Models in Hours, Not Weeks

Prasad Mahajan 01:18

Yeah, and then first of all, thank you for having me. My pleasure. Yeah, so I have 22 years with Transplace and Uber Freight combined. I started off actually writing optimization apps for several of Transplace's customers, shippers. like JCPenney, AutoZone, different customers wanting specific niche optimization salts, but also having users interact with those optimization apps to make daily decisions. So a lot of my earlier career was in writing those apps, also integrating them into our homegrown TMS systems. And then as technology started to evolve, always see that like early 2000s was a lot about optimization, a lot about solver. You could see solver growth. It's a lot about adding constraints, real world, know, multi-stop routing, gaining efficiency through optimization. And as technology evolved, you know, people started to look at supply chain very strategically. It used to be always, hey, we need to... chokes transportation because that's where a lot of the cost is and we need to make it efficient. But as globalization really spread out, information became more available, data started flowing into these through the systems into people's hands. People started making cool stuff out of data. And so this is where the solvers, I think, in democratic, precisely into so many people's hands about, can we really take a strategic look at how we are spending money in transportation. Are we doing justice to our customers? Are they getting the right amount of service? And so a lot of that 2010 to 2015, 2016 was more about network optimizations. know, my supply chain is global. Where should I have my distribution centers? How big they should be? Which customers should I service from? Where? How much inventory should I hold? All those. And then what the later half of, think, the second decade taught us of this century was really the uncertainty, right? Like, okay,

Prasad Mahajan 03:42

well, supply chain can be efficient through a lot of different steps. We optimize it, we plan around it, we design the right things, but uncertainty throws a wrench into everything. So how do you handle uncertainty? How do you handle volatility? And then one-off situations and things like that. So later on, as we started getting into these natural calamities you know, wars or disruptions through volcanoes and things like that, or whether it became important to have a more robust plan and thought process around how do we optimize it under uncertainty and how should the executives be prepared to make decisions in such an environment. And so that's kind of the evolution that I have experienced, you know, starting with that transportation and so towards the tail end of my career the last 10 years were mostly spent into consulting. For the very reasons I just talked about is executives would come and ask questions about hey how should I operate and things like that and the consulting practice that I helped start with my other peers there really addressed the need at that moment.

Prasad Mahajan 05:00

This was both during and after and so I think before, during and after because we really started in 2010 and that's where this software was starting to become easier to use

Prasad Mahajan 05:15

and the role of supply chain modeler was coming along very well and that's the before part but After you Uber Freight, think as I transitioned into Optilogic, I'm seeing a tremendous platform here.

Prasad Mahajan 05:32

Makes it so much easier for strategic decision making and you can actually deploy that decision into hands of many users through apps and digital twins and things like that. So really excited.

Prasad Mahajan 05:58

Absolutely. Although the role has definitely evolved, it's everything that I like to do and used to do in a different world. So there's really three main things I do here. One is really solutioning for our customers. Most of our customers have global supply chains or they're logistics companies and really engage with them and help them really show, hey, how can they make a much more positive impact to their ways of operating through supply chain design, through the technology we have available? Secondly, for some of our customers who are mature, really helping them with whether it's building a model, whether it's building a strategy, sometimes it can be a how do we build a center of excellence? Because we want to have this as a core capability now as supply takes central focus in organizations. And again, the third thing is really customer engagement. Once a customer has onboarded on our platform, how do we make our customers successful?

Prasad Mahajan 07:04

Because as... We want to have as many of our customers with most efficient supply chain, most efficient in terms of carbon emissions, everything like that. So that's the role I enjoy as well as to help our customers succeed and adopt our platform.

Prasad Mahajan 08:46

Yeah, absolutely. And so I think, you know, and that's what supply chain design is all about is trying to find that, you know, best balance between different costs that

Prasad Mahajan 08:57

you are faced with today. And

Prasad Mahajan 08:59

because all these different costs, whether you want to take transportation, warehousing, which has a labor and a storage component. Then you have the inventory costs on top of that. At times they move in different directions. There's different macroeconomic parameters influencing these costs. And so at any given time, the supply chain design is going to give you the best of all the combinations because our technology would look at optimizing across all of these different costs as a cost to serve kind of a parameter.

