Tackling Tariffs Using Contract Intelligence, with Anthony Tacker of Flexgen

Everybody and their grandma knows about some of the big challenges the last year has given US in 2025 on tariffs. So as those tariffs were fluctuating every month, sometimes multiple times in a month, it became very important for us to be able to adapt and actually mitigate a lot of those risks. Because if you miss a window, that tariff has risen and you may or may not have those clauses in there to be able to align them and not present either your customer a big risk or yourself a big risk. I'm Hal Marcus. And I'm Mae Maung Wudiwe and this is Meeting of the Minds, the legal AI podcast from evisort. We interview lawyers, professors, and legal operations pioneers that are pushing the envelope leaders using technology to drive great business outcomes and shape the future of the profession. How's it going, Hal? Going well, Mei Mei. We just had a really upbeat and interesting conversation with Anthony Tacker, who is the director of Risk and Procurement at Flexgen. Yeah, Anthony, he has an engineering background, but he's been working deep in contracting and the contracting space for over 20 years in various positions across different companies and different industries. He's incredibly innovative. Not only is he steeped in AI, but you'll notice he even has a 3D printer behind him in the interview. So he's got a innovation going all throughout. Yeah. I'm really glad you brought Anthony to the podcast. We had a great conversation about some really interesting work he's been doing around tariffs and saving his company a ton of money by doing really in depth analysis on that front. And we talked a lot about the different roles. He's not a lawyer, but he's been doing contracting for so long, he's bringing a wealth of strategy strategies and knowledge about how to contract well. So it's sort of how procurement and legal and contracting all work together to get the best possible results. Awesome. Well, I'm all charged up for this one. Let's hit it. Hey, Anthony, we're so excited to have you join us today. Thanks, Mimei. I'm glad to be here. Thanks for an invitation. So you've been in this space at this intersection of contracting and procurement and legal operations, legal, technology, all of this for a long time now. Give us a little bit of your background and the role you're right now. Sure. I'll start with a current background. I'm the director of Procurement and Risk for Flexion Power Systems and what we do is build battery energy storage systems and custom solutions. And one of our big things that everyone knows us for is our software that controls how power is distributed through electrical grids and stuff. My background I started a while back, you know, in contracts and risk. If I add it all up together, probably a little bit over 20 years I've worked in heavy industrial as well as clean energy. These last I would say four years. And one of the greatest things about it is how much the practice has changed over the years and stuff. And I've had the blessing to work with really good companies and really good resources. But I got to tell you, these last couple of years have been so exciting with some of the things that are happening in the market and with technology and stuff. So for me I've really taken a big part of my last couple of years and just really delving into artificial intelligence and how can it improve you with some of your contracting operations. But I've worked in 3D printing experimental stuff. I've worked in engineering research where I'd help Design and test 1D 3D Cyclone technology to clean the air. Before Dyson was popular, my professor had invented Cyclone technology. I've worked with electrostatic precipitators so they can clean the air with electricity. So I've been in technology industrial stuff. Up 12 years on the industrial advisory board at the University of Houston where we helped and build a construction project management processes and stuff so that when people graduate they know how to work. So I've got a lot of background in technology based or program stuff. But my heart's been in AI for the last couple of years. So that's kind of my background. You mentioned you've got a 3D printer there behind you. So between the AI and the 3D printer, I don't know, I'm picturing Tron now or something. I think big things are going to come out of that room. Yes sir. I like to work on things. I would call myself a super nerd but I really like AI. I do it outside of my work requirements. I love Yalls company just because of the innovation and not just innovation, but just how good a relationship it's been for our company Flexgen. It's made a difference in us and it saved us money. But it's also made us a better operating company in a lot of areas. Not just in the legal operations but in how we do things in the field. How do we digest really risky things that come at us? How do we respond to things sometimes that are like off the cuff that you would need to have information very quickly. And that's one of the benefits of having evisort, at least for Us in our usage, having the ability to get information quickly, having accurate information, having information that is cited. With evisort, the way that it's built and trained, based off of my experience for several years using it, it's going to give you a highly, highly valuable set of data that you can use. And you can use it as excluding something that you're trying to figure out or confirming something. And it could be from claims that you may have to deal with. You know, like, does this really count as a change order? And you don't necessarily cut out lawyers out of this. This is not to say lawyers are not