Revolutionizing Procurement: Pactum's AI-Driven Innovations in Business and Government

Unlock the future of business negotiations with Kaspar, the visionary co-founder of Pactum, as he shares his groundbreaking journey from pioneering Estonia's e-residency to revolutionizing the world of AI-driven contract negotiations. Discover how Pactum is transforming procurement processes for giants like Walmart, using AI agents to autonomously interact with suppliers, optimizing terms, and delivering significant savings and efficiency. With Kaspar's insights, you'll grasp how autonomous agreements are redrawing the landscape of procurement, resulting in win-win outcomes for enterprises worldwide.

As we navigate the complex world of government procurement, Kaspar sheds light on how AI can redefine how public funds are utilized, making spending more transparent and strategic. Learn why traditional government processes are ripe for disruption and how hesitancy from startups is a barrier that can be overcome with advanced AI tools. Explore the shift from indirect to direct spending in manufacturing and the potential of AI as a strategic ally in decision-making, making companies more agile and competitive. This episode promises to broaden your understanding of AI's potential as both a transformative tool and a strategic partner in business and governance.

Transcript

Speaker 1 0:01

Hey everyone. Really interesting topic and discussion today with an innovator when it comes to AI-driven contract negotiations. Who would have thought of that? From Pactum Kaspar, how are you?

Speaker 2 0:14

I'm good how are you?

Speaker 1 0:16

I'm well. Thanks so much for joining from Estonia. Would love to hear about your personal background and the mission at Pactum. Give us some background on the big idea.

Speaker 2 0:28

I'm happy to. So my background I work for the government of Estonia. I established an e-residency, become a digital citizen of Estonia and was running that for five years. So building AI for government, making government to act like a startup, for government making government to act like a startup. Five years after that. Then I stepped down and now last five years building Pactum. So we have built the largest AI agent platform for enterprises. So we have AI agents like chatbots who reach out to suppliers of large enterprises and conduct autonomous negotiations so go through agreements, deals, preferences of each other and agree on new terms and sign those deals autonomously, so that no human involvement is needed and the AIs are growing the business in the background.

Speaker 1 1:19

Amazing. So what inspired the vision behind Pactum? It's a fascinating idea. Now, of course, there's lots of excitement around autonomous agents, but what was the original vision and how did it evolve over the last few years?

Speaker 2 1:33

Yeah, it was five years ago, so back then we didn't have large language models and people really didn't appreciate chats like that, right. So the idea came together by three founders so, besides besides me, my older brother, christian, who was ai leader, starship technologies, and and martin run, who sold his last company, and he brought the topic of negotiations to our table that negotiations are like something that humans are not excelling at. Right, we have these biases, we don't have time, and when we merge negotiations with ai, this came the autonomous negotiations, the AI agents that do negotiations by themselves. And back then, of course, there were too many people who believed that AI can do business autonomously, that suppliers or other parties would trust their business relations to AI. It was called nonsense because you know you do business with humans and you negotiate with humans.

Speaker 2 2:30

But for us it was very clear. We did first pilot like two months after going establishing Pactum, and it worked out that there was a limousine car owner who got 80 euros better deal and we saved both parties some value, and autonomously. So, and then we knew that, hey, there is win-win deal possibility for two parties and it can happen without human beings involved. So it was a great milestone and after that we became Walmart, became our largest customer and after that now working mostly with enterprise customers.

Speaker 1 3:05

Fantastic. So maybe dive in a little deeper to how Pactum leverages AI to handle complex negotiations and how does it work versus the traditional approach we're all familiar with in the B2B world.

Speaker 2 3:19

Yes.

Speaker 2 3:20

So there are two main scenarios. One is deeply inside procurement end-to-end process. It's called source-to-pay or procure-to-pay, and inside that process, basically you have SAP, ariba or Coupa type of systems and then negotiations are triggered automatically. So if you want to buy something let's say an iPhone case and you want to buy them let's say a million of those then there is a requisition created in spend systems and instead of just purchasing that, negotiations automatically happen with this supplier or other suppliers, going through different terms like when can we produce? The quality matters, the price, the payment days and through negotiations, trying to find a better deal and agreeing on that deal automatically. So this is one part of the process, ren, a contract purchase of some item uh, price lists in car manufacturing, for example, very popular, like thousands of lives of price lists that you need to go through all the time and humans are not capable of inside procurement process.