Prasad Mahajan 09:40

But the opposite direction to that becomes service. And that's where your California example, Southern Cal comes into play is if this was all about service, I would put my distribution center in Southern Cal because that's where a lot of your demand is or somewhere in California due to hit the West Coast. But when people actually look at supply chain, it's not just the demand, but you also have to look at where is your supply coming from. And what is so for one of the customers I worked with in a manufacturing industry where they had a plant in Salt Lake City and all most of the market was in California, Southern California. So obviously there was a need for having a warehouse there, but how do those cost metrics work?

Prasad Mahajan 10:37

Yeah.

Prasad Mahajan 10:42

Exactly.

Prasad Mahajan 10:43

I would say, you know, start with understanding. I would say start with clarity. you need to have clarity on what is most important for your organization. People always say I want the best of both service and cost, but it's really saying that, well, I'm going to meet this service expectation. So if your service says I want to hit all my customers within six hours, eight hours, 12 hours, sub same day or one day, that becomes a different strategy. But if you want to say that, okay, I can do two or three days, but I want to really minimize cost or keep my cost KPIs in control. Then you have a little bit more flexibility around, where else can I be? So for this specific example, we looked at other options. really creatively looking at Henderson, Nevada, which is right out of Las Vegas. And you've got Glendale, Arizona, Phoenix area, even Reno. And so what else to look for is really what can differentiate between one customer versus other is how is their inbound supply chain laid out? Where is the product coming from? If you're making it in Salt Lake City, choices for modes of transportation going into California. Now, also remember that sometimes getting into California is cheaper because it's a backhaul lane. And so if you have

Prasad Mahajan 12:06

a site in Las Vegas, for example, outside, that becomes a backhaul lane. you're all the carriers and you're trucking it are wanting to go back to California. So

Prasad Mahajan 12:16

you're going to get a better

Prasad Mahajan 12:17

rate.

Prasad Mahajan 12:18

Yes,

Prasad Mahajan 12:19

exactly. So you're going to get a better rate going back to California. So even though you're incurring a little bit more miles, you might actually get better line haul. And you just have to see, does now do I now get better labor availability? Are the taxes better here? You know, compared to what I'm paying because inventory in some certain counties can be taxed and, you know, there can be a tax implication that could kill your profitability and could change your decision. as well. So all these parameters, I think you need to think about what's important and then you feed it into the optimizer and it gives you a good side-by-side comparison of all the options you have chosen.

Prasad Mahajan 13:23

Absolutely. And I would say like if it's sometimes the decisions are lot easier when you're actually opening up a new facility. versus you are relocating something because now you have to think about labor and are people willing to relocate and you know all those things so because skilled labor is an asset you know and I think it's on paper it seems very easy oh move from here to here and get the savings but when you actually start implementing this there's a lot of things you have to be very sensitive about because all that training and development that's gone into your labor in a given facility, you know, what does it cost to replicate that in a

Prasad Mahajan 14:03

different

Prasad Mahajan 14:03

facility.

Prasad Mahajan 14:37

Yeah, and so in this case, there was a sizable amount of product that was coming from the ports as well. So the Long Beach, Rialto area was where DC was selected.

Prasad Mahajan 14:51

And so although you had a plant in Salt Lake City that now cost a little bit more to move that product all the way into Rialto, but for anything that was moving in the rivers you know all these containers are coming in and you know getting stored it was lot easier to have those ocean containers bring into Long Beach and then have that product available right there in Southern California.