needed, but there's different grades of things, like most people that are expected to be able to read their contract. But let's be honest with you, is somebody going to sit down and read 600 pages in a sitting? They're not. They're going to read it over a period of time. And so if you have aids that you can give them that are pulled through evisort, I call it the cheat sheet. Everybody loves it. It's called the cheat sheet I have for this particular industry. What are the most important things? Do I have liquidated damages? Obviously, evisort pulls out a lot of these clauses automatically for you, which is a blessing. But there are some things that you can customize with. You know, you can train the AI to pull things out automatically. So when you drop that document in there, it's pulling out the data you need. So then you can take that data and give it to the right people. Does an application engineer need to know what the performance data criteria is? They can have access to the contract and read it and pull it through evisort, or if they're looking for something specific, they can go and put the, you know, ask AI and get the information out very quickly. And that allows our teams to be more finesseful and also to respond to things quicker than traditional models. So that's one of the bigger things that we've been focused on as a company and is not only how fast can we do things, how accurate can we do things? And then obviously, everybody and their grandma knows about some of the big challenges the last year has given us in 2025 on tariffs. I pulled some data just to verify what I thought, you know, last year at the starting point in January, tariffs for, say, one of our bigger commodities batteries was if you were buying from overseas, say China for one of. Because we have lots of suppliers. And this is not, you know, anything confidential that I'm sharing. But let's say you were buying batteries for your project, commercial grade, major, major batteries. You were paying about 10.9% in tariffs, which was really, really good. But when you jumped into February, it doubled. So as those tariffs were fluctuating every month, sometimes multiple times in a month, it became very important for us to be able to adapt and actually mitigate a lot of those risks. Because in our type of business, because we're kind of an integrator, we have our prime contracts with our customers that may have certain Clauses, performance guarantees, LDs, you name it, and then you have your suppliers that we buy from that we have to try to get aligned. An alignment of two separate or three or four separate parties is very hard to do it quickly because if you miss a window, that tariff has risen and you may or may not have those clauses in there to be able to align them and not present either your customer a big risk or yourself a big risk. So for us, one of the greater things we used it for was being able to identify if the contract allowed for changes in law. And sometimes when you look in a contract, it's going to be real point of fact, changes in law, this is allowed or this is not allowed, and this is how you handle it. But there's other ways to state changes in law. Sometimes you'll have something that's a force majeure, and you have to look at your force majeure contract clause, and maybe the force majeure says changes in law or changes in a government, or government changes. So you might not have picked it up. They might not have a specific clause that said change in law, but it's covered under another place. Or it could say taxes, and another section could be covered in taxes. And you have to look at the relationship with those clauses and see what's precedent. Because sometimes you'll have things that are contradictory inside your contract. And that's where evisort really paid for itself. Because when you looked at for us in our particular usages, we had maybe over 2 million pages of documents to look through. And that's tons and tons of contracts. And I did this on purpose because I wanted people to know, not fictitiously, but factually, how much does having evasort in the AI platform really help you? If you were to have a person or two or three people, how long would it take them to go through each of those contracts that are hundreds of pages long, read through it all, find those clauses, because they're not the same contract. Every contract's different depending on your customer. Yes, we use templates, but not every customer agrees to those templates, and not every template is the same. So you may have $10,000 a day in liquidated damages. And then another contract says it's a hundred thousand dollars. And I'm just throwing those out as examples, not revealing anything confidential. But the main thing that I'm trying to get at is with the EVISORT platform, workday platform, you're able to gather data quickly and disseminate really quickly. So then you know, okay, 90% of my contracts give me an out. Okay, then what you need to do in your mind is say, okay, now I've identified, I can mitigate that risk, but how do I do it? Okay, then this one says three days. And we'll category these are super high risk. We need to take care of those first. The rest of them are give you a week, and you have to look is it calendar days or business days? And so then when you segregate all your data using the platform, then you can build your risk mitigation plan. And then you say, okay, now I know what the highest risks are. These are the projects we have that we have maintenance on that may have risk. Okay, but these are less volatile. So we've already got those built out. We're just dealing with minor maintenance. So that's not big. So we push those in that side. Then you have