Using Technology for Better Negotiations with Martin Rand

Hello, my friends, and welcome to another exciting episode of Negotiate Anything. This show is produced by the American Negotiation Institute and with over 5 million downloads and listeners in over 180 countries, listeners just like you have made this the number one negotiation podcast in the world. Hi, my name is Kwame Christian and I am the founder and CEO of the American Negotiation Institute here at A&I. We believe that the best things in life are on the other side of difficult conversations. And we are passionate, passionate about providing you with the best content that will help you to make your difficult conversations easier, both at work and at home. Lastly, I want to remind you that we offer consulting and conduct trainings both virtually and in person all around the world. Our focus is in three main areas. First, negotiation and conflict resolution. Second, leadership. And lastly, diversity, equity and inclusion. Check the link in the description below to learn more about how we could work with you and your team. And now, without further ado, let's jump into the interview. Martin, thank you for joining us today.

Thanks for having me on the podcast. Kwame, pleasure to be here.

Yeah, it's our pleasure to have you. I'm excited for this one. So how about you start us off by telling us a little bit about yourself and what you do.

Oh, absolutely. So, a bit about personal. Well, firstly, I'm the co founder and CEO of Pactum and we're automating long tail vendor negotiations with AI and maybe a bit about personal background. So I used to be a product manager at Skype. I then set up my own startup which was actually in farm management. We created a farm management system for farmers and ended up selling that to the Climate Corporation, which was the digital subsidiary of Monsanto. And as a result, I became Climate Corporation's commercial lead for all of Europe. And my task was to, among other things, was to negotiate with big food and ag companies in Europe. And negotiating in Europe is a lot of fun because you can really see the diversity of different cultures. So I could begin the week with negotiating with the Ukrainians and end the week with negotiating with the French. And I thought, wow, does this really have to be so painful as it is, I think I got to experience all the cultural differences and all the cognitive biases there are to experience in Europe. And I thought, wow, this technology could really make some of these things smoother.

Wow, that's great. This is really interesting. And so the general topic that we're going to talk about is negotiating with a computer or using AI during our negotiations, artificial intelligence. So for a lot of people, we don't even know what that means as it relates to our standard approach to negotiation. So when you think about using computers and artificial intelligence in this way as it relates to negotiation, what does that even look like? Yeah.

So I'll bring you a simple example. So we work with very large Enterprises, Fortune Global 2000 Enterprises. One of our customers is Walmart, for instance. And such enterprises all have similar problems. They have tens of thousands of vendors, and according to the Pareto principle, 80% of vendors are long tail, which means that people can't take sufficient care of them since those are small deals and there are simply not enough people, even though some of these companies have hundreds of negotiators. So these deals are sort of neglected because they're small deals, and we renegotiate those deals automatically. So what we do is we do interviews with our customer and build up a bot that can negotiate. And for the vendor, the bot is a chat interface. So essentially the vendor receives an email saying, hey, we haven't talked in a while. Sometimes we haven't talked for years with the small vendor. And hey, let's look at our contract and let's find out if there's a more beneficial contract out there for both of us. And the bot in a chat interface will talk with the person just like a human would and come to this new agreement.

That is so cool. That's really cool. This is really interesting because like I mentioned at the beginning of the conversation offline, as a whole lawyer, the negotiations that we have are very traditional. And even when technology is introduced, it takes a while because of the rules of professional conduct. We need to make sure that everything's above board. And that can take a long time too. And so it's really interesting to see how other areas are able to kind of innovate in this way. And so the person who's interfacing with the bot, do they know that they're talking to a bot or is it meant to kind of make them feel as though they're talking to a person?

Yes, they, they do know they're talking to a bot. And, and very interestingly so, Depending on the project, 7, 74 to 82% of the people who have talked to our bot would prefer talking to the bot next time as well, not a human. So that's, that's sort of a counterintuitive finding because, you know, if you call a call center and end up with a bot, you know, people don't like it very much.

Right.

But interestingly, it's different in our case. So we were afraid of the same effect. It didn't come, it didn't come to force because we realized that negotiations is a high stress situation. And people, people actually tried to avoid this high stress situations. And now imagine that if you're a small vendor doing a few hundred thousand dollars per year for a huge multinational corporation, and so they have people on payroll that have been doing this for all of their lives, and they don't have a lot of time for you because simply you're such a small vendor. So that they would try to rush through the negotiation, probably using some cookie cutter ideas that they have. And so they wouldn't spend the time to really expand the pie and to really find a solution that would be beneficial for both sides. But it's different with a bot, since with the bot, the bot has unlimited time. The bot will talk as long as it takes, as long as we can agree a good agreement. And so the vendor now suddenly feels in control because they can. Nobody's rushing them. And they can really see all of the available options that the bot will offer to them, all of the available benefits for them to make the deal better for them, so they could reciprocate with what the bot asks so it's less stressful. The bot is always polite, and that's feedback we're receiving a lot. The bot listened to me and it was a pleasant conversation. So human aspects like this come into play and they can also get a better deal.