Prasad Mahajan 15:42

Yeah, and then there is one more piece before we move on to the next one, I would like to add is there is also a little subjective element which you don't see that as well documented online, but you just get it when you are interacting

Prasad Mahajan 15:56

with senior executives and leaders is what are really they going through. And that one piece I want to highlight is ease of doing business, which is very, it can differ by state. state and that's where I think you can give them a recommendation from a number standpoint but they may still decide to override it because that state or that specific geography is just not easy to do business with and there could be labor laws or certain things associated with a lot of parameters but that's a subjective piece that people may bring in at some point.

Prasad Mahajan 16:54

Exactly. And on the other side is like there could be probably nothing wrong, but somebody else gives you better incentives.

Prasad Mahajan 17:12

Yes.

Prasad Mahajan 18:13

Yeah, so I think that's a very good question. think we, you know. em And I'm glad there's, we work with some really good executives who before making a decision, they are wanting to see what the model or supply chain design project is actually telling them and whether that decision is right or wrong. And a lot of, maybe the most, one of the most common one I see in the last few years is nearshoring. outsourcing, right? So now, as you can see, one of the, I'll be very specific, but theme is fairly general. But that example I want to call out is a carpet company that had manufacturing as well as some suppliers in Southeast Asia. And so they were bringing in all these containers into North America, which was both a mix of some raw materials, but also finished products. So a few years back when there was volatility in the ocean rates, you would see the container rates were spiking up.

Prasad Mahajan 19:21

And when you are so much dependent on that one move, obviously the question comes, hey, what can I do different? And some of the executives were about to make a decision that, hey, let's just do nearshoring, move all these manufacturing units or lines, per se, to Mexico and there was not much more capacity available in the US itself. But they did this study before making that decision, which really prevented them from making a wrong decision was that when you look at Mexico, it was a good option, but the labor from a transportation cost standpoint, as well as from a lead time standpoint, it was really hard to beat the labor. So the difference came very close and that's where the other pieces come in is long-term relationships, long-term contracts with

Prasad Mahajan 20:13

suppliers and your partners. How should you weigh that versus some short-term gains? And so the scenarios that we were able to bring in front of these executives were, hey, what if you kept your product in Southeast Asia? What's the worst case if ocean rates spike and they stay long for a period of to 12 months. Versus what if you bring everything into Mexico or what if you start another line hypothetically in the US. And by seeing those different cost parameters side by side, although there was a little dip in the Mexico option, they still selected to choose to keep the sourcing in Southeast Asia because of the relationship and the short term nature of this disruption. and so forth. So they prevented making a wrong decision I think at the end and I'm glad you know they decided to use this approach to really make that decision as an intelligent decision through data and modeling.

Prasad Mahajan 21:48

Yeah. And now our answer to all of this is because we've seen the number of disruptions that have a negative impact are growing in number. They're becoming almost a way of life. I just attended a conference recently where theme of key supply chain officers was supply chain management has become more of a firefighting exercise,

Prasad Mahajan 22:12

right?

Prasad Mahajan 22:13

We don't want anyone to constantly firefight. everybody to be on top of that. And so our answer to that is having an always-on digital twin

Prasad Mahajan 22:25

in the market where you have your supply chain model ready and it's all refreshed so periodically, so frequently that it's always on, meaning I have the most current view

Prasad Mahajan 22:35

of my network.

Prasad Mahajan 22:37

And then I'm either anticipating and modeling what can go wrong. And I have a really a playbook of, this goes wrong. I'm going to take these three decisions first. If this goes wrong, I'm going to take these decisions first. And it's not too hard to understand what can go wrong because you have a good history, but you also know your current metrics, right? It's like, what are my current metrics that are exposed to risks like these? And so being prepared to... disruptions are not going to stop. We are not going to in the business of stopping disruptions, but we want to make sure that you're prepared to handle that. You can anticipate those and you have a playbook of actions ready. If that happens, you're not starting from scratch. You're really getting into, okay, where should I act?