your other bucket. These are the ones that you're building or going to build. And you have to be able to respond quickly, because if you don't, your customer is going to be shocked when you ask for $20 million more on their one little project. They're going to say, you know, pound sand. So that's why for us, with all these changes that were occurring pretty frequently, we did the work. And I like to say, and I'm not bragging, I'm just telling you the truth. My friend that worked for me and I, because he worked in my department at the time, and it was a smaller department back then, we did the work that would have taken at least three weeks of work. You might have gotten it down at two and a half weeks. But honestly, we calculated how much could a person read, how fast can a person read? And we're not lawyers, but I've been doing contracts a long time, so I know how to get to them. I may not speak very succinctly, but I can read fast and I can write pretty well. So how fast can you get it done and how much time does that save? And yeah, there's time of the labor cost, which is A small part of the fraction. But look at the opportunity cost. And so when people do their ROIs, when they want to look to invest, they need to look at what does that mean? If I'm running a legal practice and I'm trying to go through this so I can get a claim out, or if I need to close something, that means that you have other hours that you could use for other things that you need to do in your practice, or if you're a non lawyer, but you're doing stuff with claims and risk, this is how fast I can get that information to the chief operating officer so she can make an informed decision for our strategy for our company. And if you're on the field level, this is the information I need to know. Because now I need to give a change order to my customer and let them know that this change is coming that the government or whoever put in place so that I can get the monies so I can be made whole. So you know, it's complex, but with having the AI platform, it helps you get your information in a way that you would have been very, very challenged to do any other way. If that makes any sense. That makes a lot of sense. And I just want to dig in a little bit more into some of the tariff use cases you were talking about. And it's interesting because there's no such thing as a tariff clause. Right. For folks in similar industries kind of dealing with these kinds of frequent changes from a tariff perspective. Any kind of quick best practices you might want to inform them of. Sure. One of the things that I would say if you could adapt or without spending a lot of time, you could put a clause in that says that your current price is based off of the in the current market conditions as well as tariff rates. And then site cite the tariff rate that you're using or the tariff timeline. And then that way you could either leave it this, just that as a statement and then you have your change management process which would say anything that changes that's not due to your fault or one of the other things, then that would give you a chance to put in a claim for change order. You could tie it to consumer Price index, you could tie it into changes in government, changes in law that affect financial conditions. And there's a hundred ways to say that, but for me I just made it very simple or I tried to make it simple. And I worked with really good attorneys, I work with really good staff, but I sometimes come up for me, I'm just like, let's be honest let's be transparent. And then somebody has to argue for me why it would be fair for me to absorb all these costs and stuff. And it's going to be hard for you to say, hey, this is me, we're being fair, we want to have a win win relationship. But oh, by the way, you're going to have to take a big loss. Most people aren't going to do that. So what you can do is put in a clause that says, you know, our bid is based off these following conditions. And if for some chance there's a change in law that affect tariffs or financial calculations of taxes, we will address it in a, you know, win some way or refer back to the change in law provision of your contract. And I'm just talking off the top of my head because I've seen it done so many different ways. But the biggest thing is making sure you have that clause in there. Because if you don't have it in there, what I found, and there was very few contracts that I've seen that are like this, if you don't have it in there, they are assuming that you've made the adjustments when you've turned your bid in. But when things are changing on a weekly basis or even a monthly basis, you've got to put that in there. It doesn't really matter what industry you're in, even if you're buying stuff domestically. Like I can give you an example. There's a component that energy systems use, and yes, we get them from California, but they're manufactured in different countries to formulate that product. So if you don't put that language in there, you don't have it. The other thing that's really important to look at is when you get your thing shipped to you and people want to charge you for the tariffs, ask them for their customs forms and the bill of lighting. You know, you may not do that as an attorney, you may not do that as a legal professional, but as you're looking at, to validate, trust and verify sort of thing, you want to get those things in there too. Because if somebody's going to make a claim to you, say, hey, you could say, hey, I had a contract, it was based on this day and this is when you sign it. This is the effective date. But this change happened after the fact or before the fact. So you should have known it. I want to