Why AI Agents Can Beat the Incumbents

a16z’s Seema Amble and Elena Burger sit down with Lio co-founder and CEO Vladimir Keil to ask where AI-native startups have an advantage when incumbent software companies already own the customer, the data, and the system of record. Their answer comes down to the work that happens outside those systems. In procurement, a final price in an ERP can hide hundreds of emails, spreadsheets, supplier conversations, engineering analyses, and decisions across legal, finance, and operations. Vlad explains how Lio uses multi-agent systems to take on more of that end-to-end work, from sourcing and RFQs to negotiation, shipment tracking, and invoices. They also discuss how enterprises learn to trust agents with increasingly consequential decisions, why the last 20% of an internal AI build can require most of the effort, and what happens when both buyers and suppliers have agents working on their behalf.

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:00Speaker 00

  If you want to build an aircraft, you need to procure thousands of suppliers. Someone sends a confirmation of like, hey, sorry, like this part is going to arrive two weeks later. And if they missed this email, hundreds of millions of them.

• 0:10Seema Amble?

  Procurement historically may have been more in a box. And now it's like, okay, it's touching legal, it's touching finance. It's touching a bunch of different software systems and people. The opportunity for the AI native startup is to say, we're going to own that entire end to end arc.

• 0:24Speaker 00

  No company, no enterprise starts with fully autonomous negotiation agents from day one. Why? Because they don't trust us and they don't trust the technology from day one. And by having this even the loop approach, we are feeding our agent with all the feedback and all the learnings. And then they suddenly trust us for like 10K negotiations, 20K negotiations, 100K negotiations.

• 0:42Speaker 02

  When you think about what a durable vertical AI company looks like, what are the qualities that you look for? It's really, really hard to forecast your mode going forward. If you look back at all the best businesses, at the early stages, they were.

• 0:57Erik Torenberg?

  If the incumbent already owns the customer, the data and the system of record, where does an AI native startup have an advantage? In this episode, Elena Berger sits down with A16Z partner Seema Amble and Leo co-founder and CEO Vlad Kyle to answer that question through one of the most complex parts of the enterprise, procurement. They get into why so much of the actual work happens outside the system of record across emails, spreadsheets, contracts, engineering data and conversations with suppliers. And Vlad explains how Leo is building agents that can coordinate that context and increasingly take on the job end to end. They also discuss how you earn enough trust to let an agent negotiate on your behalf.

• 1:39Erik Torenberg?

  Why building an internal AI tool can be deceptively easy until you hit the exceptions? And what changes when agents eventually sit on both sides of a transaction?

• 1:51Speaker 02

  Welcome back to the A16Z podcast. I'm Elena Berger and today I'm joined by Seema Amble, a partner at A16Z and Vlad Kyle, co-founder and CEO of Leo, which builds AI agents for enterprise procurement. Seema, you recently wrote a piece called "The incumbents are coming" and then asked a question facing almost every AI application company. If an established software vendor already has the customer and the data and the model can work across its tools, where does a startup have an advantage? And we have Vlad here and Vlad can really help us understand where a startup has an advantage.

• 2:30Speaker 02

  So I think a good place to start is the cases for and against the incumbents. So if a company can connect a capable AI agent to the software it already uses, where does another application actually have a use case and an advantage there?

• 2:49Seema Amble?

  Yeah, let me back up and frame it up a little bit. So historically, we thought about there is the incumbent and then there's the startup. And it's a fight between distribution and innovation, which to take my partner Alex Rampell's phrase. However, now there's like this third piece, which you're pointing at, which is the incumbent can layer on one of the models on top. And then there is a much more formidable competitor in the market. So why do you need an AI native startup if you've got ClaudeForce, which is taking Claude plus Salesforce and putting the two together? And then you've already got all your data and your employees are all used to using the product. So why another product? I absolutely still think there's obviously still a case for the AI native startup.

• 3:32Seema Amble?

  And it really centers around the fact that the legacy incumbent is limited to their system of record and that record that they have. And they're not completing the end to end job. So let me put that more concretely in an example. So say you're a customer and the customer calls and says they got charged after they got canceled. Resolving that cancellation history isn't just the, you know, customer going into the chat and saying, hey, I got overcharged. And the response there, it has to hit billing. It has to look at all the chat history. It has to look at the contract. That's not one system of record. That's the knowledge around that customer and everything it touched. And that's something that one system of record wouldn't touch.

• 4:16Seema Amble?

  However, the opportunity for the AI native startup is to say, we're going to own that entire end to end arc. So that could be illegal. So owning everything from a brief all the way through trial. Vlad can talk more about procurement, but it's really the concept of owning the end to end work.

• 4:33Speaker 02

  Yeah. Vlad, do you want to talk about where that does show up in procurement? Like where a existing incumbent plus a model just is insufficient? And what have you seen just with the companies that you work with?