Prasad Mahajan 24:36

Yeah, great question. And I think it's more of a matter of how your exposure is in supply chain versus whether you're big or small. And that's what I think because even if you're a small company, if you have a critical mass of suppliers, if something happens to them, it can really take the company down. This becomes a critical feature to have for you. Secondly, It doesn't cost as much to build a digital twin nowadays with technology and especially our platform, is so AI enabled, that

Prasad Mahajan 25:13

once you have the data available with a team of humans plus AI as an enabler, you can easily set up your data pipelines through our DataStar technology. and having our agent API really makes it simpler to cleanse the data, make it. more structured for modeling and so forth. So having those data pipelines, connections is now becoming very standard in a digital twin so that you with a click of a button can get a refreshed data or set some automatic jobs to do that. And so your platform is always current. So anybody who needs to assess their supply chain and say, hey, what are our risks? Are we exposed to disruptions? and I would be surprised if anybody says no. And why should I not have a digital twin, right? And if you don't get an answer, that's really the path forward. And then the last thing is the reporting has become so crystal clear with our platform as well as the use of our agent AI is that you're not just telling senior executives that or your decision makers or your operations people, business users that, hey, this is what the delta is. But it's actually going deeper into saying, this is what the delta is, why this is happening, and this is what you should do. So the getting from insights. into prescription is I think the real value of

Prasad Mahajan 26:47

a digital twin.

Prasad Mahajan 26:49

Everything else is structure and you know mechanics which I think we have mastered how to do that pretty well. So if you tell me there's a company that doesn't value prescriptions I don't think digital twin is for them but I don't think you'll find such a company.

Prasad Mahajan 27:36

Absolutely, and I always use the word that this is not a small improvement. This is a game changer. So I used to hear us say a lot that, we can model in days instead of weeks. And then when I actually tried it myself, so I come with actually a very minimal modeling background. But I will, with the help of... either our agentic platform and AI, the whole burden of modeling can be done in, I would say, if you are fairly data ready. But from setting up the pipelines to the actual model to getting a reasonable number of scenarios out. It could be hours or days. I have built personally eight to 10 models in a month that got done just within hours and they were across different use cases. So some were network optimization, some were pure transportation routing, know, fleet optimization, sizing the fleet. Everything is becoming like an hour's play instead of like weeks and weeks of work.

Prasad Mahajan 29:48

Yeah, and I... Thanks for asking that question. actually that's coming from the 3PL and really 4PL now background. That was a very important aspect for me coming here. And we actually now have a pretty good playbook for 4PLs or 3PLs who would like to do these services for their customers or even themselves. So one of the things that's becoming very critical, I just saw a survey recently, is that going forward, the future of the revenue looks like for the 3PLs and 4PLs is there's going to be a lot more growth, faster growth in the services and solutions side than the trucking side itself or any other modes of transportation. services aspect here is how do you be that strategic partner to your shipper. And that's a pretty decent, I would say, thought process. But also, I would say it's more essential now that if you're a 3PL and if you have the customers, your shipper's data, there is an implicit expectation that give me insights. Be an advocate, be an advisor, be more than just a partner. or take that relationship to the next level than just being transactional with me. And this is where our 4PL playbook would come in play is these are very specific, pre-designed use cases that 4PLs would run both from their solutioning side as well as repeat customer improvement value at standpoint is they would have an array of these use cases already ready to go.

Prasad Mahajan 31:35

So whether it's

Prasad Mahajan 31:36

like

Prasad Mahajan 31:36

a. Yeah, a lot of that is tactical. Like, example, obviously the strategic is fairly obvious, but tactical is like, I want to change my network this year, but I want to see if there's anything I can do. Should I use a cross-dock? Should I use an extra space in my DC as a pool point? Should I, anything that goes into Mexico, know, should I send it through Laredo or should I have a pooling point somewhere in Chicago, right? For everything that goes south, same with everything that comes north, what should be my fleet size and so forth. So all these use cases, we're offering that as a playbook for three PLs and 4PLs.