This is so incredible. This is really great. I'm glad we're having the opportunity to talk about this. And so going forward, the typical format, we have three main points that we want to discuss. But I'll be honest, there's a big part of me that just says I want to just talk about what this, what this means and what it looks like. But I know we're going to dig deeper as we go organically through the conversation. So the three main topics we discussed were, number one, the pros and cons of negotiating with a computer. Number two, answering the question whether or not the computer can expand the PI more than a human can, and then also recognizing, like you discussed earlier, the unlimited capacity of the computer. Computer. And how does that change the computer, the outcomes in these negotiations? So let's just start off with the pros and cons. Where should we start as it relates to the pros and cons of negotiating with a computer?

Yeah. So if you think about it, there are some fundamental differences in negotiating with a computer. The bot, if it's not the general intelligence AI, which which it is not, then it can never strike an out of the box solution. And this is for humans. It doesn't have the human ingenuity and it doesn't create the repo with the other person. So these are the cons. It can't play golf and drink wine with some of these vendors. And that's why we're focusing on the, the long tail vendors, the unmanaged tail spend. So these are the cons.

That's really interesting. It makes sense. I was going to say it's the human aspect, right? The things that we can do to build rapport and bring people together. That's what's missing. And that makes a lot of sense. But at the same time, it seems as though what we're doing is we're making sure that we are only utilizing the bot in situations where these are small spends. It's not a big deal for the company. It's one of those smaller deals that have to happen, but we want them to happen in a way that doesn't take away bandwidth from the negotiators for the bigger deals. Is that right?

Yeah, exactly, that's correct.

And so we have efficiency as one of the pros, politeness as one of the pros.

We.

What else is there?

There are interesting non intuitive things when you think that a computer can negotiate with unlimited capacity. And I will touch on the third point of capacity a bit here. So firstly, humans are not built to think in a multi dimensional contract space. If you think, I mean, if I was thinking back when negotiating big deals that I had, then how it really usually goes is that you identify three most important points, try to get those as, you know, try to maximize those, and the rest is kind of a blur. You try to maximize whatever option you can out of those. In reality, there is probably a combination of items or which is perfect for one party and which is perfect for another party. Now the problem is that if there are 30 items to negotiate about, the combinatorial explosion becomes so big that people can't. People aren't efficient in finding the perfect combination for both sides. And at the same time the computers are perfect at it. So it'll find a better outcome for both sides.

Does your company invest in professional development training? If you believe that your team would benefit from a negotiation workshop, all you need to do is go to our website, fill out the workshop request form and then we'll set up a time to chat. These workshops are completely customizable and we've done them all around the country. Negotiation and conflict resolution Skills are beneficial across all professions, but they're especially useful in procurement, purchasing, sales, sourcing, and contract management. Our calendar is filling up quickly, and we even have some workshops scheduled for next year if you think you might want one. I'd suggest reaching out soon so you don't miss out. Check out the link in the description to learn more. And now back to the show. This is so interesting. And so let's actually break this down, because I've never had this experience. So let's say I'm negotiating with this bot and I'm selling, let's say, paper towels to a company, and I'm making up numbers because I've never had to do paper towel sales. But let's say $4,000 a month is the spend for paper towels for this big company. And so I'm selling it and I say, okay, we would actually like $5,000. That's my anchor, and that's how I start what would then transpire.

So the bot will guide you through the negotiation. It will start by setting the scene by. By explaining how it works. It will then try to find out as much as possible about you. So you're selling the paper towels. What would you like to improve in the current contract? How are the freight terms? How are the warehousing terms? How are the payment terms right now? So it'll try to find out as much as possible, and then it'll give an ask. So the company, our customer, would like to maximize this particular item. And then there is a negotiation that follows. And the computer will use data, as much as it knows about the vendor and as much as it can gather during the negotiation to find a perfect combination of liability, freight terms, warehousing terms, payment terms, price, and so on.

This is fascinating. So in that case where I said, okay, maybe the standard rate is 4,000 and I'm starting off with an anchor of 5,000, then the computer would counter with another number or a package offer and then substantiate what it's asking for.