• 4:47Speaker 00

  Sure. So when we think about procurement, I would assume that most people think about like prices, right? So like, what is the end price that we negotiated on? And surprisingly, the record looks always very, very simple and very easy. It's just 8K. That's the result, for example. And I mean, it's the same for sales, right? So even if I come to Seema and tell her like, hey, we now partnered up with another enterprise. Look at the signature here for like this contract. It looks like very easy, but like Seema doesn't see like all the work behind it, right? So like there are probably like 30 stakeholder meetings happened, 500 emails, 20 Excel sheets. And that's the same for the counterpart procurement, right? So you see in your ERP system, 8K for aluminum,

• 5:30Speaker 00

  but you don't see that maybe the supplier did like a pushback and asked for like 10K. You don't see that like a cost engineer had run three weeks of Excel sheets and 3D modeling to find out the prices of the part and everything else. So that's what we see. So like most of the work in procurement actually happens outside of this ERP or any system of record.

• 5:51Speaker 02

  Yeah. And I'm sure in the workflow that you described, an agent can do a huge number of things and different kinds of agents can do a large number of things too. And I actually think that's a good bridge into the next question, which is, you know, a year ago, I think a lot of the incumbents were releasing chatbots. And that was kind of the extent of what you would see. But Seema, in this piece that you wrote, you lay out four different kinds of agents. Retrieval agents, which are kind of the chatbots that we're familiar with. Process agents, policy agents and principal agents. So can you just walk us through all of them and explain kind of what changes in the kind of judgment that's necessary across all of them?

• 6:32Seema Amble?

  Yeah. So a year ago, I made this meme, which was the slap on a chatbot strategy, which is essentially like all the incumbents effectively had a chatbot that sat on top of the system record, which you could chat with to retrieve information, maybe do some like analytics. And that really was in that first bucket of what the retrieval agent is. And maybe let me walk you through an example of what each of like the retrieval, the process, the policy and the principal, back to the customer support example, just because it's very, it's easier to understand. So imagine if you're a customer and there's a service outage and you're calling in to say, hey, I want to get compensated for this service outage. The retrieval assistant, which the incumbent may have, is going to be able to pull up, yes, this was what the contract term said.

• 7:16Seema Amble?

  And yes, there was an outage and just verify that information. It's pulling up information about the customer. It's in the database and it's just sharing it back and maybe synthesizing. The second step in the agent sequence is the process agent. So that process agent may be able to pull through an approval of credit and say, okay, based on our policy handbook, it was out from these dates. Therefore, we say you are entitled to money and it can just apply the bill and just do process. There's no judgment involved. Then as you keep going to the policy agent, so the policy agent isn't just going to apply the process, but it's going to say, okay, in this situation, it was out for 20 minutes.

• 7:59Seema Amble?

  That's enough to be considered a significant outage. And they're applying that judgment because there isn't really a strict definition around it. And then the last stage, which is the principal agent, you're actually weighing, okay, should we offer more compensation? Because that was a pretty terrible outage and we want to reserve the relationship. And it's worth doing more beyond even what a process or policy. But the importance of these four things is that most incumbents started out in, if at all, in bucket one. Now at least they're marketing that they are moving towards process and policy, meaning they're be able to apply more judgment. And you know, if you look at what they've launched, these are workflow agents that will help you get a document signed

• 8:43Seema Amble?

  or input information from a transcription or things like that. They're very much still limited.

• 8:50Seema Amble?

  Two, I would say retrieval and a little bit of process. They're not gotten into more judgment. And I can get into why, like, there are all sorts of incentives that are preventing them from that. But the incumbents are trying. And I think what they're not able to do, they're like, some of them are thinking, okay, let me partner with OpenAI, Anthropic, one of the labs, and try to build out, take that model capability and complement what they have and superpower it.

• 9:14Speaker 02

  Yeah.

• 9:15Seema Amble?

  In charge, I should say.

• 9:16Speaker 02

  Yeah. Well, do you want to say why sort of some of the incumbents are holding back?

• 9:20Seema Amble?

  Yeah, okay. So I would say they're holding back, but they're held back.

• 9:24Speaker 02

  Okay.

• 9:25Seema Amble?

  Yeah, I'm sure they want to be full force, but there's probably two pieces. One, they have this advantage of distribution, right? They have the customer trust, which enables them to then sell more products with the customers. So take Salesforce. When Agent Force launched, it was very easy for customers to say, yeah, I'm going to sign up for the Salesforce agent, especially if it was offered at almost no extra cost. And so they have this trust, they have the distribution, it's often pretty seamless to turn on the product. The flip side is there's all these internal incentive issues, right? Which is, if you start getting into more complicated agents, it's an internal conflict with an existing product that's offering workflow versus you resolving the work.

• 10:07Seema Amble?

