Why does a tiny disruption shut down your whole supply chain?
Thirty years of optimizing for speed and minimum cost removed the slack, so one late shipment cascades into stockouts across a business. That is the problem Mohannad Gomaa sets out on this episode of The Artificial Intelligence Podcast, and his answer is that resilience has to be built into supply chain strategy deliberately, because the efficiency model we inherited is fragile by design.
Gomaa's line is that "Supply chain models are very brittle to the tiniest disruption." A port closes, freight reroutes, a single input arrives hours late, and the effect travels through manufacturing, inventory and distribution. The just-in-time model that made sense for three decades assumed nothing would go wrong. The lesson of 2020 was that something always does, and the constraints arriving now keep proving it.
What replaces it is not stockpiling for its own sake. Inventory is working capital, and Gomaa's point is that the question is how to hold just the right inventory at the right time while still hitting the customer service levels each product actually needs. Some products can go out of stock without much damage. Others put real revenue at risk. Those two groups should not be planned the same way.
How do you align supply chain strategy with the business?
You start with the business strategy and translate it, rather than letting procurement and planning set direction on their own. Gomaa's framing is that supply chain is a function of the strategy across products: how much you stock, how reliable your suppliers must be, how close or far they should sit. Resilience is one of those strategy choices, not an operational afterthought.
That translation shows up in concrete decisions. Whether you multi-source, single source or sole source is a strategy question with cost, lead time and risk consequences. Long-term, mid-term and short-term planning all have to line up with the same demand signal, across manufacturing, inventory, procurement and distribution.
What does AI actually do for demand planning?
It improves the demand signal and the allocation decision, not the strategy. Point of sale data and demand sensing are what let you decide which product mix belongs in which market and which store, instead of forecasting a total volume and leaving the placement to whoever runs operations.
Gomaa describes a client whose forecast stopped at the volume level. Operations then decided where the product went, without intelligence behind the choice, and the result was misallocation and transfer costs as stock was moved between markets to fix it. Historical data alone will not solve this, because trends move market by market and product life cycles turn over. Allocation has to be driven by what is actually selling where.
How can a smaller business compete with Amazon's logistics?
By buying the capability instead of building it. Amazon multi-sources even simple items, so no single supplier failure stops the line, and it balances inventory across fulfillment centers to hit delivery windows that keep shrinking. A small e-commerce brand or a single store cannot replicate that network.
Gomaa's answer for smaller operators is direct: if you cannot do it yourself, you buy it. That means third-party and fourth-party logistics providers, shared services, and consulting firms that give you visibility and predictability of demand you can act on. Some small businesses simply sell through Amazon direct to customer. Others can build the muscle over time, starting with one supplier and diversifying the portfolio as they learn to find and negotiate with more.
Should you move manufacturing closer to home?
Near-shoring and friend-shoring have already started, and the deciding factor is usually not the factory. Gomaa's point is that building the plant is not the problem. The question is whether your suppliers follow, whether they can deliver raw materials at a workable cost, and whether electricity, tariffs and free zone arrangements keep the total landed cost sensible.
Apple's announced manufacturing facility for glass is his example of what volume and know-how can make possible. But manufacturing footprint decisions also depend on where you ship: if most of your product goes international, the right location may not be the home market at all.
How do you respond when customer demand shifts for good?
You shape demand rather than waiting for the old behavior to return. Gomaa calls this demand shaping: driving promotions, working the right mix of products and services, and respecting seasonality instead of fighting it. A bar does not have to carry ten beer taps. It has to carry the ones that sell.
The payoff is working capital. Holding less of the wrong inventory frees money to invest elsewhere in the business. That is the same logic as the resilience argument: the goal is not maximum efficiency or maximum stock, it is the right position for the demand you can actually see.
What to do when your supply chain is optimized for cost instead of resilience
Decide your sourcing strategy before you optimize anything. Choose deliberately between multi-source, single source and sole source, and set the reliability and distance requirements that follow from it.
Diversify your supplier portfolio. Gomaa's advice is not to optimize purely for cost. Build relationships that give you the right price and lead time agreements, and partners who can flex when something breaks.