Yeah. So essentially the computer's task is to look at the current deal we have and move it towards the Pareto optimal line. So you might be familiar with that. So the Pareto optimal line is in negotiations. If you imagine an outcome of a negotiation as a dot on a graph where one axis is the value for one side and the other axis is the value for the other side, then there is actually a curve. And if a negotiation outcome is on that curve, that means that one party cannot get a better deal without hurting the other party. And that means the deal is now Pareto optimal. Now it's not win win, it doesn't have to be win win, but it's Pareto optimal. And every deal in the world should be Pareto optimal. So now the computer's task is, is to find ways to push this dot, this outcome of the negotiation, towards the Pareto optimal line, usually creating value for both sides. And there are a number of ways to do that. So one strategy that we use, for instance, is multiple equivalent simultaneous offers. We know, we know the value function of our customer. We call this the value function. This is essentially all of the interactions between all of the positive and negative levers. And so from these levers or from these negotiation items, the bot will pick two offers which should be as different from each other as possible, but both are beneficial for our customer. And if the vendor chooses one, it is obviously better for them as well. And we've created value for both sides. Now we move this dot towards the Pareto optimal line. So this is a bit of kind of negotiation, sciencey explanation, but I hope listeners of this podcast were able to follow.

Oh, they followed. I know they get this, and this is really exciting for me because essentially when I started this conversation, I was ready to be overwhelmed by the science and the technical aspects. But as we started to talk more, what I'm recognizing is that essentially you've created the perfect negotiator, because the bot that you've created isn't trying to take advantage of the other side. It's using the core fundamental negotiation techniques that people, humans would utilize in best case scenarios. And so they're gathering information, they're asking questions, they're figuring out what it is that they need, understanding what the other side needs, and then find, finding those combinations. And then even strategically, the different tactics that are used. So offering package deals, that's a really powerful technique that people use. I don't think they use enough. So we say, I don't know what people want. I'm going to get different options that are equal value to me. And then I get a very clear signal of what's important to them.

And I've tried to use this myself in negotiations. Multiple equal simultaneous offers. It's very tough because, you know, your stress level is high during a negotiation and you have to calculate what's kind of what is the combination of those items that are both good for me that I could offer. And they both should be as different as possible from each other. So I get information out of what, what the other party chooses. It's really complicated. It requires a lot of mental effort. So the computer is way more efficient at it.

This is great. And so we are using this right now for the smaller negotiations that we have, the smaller vendors. Do people also utilize this in some capacity with the bigger vendors as well?

Not right now. Maybe at some point. And at some point it could also advise human negotiators. So in some cases, people want to keep the human relationship. So the bot could provide information during the negotiation for the negotiator on the enterprise side. So this is one option, but sure, as the system evolves. And one huge benefit of using AI in negotiations is that every negotiation makes every other negotiation better. So essentially any customers, any customer's negotiation makes the negotiations better for any other customer. So we don't store the business sensitive information and we don't collect that from negotiations. This is highly confidential. And this we delete right after the negotiations. What we do store is the negotiation psychology bit. So we learn how people negotiate. The system learns what tactics and strategies work, what ways of phrasing sentences work, how many seconds does it takes to say no versus a yes. So all of these things make every next negotiation better. So the bot can choose the right tactics and strategies dynamically based on what it learns.

This makes a lot of sense. And I. I'm a chess nerd. I talk about this all the time. On the podcast. I was reading Garry Kasparov's book, and he was talking about his first few chess matches against Deep Blue, which was the artificial intelligence created. I think it was IBM, where it was the first time that a chess master went head to head with a computer, and then the computer won after the fact. Gary recognized that there were a couple times where he made a mistake. It wasn't necessarily that he was outplayed, but he made a mistake. He didn't calculate it correctly. And so now we're at a stage where since the computer can do infinite calculations, the computer is going to beat the chess master. But what they found is that if you pair the chess master with the computer and combine those various types of intelligences, then the combined effort is greater than what they would do individually, because the computer can have unlimited calculations. And then the human has ingenuity, that creativity that would otherwise need to be programmed. And so it's interesting to think about what the future of negotiations could be when you have the bot, the artificial intelligence, paired with an experienced negotiator, actually at the negotiation table.

Absolutely. So this is very exciting for the future. The bot right now learns from These long tail negotiations. And it could provide a lot of value to strategic and mid tier negotiations. And since those are also more complicated negotiations than the long tail ones, this is great.

So interesting.