  And those are two products that are, if you're solving a customer support issue end to end versus providing a workflow for a support agent, a human agent, those are different buyers. How do you, you know, those two teams are in conflict. And then on top of that, I think from a sales perspective, like, what are you selling to the customer? And then oftentimes in a classic, like incumbent issue, right, is like, there's two, two VPs, right? And they're different orgs and they're selling different products. And like, they will never be able to figure out what the, the right set of incentives and the right person to sell to. But anyway, so there's all these sort of classic incumbent issues, I think, that the incumbent also runs into.

• 10:45Speaker 02

  Makes sense. Vlad, where across this, this kind of spectrum of retrieval agent, process agent, policy agent, principal agent, where, where does Leo sit?

• 10:57Speaker 00

  Yeah, so like, we span across like, I would say like, all of those categories. And it really depends on the complexity and the risk our agents take, right? So yeah, sometimes we can, like, we already have use cases where we run fully autonomously. Sometimes you have the human in the loop. It really depends on like, the budget approval, how like, how complex this is, how risky this is. But maybe like, coming back to what Seema said about trust for, for, for the incumbents, like, you talked about like, internal trust. I think there's like, also an external trust thing, right? So like, how do you convince someone to go through from, hey, just like an agent that retrieves some information and maybe runs processes to do, to do like, something completely autonomously is like,

• 11:40Speaker 00

  obviously, there's like, a product component to it, but mainly is a people component, so they need to trust you. And like, we as a startup, scale up, we have to earn the trust. Incumbents already have the trust, but this also means they, like, they can destroy the trust if they ship something like, too early and the product maybe doesn't work or it's bad or it's like, works decisions in a bad way. So that's, that's, that's what we, we can do. And then interestingly, what was like, really surprising for us, like, those process agents actually became kind of like a side quest for us. Because it, like, when we talk about, when we look at the invoice process, or like invoice agents as a process. So you, we retrieve some informations, you match this across like, other documents,

• 12:24Speaker 00

  and then you push this back to SAP or Oracle, like a very clear, clear process. And then we figured out, okay, there's like, 100% of the software market, right? So that's, that's invoice software. That's how you build it today. But it's actually only like, 20% of the work of the job to be done of the problem. Because 80% of the problem is like, what if, like, what if the invoice is fraudulent? What if there is like, a mismatch? Like, what did, what is like, if we don't take the happy path? And so we like, very fast shifted to, to the next step of agents, like, doing those, doing those exception, um, exception handlings.

• 13:05Speaker 00

  And we convinced the customers by, I think we got like, lucky being like, always slightly ahead of the curve. So we were able to, to pitch the next generation of agents. So for example, like, when we started out three years ago, it was like, just retrieving a document. Like, not, like, not impressive at all today, but like, three years ago, this was in, like, crazy impressive. So we pitched this to customers. We find out, okay, that's a real problem, that's a real use case they would pay for. And then we were able to ship this some weeks later. And in the same way, we're doing this now for the next step, for process agents and for like, fully autonomous agents and for long-running agents. We are pitching this to them, finding out it's a problem, and then we're able to ship this very fast.

• 13:48Seema Amble?

  I think an interesting point on trust is this internal versus external trust. The other lens on that is, yeah, so you need your customer to buy the procurement software and trust to use it for their internal processes. But one of the really interesting things when we first met Vlad was that they're doing, they're also doing the negotiations. So you have to trust that the Leo agent is going to then interface with a third party. And, uh, and there is that piece of trust. And of course, I think a lot of people feel burned by the incumbents and like, you know, they're pretty limited and haven't been able to do what they've marketed, they've had in the past. So, but putting that aside, like, I don't know, maybe Vlad, I'd love to hear a little bit how you convinced

• 14:29Seema Amble?

  the customers to, to trust an AI agent to now take on negotiations.

• 14:34Speaker 02

  Yeah, and can, when you do that, can you also just like, paint the picture of like, what's involved in procurement? And who, who are your customers? And, and what kind of sort of legacy systems are they used to?

• 14:48Speaker 00

  Yeah. So like, when we, um, when you think about procurement, you like, maybe just think about like, purchasing or like, you, when you look at like, B2C world, like, just, you buy something. But actually, it's like, a very intense process, process which like, runs the economy, right? And it includes multiple stakeholders and a lot of stakeholders and a lot of, um, a lot of departments. Legal, cost engineering, obviously procurement, finance, all of them have to work, have to work on this, um, on these decisions. And essentially, like, when we, um, like, I think, I think the reason is like, how, how we convince them is like, what I already said, said earlier on.