Push your forecast below volume level. Use point of sale data and demand sensing to decide which product mix goes to which market and which store, so placement is not handed to operations without intelligence behind it.
Buy the capability if you cannot build it. Third-party and fourth-party logistics providers, shared services and consulting firms can supply the visibility and predictability a smaller business cannot staff.
Set service levels product by product. Some items can go out of stock; others put revenue at risk. Plan inventory against that distinction rather than a single blanket target.
The uncomfortable part is that you cannot predict which shock is coming. You can only decide, in advance, how much slack you are willing to pay for.
Original show notes
We think of AI as living in the cloud — but it runs on physical chips, factories, and supply chains that break in the real world. Jonathan Green talks with supply-chain expert Mohannad Gomaa about the logistics of AI: why modern supply chains are dangerously brittle, how AI helps spot trends and de-risk sourcing, and why resilience — not just efficiency — is the strategy that matters now.
Key Takeaways: • Modern supply chains are brittle — the tiniest disruption can cascade into shortages • AI helps you spot trends and decide where (and from whom) to source • Forecasting that stops at volume level leads to costly misallocation • Diversify your supplier portfolio; don't optimize purely for cost • Resilience, not just efficiency, is the supply-chain strategy for the AI era
Notable Quotes: "Supply chain models are very brittle to the tiniest disruption." — Mohannad Gomaa "It's that resilience part — one of your supply chain strategies — that's what really matters." — Mohannad Gomaa
Connect with Mohannad Gomaa: LinkedIn: https://www.linkedin.com/in/mohannadhazemgomaa Website: https://mgostrategy.com/
Enjoyed this? Follow The Artificial Intelligence Podcast and share it with someone running operations or a supply chain.
Auto-generated transcript, 4,579 words. Timestamps link to the moment in the episode.
the Logistics of Artificial Intelligence with today's very special guest, Mohannad Gomaa. And Mohannad, I'm excited to have you here because we so often think about technology as living inside of the TV or inside of the internet. Like we think the cloud is in the sky, but it's actually we're learning now that all sorts of things we've learned last
few years that like computer chips matter, right? If there's a computer chip storage. remember 20 years ago when I was working for a computer company. The company that made almost all the LCD screens in the world was in a storm. And suddenly, everyone had a bet, like nobody could make laptops for three or four months. And we suddenly realized how important certain little pieces of infrastructure Mm. if there's a certain resource that every phone needs. we're starting to, I think for the first time,
appreciate supply chains once again, which is really important. And with artificial intelligence, Yeah. causes these effects, these current effects with supply chains. I think it really matter. In fact, water usage and there's computer chips and all these data centers being built and then you build this data center and then two years later the chips are obsolete and so then you have to ship them off and sell them to someone else or sell the AI processing power to lower rates. So there's these supply chain issues that
are starting to really matter Hmm. and that's why I'm so excited to have you here today. So maybe you could start by just telling us big picture why supply chain matters Thank you. and then we'll kind of dive into some specifics. Yeah, absolutely. And thank you so much for having me, Jonathan. Appreciate you. I think supply chain matters because of everything that you just mentioned. It's a function of how you first of all strategize on the right strategy across
products and how you build how you really translate the business strategy to supply chain strategy so that the supply chain can definitely align with the business wants to do. And that reflects in examples like you mentioned and how much you stock, how reliable your supplier should be, how close or far they should be. It's that resilience part, you know, as one of their supply chain strategy that matters. But then it matters because everything that you do on the planning side
is very critical on the demand side and on the supply side on the demand side to improve your forecast and be able to align the right demand signal to the company so that you can produce against it and align your supply chain strategy, be it manufacturing, be it inventory, be it procurement, be it distribution logistics. All of these functions need to align to what you really want to sell, and that's definitely key. For a long time, maybe the past 30 years,
we've really focused on optimizing for speed and efficiency and minimizing cost, Mm-hmm. which is why you kind of want to have nothing in your warehouse, Hmm. right? It's always like, I just want just enough for one day to arrive today. And while this was a really good strategy, we discovered, especially in 2020, that when there's any shock to the market or a delay, Yeah. it can cause a cascade of problems. actually these super efficient supply chain models are actually very vulnerable or very brittle
to the tiniest change. Like if a port closes and they have to go to a different port and suddenly one thing's four hours late, well then it cascades to everything else. So how are you seeing people change what they optimize for? Are we starting to see a little bit more stockpiling, a little bit more like flexibility built into supply chains? Hmm Yeah, absolutely. And that's a great question. Absolutely. The flexibility now is very important. I think everybody, you know, got caught in