tell you what customer, so I'll just call it Brand X. Brand X tried to come to me and say, hey, we got a trade levy against us, blah, blah, blah, And I was like, well, we asked you for a quote on this day. That was a month after you knew that this was in and you didn't include it. So even though I'm sympathetic to your problem, I can't go get this justified with my customer. And I know that because I've looked at your contract. And that's where having an AI platform can pull up your data quickly. You can look at your prime, you can look at your supplier agreements, and you can do comparisons really quickly with the AI product. But those are the things that I would put in there. We actually put our supplier agreements and then a lot of the other documents into the workday platform, eviswart. And what it allows us to do is do a real quick comparison. We can look for keywords, we can look for key clauses in there. And that's how you can segregate your data really quickly so you don't have to go and click every single one. You can use the Ask AI and you can select the folder or tell it the folder that you're wanting to do, and it can pull all the data for you. And then you can either analyze it, you can even export it out to Excel at times if you want to. And then when you export it to Excel, you can use that to make graphs and charts if you need to, depending on what your preferences are. Because the platform also has the really good graphics that it shows you. And I'll give you an example of one thing that was really interesting that I never knew that it was at the very beginning, there was a contract in Spanish, Castilian. And I was like, I didn't even know that one existed because I never did the contract. And it actually read the language and classified it correctly. So when you looked at the pie chart, it told you like 99% of everything was in English, but this one was in a different language. And so EVISORT correctly identified that for us. So, Anthony, a lot of these points that you're raising get to the heart of a really interesting issue, which is sort of who owns the contract? You really need legal procurement. The work that you've done over the years has informed a lot of these strategies you're talking about. Like, what have you seen as the best way for those roles to work together to get the best results for contracts? That's a great question, Hal. And I would call it the million dollar question for me. What I found, and this is just, I can talk about this particular company. I believe that our lawyers are very, very, very sharp. They maintain a Template repository themselves, and then they issue that to us based off of, you know, depending on what the business case is. And so from there, what we found is using the platform, and this is why we use it. We use the platform to pull data, mostly for us, after it's signed, but when we work with the attorneys, way we've set it up was they have access to all the data, all the contracts in the platform, so they can pull it and use it for their needs. But what we do is we have like a double, I'd say a double corrective thing. So the way we set ours up, when we drop it into SharePoint, it automatically goes up and does an upload into Eversort. And so how we manage that as a team is on the front end. The lawyers give us the particular agreement. And I know this stuff, but to keep things simple for our company and not to alienate people, because I've worked internationally and I know a lot and I could select the own forms if I wanted to, but I don't want to fight with them because I want them to be on my team to help me get stuff done. But they like to give you the forms. And so we use that with our commercial group, our operational group, as we're doing the formation part, and then after the formation part's done and it's signed, that's when it goes into the platform for us. And then from the platform, then we start using it to digest information out for our operational teams. So we start taking out the things that we need. Now, what I started doing, but I haven't fully finished, is building some of the AI models inside of the platform so it'll pull it out automatically. Because right now, one of the guys that works for me, he knows what it is that he's looking for, and so he uses the Ask AI and pulls the data out. He manually types it all in. But I was going to try to cheat and put in the clauses in there and train it, and just when you drop it in, it'll just pull it out and it'll save us time. But I will tell you this much, and I know I'm not answering your question directly here, but if you were to look at what way that we used to do it, it used to take us seven days, sometimes five days to get the contract digested and all the information pulled out manually before we had Eversort workday platform, and now we're down to a day and a half, sometimes a day. So that's how much time difference it takes us to digest the information. So working with the legal team. The legal teams now have access to evisort. I give them whatever they want. I set them up for training. I would say this year, some of the hardest. You're going to laugh at me, but this is the truth. Some of the people that were the most resistant to change have finally drank the Kool Aid. They're like, what is this? What is this, Anthony? And they're finally, because they used to hate me, because I'm like, dude, I'm not taking your job. I'm trying to make. It's a tool. It's a tool to make it easier. AI multiplies your abilities. It multiplies your abilities. I like books. I like a hard book. But if you tell me to go look up a contract, I'm going to use the AI because there's no reason for a person not to use the AI. The only way I