And so there are a couple of other kind of non intuitive pros to having a computer negotiate. One of them is that the computer negotiates parallelly, not sequentially as a human. So it can hold 500 negotiations at the same time. And thus it can do things like dynamic pricing across all negotiations. If there is a limited resource of some sort, it can use it dynamically within the negotiations. And this resource is. And all of the other kind of numbers and the values of all of the other levers, they change automatically in other negotiations based on what deals have already been accepted. So this is another kind of non intuitive advantage of that. And it also brings speed, of course. So for a human to book all these calls to have those negotiations and the human can't do them parallelly, it takes a lot of time. For a computer, it takes exponentially less time. And one other benefit I would bring out is that the computer has no biases. So we don't learn negotiating from humans. We learn how people negotiate. And this is different. So our negotiations are actually designed by the best negotiation scientists in the world. So Jean Brett from Kellogg School of Management, she's a professor of international negotiations, is our advisor. And with, with people like Gene, we design these negotiations. So we always use the best practice strategies and tactics. And because these people are professionals, they're able to realize biases and cultural differences. So the typical problem with AI is that it also learns the biases of people. If you feed the AI with data about negotiations, it will automatically learn all of the cultural differences as well, which is not very beneficial. That's why we do it in two tiers. We design the negotiations with professionals and we then tweak them based on AI, based on what it has learned, based about people.

This is great. Yeah, this is really opening my eyes to what the possibilities could be. And in that last section, I think we handled pros and cons. We talked about whether or not it can expand the pie more. When we talked about creating these deals that are Pareto optimal. And then we also talked about how the capacity impacts the negotiation. And so I guess really the next question is the future. Because right now you said you're the only company that's doing it, which is great to have a monopoly, but as things start to develop more, I'm assuming the market's going to change. And the way that we interact with each other, that's going to change too. And maybe we see a future where the negotiation books that we're teaching lawyers and people in the MBA courses includes how to negotiate with a computer, with the assistance of a computer, those types of things. What does that future look like?

So the future is very exciting. I like to bring the analogy of self driving cars. If there's only one self driving car on the street, it's not so beneficial as if there were. As if most of the cars were self driving. If most of the cars were self driving, they could communicate with each other and make the traffic much more safer and much more smoother. And the same with negotiations. So we're actually what we're seeing now because we're mainly working with large Fortune 500 enterprises. Most of our revenue is coming from these companies. We see that the value chain is taking notice and they want to use the system as well. And it won't be long until a bot can negotiate with a bot. And there are a number of additional advantages there because the bot can definitely find the Pareto optimal outcome in a negotiation every time and very fast. And the bot doesn't have a limitation on renegotiations. It could renegotiate a large deal every day if it needed to. Right now, if we're looking at the stock market and the commodity prices, you know, the price of copper has gone down 20% and then has gone up 20%. Maybe the copper based product prices should have been renegotiated. And if a computer negotiates with a computer, they could literally renegotiate deals based on every movement of a commodity price.

This is incredible. This is really, really interesting. And now that potential future where a bot is negotiating with a bot finding those optimal outcomes, it makes me wonder about those no deal scenarios. Because thus far, a lot of the times when we've been discussing it, we've been talking about what happens when we find the optimal outcome. As we know, sometimes in negotiations we run into a situation where not all deals are meant to be made. There's a no deal option. So if a bot runs into that type of situation, how does it respond?

So the good thing here is that we're not replacing humans in this equation. We're negotiating the contracts that aren't negotiated since they're long tail and they're sort of neglected. So if a deal, if we renegotiate those deals and if there is a fraction of those deals which aren't able to reach an outcome, then there are two possibilities. Either the old deal remains or we take the it's called human in the loop. We take a human to continue, and then the humans can use all of their human ingenuity to in the exact place where it's needed. Not on the 90% of these long tail deals which the bot can renegotiate successfully.

Yeah, this is great, Martin. Thank you so much for sharing this with us. We really appreciate it. But before you go, can you let the listeners know how they can get in touch with you and tell them a little bit more about PactMobile? Sure.

So we're on Pactum.com and everyone can leave their email in there and we'll be in touch. And my personal email is martinactam.com so I'll be happy to. If you also mentioned that you've got the contact from this podcast, you've definitely gotten my attention and you'll take special care of you.

Thank you so much, Martin. I appreciate it and I hope you have a good rest of the day, my friend.

Thanks so much, Kwame. Thanks for inviting me.

Congratulations. You've just joined an elite club by listening to a full episode. You're now officially on the Negotiate Anything team. So welcome aboard. What most team members do is they support subscribe to the podcast because that allows them to automatically get the latest episodes of the show. The best things in life lie on the other side of difficult conversations. Keep learning, keep practicing and keep getting better. Your relationships will improve, your career will soar, and you'll have the confidence you need to get the most out of these crucial conversations. Again, thank you for joining the team. We're excited to have you and I will see you in the next episode. Episode. I'll catch you later.