that covid and was a lesson learned. But also a lot of constraints that popping right now is definitely changing the strategy. It's not about, you know, just in time always. It's about how agile you are, how resilient you are. Hence how smart you are. It's about how you have just right inventory in time is what's important because your inventory is your working capital and very strategic. So how do you really optimize it for sure? But how we align to the right customer service,
you know, levels as well is very important. Some can go buy it and do not really have to. aligned to really restricted customer service levels. . But some definitely cannot get into stock out scenarios where they would jeopardize a lot of sales and revenues coming in. So now it's all about how you really look at all those levers that control your inventory. about how you synchronize with manufacturing. How do you really plan ahead? So we always talk about the long-term planning,
the mid-term planning, and the short-term planning, which always involve inventory for sure, manufacturing definitely. So how do you really factor in the right amount on the working inventory or the web? is very important, but also on the finished goods side, you know, definitely. And this is, you know, part of the question is how do you really set up the right service levels again so that you won't run into those stuck outs and how you distribute it as well is very important across your distribution channels.
So different, you know, strategies, different network designs for distribution is going to really matter in terms of where you allocate your inventory to which market is becoming very important. Yeah, my grandfather was a manufacturer of women's coats. And one of the big challenges, this is 40s, 50s, was you could run out of mediums and have too many larges. So even something so small as that can make a huge difference where you... Yeah. Absolutely. have to, so all of these little things
that we kind of don't think about with manufacturing and we think that AI can solve every problem, but if you optimize for the wrong thing or you put in, if it has one bad piece of data or one miscalculation about how things happen Thank or how things are interpreted, it can be something as simple as how sizes means something different in Europe and America. The same numbers mean something different in Asia, the numbers mean something different. A large is not a large in every country. Yep.
these little things, what's really wild is like genes are measured in inches, but like two genes will both say 32 and they won't be the same size. Mm-hmm. So these little issues can really kind of magnify. Sure. And I think that one of the good uses of tools and artificial intelligence and research is kind of looking at, well, you can really look very good at the past, Hmm. but we sometimes over index the past. Hmm. like, If you look at 20 years ago, wow,
everyone wants Beanie Babies. Those are going to go up forever. And then it was, like Funko Pops and now it's Labooboos, but these things have a life cycle. And so one of the challenges is kind of guessing when the waves are going to go up and down. Are you seeing that predictive models are getting better as we have more artificial intelligence tools, they have more data sets, or are they still over indexing the past and not kind of like, cause the core problem with
~ economic models as they assume everyone's a rational actor. And like, it's not how anyone is people buy stuff they shouldn't buy all the time. Right. And like, that's why those models always have problems. No. So do we have the same issue with supply chain modeling? Yeah, yeah. And again, this is actually a common issue with or a challenge for retailers is how do you really you do that, you know, assortment forecast on your shelves and what makes sense even in,
you know, in e-commerce is the same thing. So how do you allocate the right sizes in the right market or in the right store is definitely something that is always challenging. And that's why you see a lot of those, mentioned AI and you mentioned a lot of technology that helps with that And that's why now the point of sale information is becoming, our demand sensing is becoming very key in terms of really allocating or really aligning on what product mix should be in what market,
in what store. So it's driven by, some of it is historical, but sometimes it's also the trend that goes into specific market that really the allocation. And that's very important. It's, you know, you see a lot of transfer costs as well happening as a result of missing those, you know, particular forecasts or sales. Different market that you have to shift to a different market which is an added cost and actually one of my clients had to deal with that a
lot because of you know misallocation of their product in specific stores or the forecast stopped at the volume level and Left it really to operation to say where should that product go to which market without really an intelligence in there? So I think the intelligence that comes either from AI or basically Whatever demand sensing technology have definitely helped understand where the trends are and where should you align or replace your products in which market. And you've seen Zara do that. You see
Amazon doing that every day. I mean, they're using a lot of algorithm to forecast before you even buy based on habits, on different things that aligns with your profile. So sort of reduce the lead time. market or to customer. there's a lot of, you know, and now we're talking about, you know, ultra same day delivery as well. That's going to even require a lot of different technologies or advanced technologies into the forecast algorithm so that you place inventory
in the right place. So there's a lot of pollution there for sure. Yeah, as we change this expectation of delivery time, like I remember, and this will age myself when you would order something and they would say to be there in four to six weeks. And that was like, and now if you told someone to be there one to 14 days, Mm. Mm. right, the same two week when people would go crazy, they don't have that patience anymore. Okay.
