to build pallets for your selectors. That's going to be another savings potential, not as big as transportation, but that's going to be a big savings. Where I think US Foods has been leading in a long time is marketing, taking customer-specific data. You could have AI that finds category or product voids that individual customers are not buying from you and work with the vendors for marketing to those specific customers so you can penetrate. I think that's an avenue that can be very profitable. I would tell you, PFG might be working on it, Sysco might be working on it.

Client [00:27:47]: How is transportation routed today? Are the software programs for routing as good as they could be, or can AI make them better and more effective? How does that work?

Expert [00:28:18]: Yes, and I think most of them are all still using Roadnet.

00:28:30] Is Roadnet a software program that does routing?

Expert [00:28:31]: It's a software. Yes. It was actually created by UPS. It used to be a part of UPS, but it actually became its own company. It is a routing tool that allows you to build routes relatively quickly with knowns like the USD 10,000 account that needs to be there before 6:00. Then you have windows and you can have customer ratings, A, B, C, D, and it will put together routing the best it can, but it's not the most optimum. That's pretty laborious. Routing takes several hours for people to actually build the day's routes. The bigger the company, you may have two people doing that. The reality is you get stuck. In Roadnet, you're just going to follow the previous day or the previous same day. You have your Monday, Thursday deliveries, so your Monday route will be pretty much the same the next week. It will be pretty much the same next week until somebody goes in and manually changes it. AI could do that instantly. That's what I'm saying.

Every day would be a slightly different route, but it's optimising. If you can shave 50 miles off a route, that's 10 gallons of fuel. It depends on where you are, if you're in California, that's a lot of money. It's not USD 80, 50 times eight, maybe USD 400 per route and the bigger companies are going to be sending out 100 routes a day. It can be significant, really significant savings. Customers are not flexible in this industry. They're not. Until they become an outspoken person that goes out and sells how good your company is, they're not flexible. You need to meet their demands. Now you do get some customers that eventually will become your salesman. What I mean is they'll sell you to other companies, but that's not most of the customers. That's some of them.

Client [00:30:56]: Could it be an issue if you're changing the customer drop-off times every week and they might be opposed to that?

Expert [00:31:05]: Very much so, because a lot of times, they'll staff somebody to be there to actually receive the product. If you're moving that time all around, it's going to cost them money. Then it impacts your relationship with them.

Client [00:31:17]: By slotting, I assume, you mean putting things in the warehouse.

Expert [00:31:37]: Right, exactly. Depending on the size of your warehouse, if you're a Sysco that builds warehouses more than big enough for what they need, most of their products are put in full-size pallet picks. There's no manual work that happens. When you receive product and you get a pallet of it, you put it right in the pick slot or you put it in a reserve slot and it just goes down. When you get to some of the smaller companies, US Foods, PFG, there's a lot of those to where they'll have some area of full pallet, they'll have half pallet, they'll have third pallets and they'll have hand stacks, where they'll have three-level hand stacks. Having everything in the right location based on its cube and volume is what's going to save you time from the receiving from the putaway to the selection. You can also get efficiencies on your pick path based on the cube and weight and density of items and what needs to be on the top and the bottom of a pallet, you can do that as well.

00:32:48] How often are slotting positions changed? Do you use AI? Would it be a one-time improvement? Is it monthly?

Expert [00:33:09]: Continually, because volumes change so much. The volume changes so much based on seasonality. If you sell a lot of ice cream places, that ice cream stuff, it's going to be slotted differently in the summer than it will be in the winter. I'm just using that as an example. It almost needs to be continually that AI is touching the slotting. Like I said, you're going to get some benefit from your selection, but you're going to get more benefit from your receiving and putaway to where you can reduce the number of inbound people you have. You'll reduce damage, you'll reduce missing product when you do that from the night perspective, but the night is your outbound. Your day shift, the inbound crew would actually require a lot less time. You would be looking at reductions in force on day.

00:34:19] You said AI helps with marketing. What is the application there? What does it do from a marketing perspective?

Expert [00:34:45]: What I'm trying to push through is I want AI to go out and look at a batch of customers. I don't care how many is in the batch, pull their menu offline, and then based on that, identify the products that they're buying to make the products that they're selling, bump that up against what we're selling them, find, okay, we have these 72 accounts that aren't buying sour cream from us, but we know they're buying sour cream. That gives us a tool. Go to the vendor, "Hey, we got 72 prospects to convert sour cream to your brand. Give us USD 3 per case for what we sell on them for the next six months. We'll lower the price, we'll go in, we'll switch it to your product." That gives you an opportunity to now market, penetrate and grow.

I'm saying, from the old Kraft days, customers they had, they had, okay, it was going to be Italian, what type? Then there might be seven different types of Italian restaurants or Latin or bar and grill, pub food. There was enough of a breakdown of the customer that you could identify what you needed to sell to them, what they would be buying, in general, it's not always the same.

00:36:18] How advantaged and far ahead are the big players? How much better are they positioned, and how much quicker can they adopt AI vs an average mom-and-pop regional player?