Speaker 2 4:20

Other is like strategic initiatives. Let's say, president, president-elect, donald Trump comes in and says that Mexico has 30% more tax on all import goods. What do you do then as a, let's say, retailer, like will you increase prices 30%? You need to find alternative suppliers. You need to negotiate deals with everyone, again with local suppliers, do you kind of commit on volumes with new suppliers elsewhere, and so AI agents like Factum are used across large enterprises for these massive strategic initiatives Like let's go to these 10,000 suppliers, let's renegotiate these deals, let's introduce new allowance, new payment, let's commit volumes of logistics rates, because environment is changing rapidly around us and you need to react very fast to these changes, otherwise competitors take your place.

Speaker 1 5:12

Amazing approach. Let's talk about customer impact, clearly important for an early-stage company. Do you have any examples, stories or anecdotes on how you're transforming procurement, or vendor negotiations, as you describe them for clients?

Speaker 2 5:37

our buyers, who are the owners of this extra value. Because if you know that you used to pay a million for this dollar, uh, for this iphone, um, iphone, uh, iphones, let's say, now you pay 100k, you say 100, 100k, right, uh, or other way around, and, and hence you can measure how much each buyer generates savings. This is pure bottom line savings, course. There are like automation savings that buyer has more time to deal with strategic deals, buyer has more time around the bureaucracy of managing all of these suppliers and stuff like that. But mostly it's about pure savings in value in dollars for bottom line, for the buyers and for the other side. They need to get something as well, otherwise, you know, you just refuse, right, you don't do deal. No one can force you to do a deal right.

Speaker 2 6:27

So how the magic happens through Pactum is that we find some terms that are valuable for the supplier. Some, I don't know. Some prefer early cash over longer payment days, you know. Some prefer commitments of volume of business for next year. Some prefer they can expand business, new product lines. But there are various different use cases. There are different incentives and growth plans and when we offer, then we get something in return and then the AI is valuing this like what are the current terms, what are the new terms? And, on net, how much value did we save when we count everything together? And if savings are over threshold, then we agree and we sign the deal well, fantastic.

Speaker 1 7:04

And so what is the human ai collaboration methodology here? I mean what role humans play in conjunction with your ai? Is it sort of human in the loop, is it collaborative, and what's your vision there today and for the future?

Speaker 2 7:21

Yes, for today, in some use cases, no, humans are in the loop. Like very low dollar spend, let's say you just requisition order goes through and then automatically goes through negotiation and we update the requisition price, for example. But for deals like 500K up, $1 million up, especially if it comes to $10-20 million deals, then usually humans are in the loop. So you can modify the negotiation parameters per negotiation, like what are terms, what are the incentives, checkbox list, for example, what are our preferences of anchors, acceptance levels. All of this is kind of configurable then by the buyer per negotiation level. So there, humans usually want to be in the loop.

Speaker 1 8:07

Okay, got it. And how do you think about scalability? I mean supply chains. This is trillions of dollars of economic activity. There's lots of nuances there. How do you think about this AI scaling across if you're a Walmart or an Amazon and you put AI into this modality?

Speaker 2 8:27

Yeah, so in procurement landscape there are two terms indirect spend and direct spend. So indirect spend is everything you purchase for yourself, like IT licenses or cleaning services around the office, and direct spend is you purchase something into production or into resale. So in indirect it's today kind of plug and play you use Ariba or Coupa spend management systems, you activate Bactium and negotiations can go out after some configuration settings. But in direct spend it needs like a couple of months of configuration for manufacturing, for retail, for logistics, lines of businesses. They are more strategic. But all in all, like AI, as we now have proof of five and a half years, is able to conduct negotiations across the spend, across direct and indirect, across industries and across geographies, for example, from australia to us, to germans, to turkeys, you know, you name it. Different cultures have different preferences of of the negotiation, language and and tactics, kind of. But all in all we have now proof that across the globe people not are willing only, but sometimes even prefer to speak to AI.

Speaker 1 9:44

Fantastic, and what are some of the challenges when it comes to adoption? I mean, if you're a company in manufacturing, you might not be at the leading edge of IT or AI. What are some of the common challenges or misconceptions or other roadblocks that you have to overcome to adopt AI for contract negotiations?

Speaker 2 10:06

Yeah, clearly it's a change management problem more than it's a technological problem, like how do you get by in a cross-tech kind of to trust now AI agents as kind of our colleagues for our workforce, right, like we are going to coexist with robots, not like terminators so much, but like virtual chats and agents who work with us, sometimes internally, sometimes with clients, sometimes with suppliers and and what's the role of them and what's the role of us? So it's a real kind of change management happening. But in some sense I find because I have sold to enterprises before as well and worked for governments, and it seems to me that it's easier to accept AI agents than it is for European companies, for example, to accept any other type of technologies. Usually technologies are sent to startup hubs, innovation centers or hubs where you know we need to do innovation as well because it's cool.