And we went from expecting three to five days with Amazon and then two to three days and same day. I can't even imagine. similar. We're seeing these shifts in the market where now it used to be you would go to Amazon if you like would wait for the deal. But if you want it now, you go to the mall. But now that Amazon has caught up with the immediacy, because they're so good at load balancing and balancing between their different fulfillment centers. So how can smaller
e-commerce brands or physical stores kind of compete with that when they just have this massive edge in the logistics area? Like their supply chain is so strong, their ability to Even the smallest things like they even like, you know, if they're buying dial soap, they might buy dial soap from 50 different vendors, because they're also factoring in what if one vendor gets sick, Mm-hmm. right? And they're not delivering that they don't want to have. Nope. So they're always thinking about these
different layers to it that smaller can get hit by when they all have one supplier for each different item. I would never have three suppliers for one item. Yeah. They don't even think about that kind of that reseller or wholesaling aspect Ahem. of working with multiple wholesalers. So how can. Hmm. smaller businesses, whether it's e-commerce brand or a physical location, find efficiency and compete with these larger behemoths. Is there a way to use technology to kind of even the playing field at this stage?
I mean, you can always use technology to really bring in data from your point of sales or from your customer or align to their forecast just like anybody does if your customer is that sophisticated. But like we say, if you cannot do it yourself, you buy it. And that's what small businesses are doing. They're outsourcing those kind of capabilities. And you see that a lot in those shared service. shared services that a lot of companies, whether they 3PLs or 4PLs or even consulting firms now,
they do shared services in terms of how they help. give you that visibility and predictability of demand and be able to you data that you can depend on to sell to the right markets. always, there's a lot of small businesses that definitely depend on Amazon right now to sell their products direct to customer. And that's a model that a lot of them use. So whenever you don't have that capability, definitely you reach out to the right you know, ones that can do it for you.
But if you're doing it in-house, and I think that's just of your question of, you know, how you can help, how can, you know, small companies procure the right technology or use the right technology to really align with the market and fast deliver, it's going to be a function of, again, how they prepare for the demand in advance and buy in advance. is one, but then how can they really predict again where your market is and distribute it in the right place,
set up the right distribution centers is one of the key things that you need to look at. And you know, and the supplier side to your point. . You know, again, you start with the strategy. How do you build the right, you know, strategy? Are you going to multi-source? Are you going to single source? Are you going to sole source? A lot of small businesses might find it challenging to procure the right suppliers in the beginning, but and they might use one single source.
But as they build the muscle to really explore and and find different suppliers, can definitely diversify the portfolio of suppliers to get them the right price, the right lead time agreements that they need from the suppliers so that they can develop a partnership that provides them competitive advantage in the market. And a lot of the power Even if you're a small company, the power that you get is from your partners. How good your partners are, how fast are they able to respond,
how much they can flex. Of course, leveraging their technology as well is definitely important in terms of how do they really respond fast to those orders and how they deliver those orders. I think there's something critical about adaptability that Hmm. we kind of overlook sometimes because sometimes a business's first ideas just it knocks it on the park right and then then when something goes bad they can't handle it because they've never had to adapt to it and there's an interesting trend right now
that is a little bit of a supply chain issue and it's that the younger generation drinks a lot less And so a lot of restaurants are now complaining because that's kind of how they drive the majority of their revenue, right? Sure, the food is there, but there's such a markup on alcohol. Sometimes it's a thousand percent like it's so massive, same for nightclubs. And it's like, Mm-hmm. well, they have to figure out something to put in that section that to fill Mm.