would say you wouldn't want to use it is if you were using the wrong platform and it gave you really bad results. That's the only reason why you shouldn't use it. I mean, if somebody gives you an option, you should use it. What I'm saying is the attorneys are adopting it. So to answer your question really quickly, the AI platform is open to everyone. We interact with the lawyers through the contract basis. While we're doing the negotiations. We do a lot of the commercial portions, the legal T's and C's. I can give them feedback because I know indemnity and insurance really well. But most of the people don't, other than the attorneys. But I respect the attorneys that I work with. So I might say, hey, you might consider this because there's been a couple of times where they're not licensed to practice in a certain area and they may not and don't necessarily need to put this in here because I don't want to disparage them. But they don't know. Like when you work in Texas and it's construction related, you got to follow the law of Texas. Say, I want the New York law. You can't go and do forum shopping. There's certain, like statutes and then there's like four states in the union that require like statutory lien forms too. So with AI and workday, we can find those anomalies a lot quicker. So if you know the four states, you could say, give me the four states that have these clauses or whatever and it can pull it out for you. So we work interactively. The attorneys at my current company don't Use it as well as they should, but they have started embracing it. My teams, my procurement teams, people that work on risk, all the operational people use it and they love it. And so when I tell them that I'm doing stuff with workday or ever sort of, and what we're working on, I have no problem getting the renewals because we've already paid for ourselves multiple times. When you save a company millions of dollars and you look at, okay, what did it cost me to invest in this, it would be dumb not to invest in it. And our lawyers, although they were very resistant at first, they are coming around with it. So the quick answer is some attorneys and other places that I know that work on AI for use AI, they love it. If you want to go fast and you want to be a competitive firm or a competitive business, you need to utilize the AI because they're going to get the solution quicker, they're going to close the deal quicker. And if time is money, which everybody believes it is, then you need to use the AI. It's a no brainer for us. That makes sense. And that kind of touches on some of the questions around. You've got dedicated legal AI solutions like workday contract management, like, you know, other tools in the space. Do they also have these kind of frontier models that are more just kind of generalized generative AI tools? How do you see the kind of breakdown on which tools to use when for folks on the procurement or the legal, anyone kind of touching kind of contracts and those other kinds of use cases? Well, that's a great question, Mama, and thank you for asking my opinion. In my observations, my experience, you have to look at the platform and what is it trained on, what is it most efficient, what is it most effective at? And when you go back and look at that, you have to look at, am I asking for a single question and does it matter if there's a lot of variability in my answer? Now with contracts, yes, there's things that outside of contracts that may influence. There might be statutes or laws that can impact something. But when you're going to look for something or you're going to try to analyze something, you need to be really careful. So in my experience, chat GTP is good if you want to do like pull up some, you know, make a chart, pull up some something based off of some facts. But what I've found is when you have tried it, I uploaded a contract and say, okay, analyze this. It's going to give you some data and it's going to look pretty good. But the problem's going to be you might have two or three things in there that's totally made up. It's hallucination. And that's the concern I have. Not because the platform can't give you some really nice looking output. You're going to have to really, really check hard to make sure what it's giving you is actual, factually based. It's not because they purposely make it that way. It's the way that the data is processed in those types of platforms. And what I mean by that is workday in evisort. It's trained on, it's built on its underlying structure. Even though I haven't looked under the hood, I can tell by all the different iterations I've done. It's built for contracting and legal operations. Those frontier type software like Claude Cowork and all these other ones. They're good at certain things, but it's more of a generality. It's like a general output. I go with the thing that I know works, that I know is repeatable results. It's going to cite the contract, it's going to tell me what I need. You've got such experience here with supplier agreements as we get better and better and we're getting pretty damn good at turning all the artful language of these lengthy contracts into specific data points, specific terms, requirements, obligations and making them more and more standardized. That does create a path where you could have more of the terms. Not all of it by any means, not the more nuanced things, not the real contentious things. But a lot of these points could get aligned automatically by the tech between the two parties without a human going back and forth on the artful language of it. Honestly, I see that happening and I'm glad you brought it up. I don't know how society will take it, but I think to me if it gets