Speaker 2 11:11

But now with AI, people have realized from sea level that, hey, we won't survive. It's mandatory If we don't survive. It's like mandatory If we don't do it, others will do, and they bring AI and they outperform us totally. So it has become like really strategic bet for even in Europe for more progressive enterprises that we need to double down on AI or we're not here for long.

Revolutionizing Government Procurement With AI

Speaker 1 11:32

Yeah, no, it's a great point. The time is now. Let's talk about your vision for the future. I'd be really interested on the government side as well. Everyone in business knows Estonia is kind of the gold standard for e-government, digital government. You were way ahead of the curve there. Even now we're just getting caught up in the US talking about government efficiency and Doge and these kind of things. The US talking about government efficiency and Doge and these kind of things. What is the opportunity for governments to leverage AI for contract negotiations and procurement? Is this going to revolutionize the way governments spend and procure?

Speaker 2 12:07

Yes, I think it's a great example to draw how we build e-residency as a government startup, kind of how the alliance of trump and president, president-elect trump and kind of elon musk can kind of do the same in us government, kind of how to kind of build us government as a startup in a sense like, and what are the steps? And? And it's just super fascinating because because first, when you kind of release data, let's say let's speak more specifically about procurement, because current procurement is also like exciting things. Usually procurement is quite boring for many, but government procurement is very exciting for everyone because you can't replace your taxpayers' money and you want to see how it's spent, who spends how much, and you can soon realize how silly things maybe people spend money on and has been for decades. Right, because it's not visible. So if first of all, you can make spend visible, the same as in enterprises happens like you make spend visible and you realize what's going on in the company. Spend visible in government side kind of it opens new dialogues and discussions where and how much we should actually spend, and then what it opens, then it's realization oh, we should spend less. Okay, and how do you spend less? Also in government? You need to start adapting to technologies like in private sector, right, you need your spend management tools, right, like it's silly that in government you don't use the best tools out there that private sector has been using for decades already.

Speaker 2 13:39

First is, put the basic systems in place to understand, but then improve processes of procurement, because government procurement, I'm telling you even in estonia, is not that great. Right, you have this long process and then, like minimum bidders, win their beats because they offer less and you get worse results through that and and and it's so, so long process that both us government and the stoning government have asked us back to to offer back to them as well, and so far refused because it's just so long process for a tech company you don't want to be involved. So, even if your startup doesn't want to sell their products to government, it shows how messed that is right. So there needs to be that disruption, and AI, of course, is a layer. Maybe a year after core technology is in place, core data is in place, and then we can kind of adapt AI layer on top of that to kind of help, like Bacton does help, with autonomous negotiations and autonomous procurement.

Speaker 1 14:40

Fascinating and important topic. What are you personally, what are you and the team, excited about over the next few months? Next year, 2025, you obviously have a lot going on. What's what are you most excited about? I mean, you know, is it some of the go-to-market activities? Roadmap features what's on your mind?

Speaker 2 14:58

Yeah, mostly last few years it's been about indirect spend, the less strategic spend, and we have been very fully standardized and automated end-to-end process there. And now we have entered a direct spend, especially, let's say, in manufacturing, in production lines, and that's super fascinating because you know if one car part is not delivered, the entire production line stops Right. So it can be a small supplier, but it's every. Every supplier is very important there. So, and to have AI agents to negotiate these deals and affect the direct line of business, which is like 60 to 80 percent of the total revenue of a company, and then AI becomes strategic revenue of a company. And then ai becomes strategic.

Speaker 2 15:40

And once now ai becomes strategic, then, year or two after that, it starts thinking and advising strategically. It starts thinking as an our board member, colleague or our c-level exec next to us, not only to deliver transactional savings and low tail supplier negotiations, but actually help us to conduct strategies for the company for future, which markets to go, which suppliers to reach, and then execute on those plans. The next day and then next week come with new strategy, new execution, become very agile and very strategic level. With new strategy, new execution become very agile and very strategic level and this kind of changes the entire way how I think companies evolve, kind of that they become like so fast growing, so agile. Imagine of these thousands of agents who are autonomously building up the companies, while humans have different roles there, while it's kind of humans together with AI that will race against other companies and kind of the growth numbers I predict will be very different for those companies who are early on adapting AI.

Speaker 1 16:51

Fantastic. Well, thanks so much for joining. I never thought procurement was a topic that would be the least bit interesting to me. You've changed my mind. It's a fascinating opportunity that you have ahead and you're reinventing this space. It's fantastic to see. Thanks so much for joining, kasper. Thank you and thanks everyone for listening, watching and, as always, sharing and take care. Bye-bye.