in the gap, right? Because the first thought is just raise the food prices. And it's like, well, you can only push that so far. They're all trying that, right? So from a supply chain perspective, Mm-mm. what do you think? is like something they should do like should because what they're kind of thinking is like, maybe they'll start drinking like that's I think Hmm. the sub strategies is like maybe it'll turn back around on its own. But if there's a generational shift, Yeah.
and there's a culture of kind of like restaurants don't have smoking sections Mm-hmm. anymore. mean, planes used to have a smoking section, Hmm. right? So it's like, there's a change. And now, that's why everyone vapes. And now I guess everyone like, puts tobacco in their mouth like the chewing tobacco or something different. So there's a a shift, do you think that there, how is the market can handle a shift like this as an example, as someone who's like, well, I need to
change my suppliers because I need to change what's here because I need to adapt. How much of it is it like, well, I stuck because I have an alcohol contract for the next year. Like what is it like for a smaller business in that type of situation? That's a very interesting and unique situation for sure. And, you know, aside from, you know, the drinking part or not, I mean, it's definitely there's always going to be shift in customer behavior, right, or preference, right. So this is
more of a situation where how you really diversify your portfolio of products and services, for sure. And you know, depending on the flavor. mean, it's no surprise that, you know, beer, for example, has definitely been a market that has been squeezed in terms of margin and sales, with the exception of, you know, craft beer. I know that because I had a couple of clients in the industry. And definitely it's been a push system. So that's entailed a lot of adjustment in terms of pricing and volumes.
Now from a store side, mean, I think your question is more about, you know, from a restaurant or a bar, you know, standpoint, how do they shift that? That would be interesting from a supply chain, you know, situation where you will have to definitely shape demand and, you know, demand shaping, technical term that is. is all about how you really drive promotions, how you drive, you know, the right mix of seasonality of, you know, your beverages or services. And you see that a lot now here.
It's, you know, it's a matter of, you know, what's what's, you know, you don't have to hold the 10 different beer tabs. You are always going to make sure that you select the ones that actually sells the most. It's not about variety maybe, it's about focusing on, or maybe it is about variety, it depends, but it's always about how you drive the right mix of your products and services to grow demand. It's one thing. And the demand variability of restaurants is always...
or bars is always interesting and it's seasonal, So how do you adjust and stock the right products in the right time is one of the key things so that restaurants also don't have to hold a lot of inventory. So they find opportunities to also optimize on their working capital and take that money and invest it in other facets within their business, for sure. I kind of think about how 20, 30 years ago, every bookstore started turning into a restaurant.
They added a cafe and the cafe was taking up more more foot space. Let's go. And I wonder if restaurants are all going to turn back into bookstores or something else. That'll be interesting. You see that model actually. think a couple of them do that now. Yeah, it's like, I've seen this thing now where there's like, barcade, like an arcade with alcohol in it, like they're trying a lot of these mixing of ideas. And I think that's an interesting part of innovation. Mm-hmm.
like, there's an asset, all these old vending machines, Mmm. all these old vending machines, or all these old arcade machines. We can mix that and create a new experience, Yep. ~ nostalgic experience. I know there's something huge going on with bowling right now where there's like a big market shift there. Mm-hmm. because everyone hated MC and then they got bought by someone else and now it hits a new company. I was reading about that today. These are the types of things I read about.
So I'm always interested in these market shifts. Yeah, yeah. Where do you kind of see the futures? I know there's a lot of talk right now about near-shoring and on-shoring, is like starting to bring manufacturing back to America, starting to bring more warehouse to America, simply because having seen a couple of different market shocks in the past few years, which is like, what happens if there's a... tear for between two countries, which used to be so common, right? 200 years ago, was countries would tear
for each other all the time. And you would try to have everything manufactured internally. Hmm. And then we try to shift everything to Asia, first China, now India. And now we're starting to bring it back. I've even seen work with companies that do neoshoaring even for tech support, or even for phone teams, especially now with AI, you can accelerate and actually get a better price for one person running 10 AIs in America versus 10 people in another country. the pricing has kind of changed.