Prasad Mahajan 32:22

yeah. I would say that's really recent and maybe always had the capability to run it through our platform. But then, you know, what our recent advancements in deploying these apps is much of a differentiator. So when you package all of these use cases into a ready to use app, then a lot more broader people can use it. They have a good... screen to understand they have clicks of buttons and then they have different tabs where they can go and see the output and take actions.

Prasad Mahajan 33:46

Yeah. And I, it's, I see more of this as a, we, we are, our customers always ask for some things and I think, I'll address that, but from a more broader space side, I see us more as a, are leading the customer into, Hey, how should you do things best? and what is out there. So I do think that we are definitely, and I've experienced this coming in fairly new. I started in February here is, you know, the pace of innovation is just so high. that

Prasad Mahajan 34:21

we had this agentic AI initially a few months back. Then we have this Claude connector that went in. We were the first supply chain design company to get there. Now we have the digital twin, full approach, step-by-step walkthrough. And then we are also deploying cool things around apps. And so I think my focus would be also, hey, let's just bring our customers along, focus on adoption, show them the value that it brings. And then finding the good balance between human and AI to make sure that we are bringing the best to the industry as a whole and

Prasad Mahajan 34:57

to our customers, both in terms of efficiency, productivity, and then the accuracy of the decisions that they are making.

Prasad Mahajan 35:35

Absolutely and then You know, we also have a ton of progress in simulation, which is really

Prasad Mahajan 35:42

running hundreds

Prasad Mahajan 35:43

and thousands of scenarios very quickly. And

Prasad Mahajan 35:46

the way I see the product evolve there is, you know, that you have classic sensitivity analysis, which used to be fairly limited to certain parameters, certain what ifs and all those things. But can I just run a thousand of these very quickly as processing times become faster and then give you, instead of you choosing, oh, I'm sensitive to this or that, but the twin or the tool actually tells you that, okay, you're going to run out of capacity in June, 2028 at this speed. with this set of demand and this criteria, if nothing changes, because those are the questions people always want to know is at

Prasad Mahajan 36:30

what point, you know, business continuity is a great use case here because you want to have business continuity as you are growing. And even if you are a mature company, you want to still maintain that in a profitable way at no point. And we saw that during COVID that people really had to sign leases on warehouses at a really expensive rate because they ran out of capacity.

Prasad Mahajan 36:58

Yeah, so that proactive feedback is what I would see people focus more on is, you know, not as much as how do I model and, you know, which model do I create, but for my supply chain, what is that proactive feed that will alert me, you know, what are my sensitive parameters in the next three to five years?

Prasad Mahajan 38:25

Yeah, and I think the...

Prasad Mahajan 38:44

I think I see a good mix between the two. We have a pretty sizable that... are definitely coming to us from others just citing the pace of our innovation and so much investment we are doing in our products and services. So, and then for that entry level, I would say, I'll quote our CEO here is he always says, be curious. So there is with that at this pace. and where we are technologically from a feature standpoint as well as from an AI standpoint. Being curious is really important to really start that, whether you want to start a center of excellence or whether you want to start small with a quick improvement. There's nothing wrong in either of these options. But be curious right now. You just have to think what you want to do. And there are our platform would allow you to enable that. And especially with AI, you don't have to worry as much about, how do I make this happen? know, it's the technology is going to guide you through that. So what's really important as an entry level is to really be thoughtful about where that value is, how much how quickly I want to deliver that for my organization and not get lost into too much of the weeds. You know, like if your data if you're not data ready. Don't worry, start somewhere, start small, get that quick win, get that ROI and that will drive more investment from your stakeholders, your leadership, your ELT team. Because when they see that as a success and they see that as an ROI, that goes a long way and then you'll get more investment to scale big and so forth.

Prasad Mahajan 41:15

Absolutely.

Prasad Mahajan 41:16

Thank

Prasad Mahajan 41:17

you for

Prasad Mahajan 41:17

having me when I thoroughly enjoyed the conversation with you as well.

Prasad Mahajan 41:24

Yeah, talk to you. Thank you.