Expert [00:36:47]: I'm going to tell you, I don't know that they'll be faster, seriously. It depends on the individual operator at an independent or what you call mom-and-pop and how big their vision is. I'll give you an example. Implementing a different corporate card and AP payment bolt-on to our ERP system. That is giving AI on an AP side. I can just write the prompt, "Show me my repairs and maintenance on my trailers by highest the lowest," and it will show me which trailers cost me the most money. That's a concept.

I think some of the old dogs in the independent market are going to be slow to adopt. I will tell you who will win the AI battle. I can tell you that. It's the company that uses AI to improve productivity, not to replace cost. If a company is solely out there to try to cut cost, they'll cut cost, but there's going to be a benefit that those are taking a different approach. They'll lose competitive advantage. If you disrupt the computer to do it all, you're going to miss something.

00:39:32] Why do you think the US Foods Pronto Delivery service is so quick and instant?

Expert [00:39:50]: Same-day delivery. Yes, and Sysco has something similar in major metropolitan markets as well. It makes it really hard to compete with them. Let me rephrase this. I'm going to put it this way. It depends on the operator, the restaurateur. If the restaurateur is a good restaurateur, they understand their inventory, their cash flow and their cost, and the bigger your drop, the lower your cost will be, whether that's from US Foods, Sysco, PFG or an independent. The more I'm stopping the truck to deliver you, the lower your cost will be per item. If somebody can't manage their inventory well enough that they need same-day delivery, yes, there will be those that will take advantage of it because I got to tell you, half the people that own restaurants don't know what they're doing. They either know how to cook, they know how to present, but a lot of times, they don't understand all the financials and the cash flow concept of it. It's a nice to have, but is the win really there? I got to tell you, US Foods is not delivering that at the same cost they would deliver 150 case drop that's scheduled out the day beforehand.

Client [00:41:26]: Your first thought is that it makes it hard to compete with that company. Can you elaborate on that?

Expert [00:41:41]: Yes. I'm going to give you an example. Now I'm in rural Central West Virginia. Our truck getting to Pittsburgh, it's three hours to Pittsburgh. The driver only gets 10 hours of drive time. As we're trying to grow that market, boy, it would be nice if we could load the truck Sunday night, deliver half the truck Monday, the other half Tuesday and he comes home. The problem is customers don't want to give you their order on Sunday for Tuesday, especially when you have a US Foods that can say, "Give me your order Tuesday, I'll deliver to you Tuesday." That's a hurdle to jump. Now that hurdle is ours because we have to grow our business in Pittsburgh so that we can put a shuttle yard and we drop trailers and then the driver delivers from there. That does create additional complexity for growing markets is what I guess I'm trying to say.

00:42:42] Under its new CEO, US Foods has been pushing out centralisation over the last few years. It used to be centralised but is now trying to decentralise a little more, just like PFG – not in cost and local management, but in sales and autonomy. It’s moving towards a full commission model that started a few months ago. What do you think of those initiatives? Do you think it could be successful and continue winning or accelerating market share wins? What are some risks of doing that? Do you think that's the right decision?

Expert [00:44:08]: I absolutely think it's the right idea because like I said, PFG proved it. It's been year after year, PFG has been taking market share. I think it's the right thing for US Foods to do. My concern is it will be a slow burn. Will people come back? It's been long enough that some of those people aren't even still in the game anymore. They're going to need to invest heavily in training and really, they got to do a lot to get it back. I think it's the right move. I'm very happy for them that they're doing that. It's not going to be a light switch.

00:45:10] If you were US Foods, what would you watch out for? What could go wrong, or where could there be hiccups or disruption?

Expert [00:45:28]: Lack of training. It's a totally different concept to getting a sales rep that's fully commissioned and how they sell to somebody who's salary plus bonus. It is a totally different. The people who are good at one aren't going to necessarily be good at the other.

Client [00:45:43]: Why is that?

Expert [00:45:52]: From a sales rep, you want somebody, who's going to push the envelope. When they're full commission, they're going to push the envelope to make as much money as they can. Without the training or the deep, long understanding of how that works, it's give to get. You can make as much money from a customer this week, but you price yourself out of the market, you won't have an order next week. It's a learning experience on knowing the customers. What are their sensitive items? Where do I need to be priced where? That's something that comes through training and experience. If they move more to a commission-based model where the sales reps are going to have more control on that, if they're not trained properly, they're either going to be giving it away or they're going to be overpricing themselves. Very few will price it properly to begin with.

00:47:46] PFG and US Foods both look interesting. US Foods is more exciting, given its technological leadership, which allows it to implement AI in a quicker and better way. It's also shifting its sales force to the likes of PFG. Do you have anything to add?

Expert [00:48:15]: They definitely have a bigger ceiling and definitely room for growth. I don't know if they're still doing their Scoop marketing. I don't know if you're familiar with that. I thought that was tremendous when they rolled it out. I thought it was a great concept, but I don't know. If you hadn't wrote it down, do research on Scoop, I would suggest you do that. Just giving you some additional tips, if you don't mind. They also do a better job of marketing their own private brands. I'm going to say that. PFG is trying to do it. They don't do it as well as US Foods does. US Foods has their own brands very solidified in the marketplace. That's another thing. Sysco doesn't really have brands.