And are you seeing, do you think we'll see a shift in manufacturing and just because it doesn't have to travel as far, kind of, there are start, we're starting to see some advantages to actually building back in America again. Absolutely, I think it's already started. mean, whether it's near shoring or French shoring, we're seeing a lot of that happening right now. And it can happen. It's just a function of time and capabilities for sure. Of course, there's a lot of automation.
So it's a function of definitely skill set. No doubt that automation is also driving manufacturing right now. And the cost. there are specific costs with the manufacturing that might not make it as cost effective as it is offshore. Electricity is one of them, for example, and depends on the industry and how much they use of that is definitely a factor. I think every... manufacturers looking at their or company is looking at their manufacturing footprint right now. Should you know we continue to use CMOs or
we're using the right CMOs or contract manufacturing organizations or should we do it ourselves? Apple you know recently announced their new manufacturing facility for Glass recently and it shows you with the volumes that you're doing and with the top of knowledge that they're doing there definitely can make it happen. And it's not just a function. if we're talking just about the factories, that's not a problem. I think the most important elements out of this is your supply chain. So are your suppliers going to follow?
Are your suppliers able to provide you with those raw materials at the right is it still going to be cost efficient to do that versus, you know, if they are set up somewhere else and they have to incur even higher on moving the raw materials to country, that's a factor that definitely they are going to be looking to consider. With tariffs and everything, it's going to be interesting as well, right? In terms of how much is that going to cost them? to bring those through materials.
There's definitely a lot of, you know, free zone talking right now that would definitely help a lot of those manufacturers bring in those through materials with no tariffs. So that definitely can be a factor. And in general, it's, you know, when you think about manufacturing or when I think about manufacturing, it's just not, you know, inside the four walls type of, you know, setup that is an issue. It's your supply chain in general. You know, your distributors as well. You know, you mind you, you definitely
thinking about the market here, but we still ship, you know, most of our iPhones or other products or Nikes or whatever, international. So it's a function of what it makes sense to learn to which market our manufacturing, that's also going to play, you know, a big role in that and those kinds of decisions. I think I've always been interested in it. was a, I think it was a Tom Clancy book like 30 years ago where there was a car accident caused by Japanese car.
And we suddenly said, you know what, we're going to inspect every car really big time. And it slowed down the entire supply chain. Mm. caused like a massive trade war. And it was like, we'll just inspect the way you inspect. Yeah. And it can be something that small that you don't realize actually changes Mm. everything because we spend like one minute inspecting Japanese cars or they spend four hours inspecting ours. So if we switch to four hours, now it's like, you know, that's a 100 X
increase in revenue for them and it makes a huge difference. It could be all these little things. So you, the one thing that you know is that there's always going to be a Mm-hmm. You just don't know what the surprise is going to be. that's why I think it's important because we're forgetting that supply chains Yeah. are important and we're forgetting that the artificial intelligence, just the real world. And you talked about electricity costs, which is a huge topic for me. So I think
this is really important and things really excited people who want to know more about supply chains. I know you're teaching at a university, which is so exciting. Where can people kind of see more of what you're doing and kind of learn more about the type of companies that you help. And maybe even if they need your help, they can reach out to you. Absolutely. I always, you know, I mean, this is my job, right? I'm definitely working with, you know,
from private equity to corporate on terms of how I can help find those, you know, opportunities. to improve the supply chain first, but as you can imagine, my work has taught me to, because I work with private equity mainly, is to really how to align that with their cash flow cycle time, with the value creation thesis. This is definitely one of things that I enjoy is how to find those opportunities and be able to focus on where the value is and bring that value across
the value chain. ~ thank you, Sochring. I'm gonna make sure to put your LinkedIn profile and your website and all your information in the show notes and below the video for everyone who's watching this. Thank you, Sochring, here at Dendtoday, Mohannad, for an amazing episode of the Artificial Intelligence Podcast. Thank you so much Jonathan for the opportunity. I really appreciate it and thanks so much for reaching out. Appreciate you sir.
More on ai in manufacturing, energy and infrastructure