you the deal done quicker and it still gets you your achieve your goal, I think it's a good thing. I wish I could do the experiments with this because I will tell you that some of the firms that we've worked with are in my past, they're one sided like boilerplate terms. So does that mean now society is going to take it start off with a better boilerplate because they know the AI is going to flag all this stuff and the procurement person that great at purchasing things but doesn't know that kind of information, is that going to sit in front of him or her and help them get better? Are you going to have the big players that are used to, you know, having great lawyers and great teams that have engineered. Because there are companies that I know for a fact, and I've dealt with them so long that they engineer like psychological stress points and hoping that you'll adapt to bad terms. So I'm wondering if how that's all going to play out because maybe they take their competitive advantage when everybody can get a good deal. We'd be remiss before we close today's interview if we didn't talk a little bit about a book that you're writing around how contracting has changed over the years and AI is accelerating. That anything you could say a little bit about kind of that book, what inspired it and, you know, maybe some of interesting issues you're getting into. Sure. Thank you for asking about that. I've been working on it since probably August timeframe. So, yeah, what inspired me was as I got more and more and I've been working with contracts and legal persons for, like I said, close to 20 years, if not longer. And as I learned more and more about AI, I saw some things that I thought, well, I know people talk about AI in general now, and everybody's on the AI bandwagon, but a lot of things, as you well know, they don't always understand, you know, the different types of AI and how it really relates to things. So after, you know, working with workday and evisort for a while and also experiencing personally seeing how sometimes even great professionals having resistance to change, I thought, well, let's think about it. In a business model, how could I help people? You know, obviously everybody wants to make money when they do things, but I have a really good reason for doing it because I want to make things better for people. I want people to get the solutions they want better, but I also want to help the legal practice. And so that's what was my impetus, wanting to look at how's the business model. A lot of lawyers I've interviewed, they know about law, but they don't know where their practice came from. They don't know where the first lawyers came from. They don't even know how it came to be. And that's important because once you find out where you came to be from, you see that it was so noble, it was such a good thing. And it still can be, and it still is in a lot of ways. But also looking at the business model, when you provide services to people, you want to look at things that you can be efficient and effective. But that's the opposite with most people that bill for an Hour, an hourly TNM model. So how do you look at that? So this book challenges the traditional time and material model not for all practice things, but for a lot of practical things, and how AI leverages that ability and how you can build business models inside your practice or your legal operations using platforms such as Eversort to help you with your contracts and stuff. But it's not just contracts. There's other platforms out there that help you with discovery and all these other things. That's really magnificent. And how do you build a case, not a case for law, but how do you build a business case to adapt that? And how do you adapt it and how do you implement it in your organization? So we talk about how that changes and what using AI can do and what in examples of how you could use it. One of them is about tariffs. Other things can be, you know, other items that you need. And it doesn't always have to be legal operations. But most of this stuff is all about how you could take your practice use AI to become, you know, up to 50% more efficient. And so my argument to attorneys are instead of doing the hourly thing, do it lump sum on a lot of these things. Lump sum it, because you're going to be so much more effective. You're going to be faster. You build your models beforehand. And I talk about how you build models and how you test it to make sure you don't lose money. And you could do 10 of those things in the day and make more money. And guess what? You're giving your clients, you're giving them risk mitigation. Because now they're saying, hey, I know what this is going to cost me. You're able to do more work, you'll make more money, and you'll have an easier, better life, you know, in the way that you're doing stuff using the AI. And that's kind of what I'm boiling it down into a short thing, but that's what it's about. It's about taking a chance and looking at not just for what you need, but what your client needs and how do you adapt to the technologies that's out there. Anthony, it's been an absolute pleasure speaking with you today and picking your brain. Thanks so much for your time. Yes, sir. Thank you. Thank you all for your time. Hal and me. Pleasure, Anthony. Take care. Meeting of the Minds the legal AI podcast is brought to you by Eversort. To learn more about Eversort and how we can help you contract embedded with AI, visit evisort evisot.com you can find meaning of the minds on Apple Podcasts, Spotify, or wherever else you listen to podcasts. Don't forget to click subscribe so you won't miss future episodes. On behalf of everyone here at Eversort, thanks for tuning.