• 15:29Speaker 00

  So like, we were like, always a little bit ahead of the curve. Okay, so we knew the technology is coming. So we, like, even before ChatGPT came out, um, like, we, we just started a few weeks before like, this ChatGPT breakthrough. So we already heard all of the, all of the problems. Then we hear the hype and the, let's say, like, tech bubble. And we were able to pitch this enterprises and we quickly figured out, okay, like, this is a, this could be like, an interesting use case, like, just chatbot applications or retrieval agents, document processing. And then we, we are able to find out the problem and then ship this quickly to them. And then, obviously, like, this is like, the people factor of like, trusting. So we, we are telling them something about and then we're able to really ship something in production.

• 16:14Speaker 00

  But then there's also like, this product perspective where we have a lot of like, evals in place, right? So like, we, um, the, like, no company and no enterprise starts with fully autonomous negotiation agents from day one. No one does that. Why? Because they don't trust us and they don't trust the technology from day one. So we have like, a very easy approach of like, having a, like, human in the loop. And this is like, extremely helpful for us because we see, like, we have like, let's say, like, the perfect negotiation agent, which is like, overall, the perfect procurement negotiator. But we don't know exactly how a Fortune 10 enterprise, like, the specific Fortune 10 enterprise operates.

• 16:56Speaker 00

  And by having like, this human in the loop approach, we are feeding our agent with all the feedback and all the learnings. And then they suddenly trust us for like, 10K negotiations, 20K negotiations, 100K negotiations. And then you also have like, other, um, other agents that are like, more long-running. So when we talk about like, multimillion dollar negotiations where you analyze complex 3D models and technical drawings, there we on purpose have always experts in the loop, right? So there's an agent running for multiple hours and then we ask for feedback of the cost, like, cost engineer. And then it does the next, um, the next work and so on.

• 17:33Seema Amble?

  Maybe just to double click on that, how do you, what, where do you put the human in the loop on the, like, negotiation side? You mentioned the cost engineer, but like, if I were, you were going back and forth on a deal, is it mostly around the like, uh, you know, data for, you know, something like cost engineering? Or is there anything else where you have humans in the loop there?

• 17:52Speaker 00

  So again, depends on the level of negotiation, right? So like, we have to distinguish between like, Um negotiations where you just like a negotiation where in uh like where enterprises, they never did those because they didn't have the capacity. But by by by deploying agents, they can just um capture savings that they were unaware of. Right? So that they like before agents or before Leo, they just didn't care about everything which happened below fifty K. Right? So you can just like this is maybe a hack for like other startups. You can just send um an enterprise an invoice for forty K. They will probably not negotiate because they don't have the capacity to do so. Um um except they have Leo agents, then we are going to negotiate um against you. Um but other than that, they are just like just um just paying that.

• 18:37Speaker 00

  And obviously, there the risk of like, like, what is the risk of like you don't you didn't negotiate at all? So what is the risk now of having a bad negotiation agent? Nearly zero, right? So um maybe we miss out on some negotiations, but like it's better than nothing. Um but still, um in like in we're talking about business relationships and business relationships are not always about the cost and the money, right? So maybe you're not spending a lot on a vendor, um but maybe you you're like you need this business relationship, right? So a good example might be like podcasts or marketing services. Okay, that's like probably like a friction of of of the spend.

• 19:17Speaker 00

  But you don't want to like you have a clear business relationship with with someone like setting up the studio. Um and you don't want like like a random some random people doing that because they already know how how A six since he operates and how you want to record all this stuff. Um so there we have human in loop um approaches um where like procurement people care about the relationship, so they care about the voice of tone and and how it works. Um but mostly autonomously. Um and then we have the other set of agents where we always have um a human in the loop approach. And it's like um they're like multiple steps and like negotiating. It's also like it's not only the price, right? It's also like um how is the contract designed?

• 20:00Speaker 00

  So we're talking about legal, how is like um collection design? So we we talk about finance, obviously like cost structure. So we're talking about really like cost engineering. Then we talk about commercials, um that's procurement. Um and those are not back office people. Those are like highly trained um people where they have very specific knowledge of a very specific process, of a very specific company and very specific industry. Um and they feed are those long running agents of Leo with with those insights.

• 20:31Seema Amble?

  That's another example of how procurement historically may have been more in a box. And now it's like, okay, it's touching legal, it's touching finance. And it's touching a bunch of different software systems and people and both specialists and more generalists.

• 20:46Speaker 02

  Yeah, um can we can we map this on to a specific uh customer? Like not a specific customer, but a specific vertical. Like I'm a drone manufacturer. I manufacture humanoid robotics or something like that. Like, how many parts do I have to, you know, order and procure? How many factories am I touching? How many suppliers am I touching? Just all of all of those things. If you want to pick maybe Vlad a vertical that um, you know, is is just managing all of this complexity with Leo and and just kind of take us through what their experiences. I think that would really help just illustrate exactly um just everything that you touch.

• 21:29Speaker 00

  Again, like when we like when we as Leo, when we talk about procurement or purchases, like we don't talk about like laptops and pencils, okay? Like, we think like that's solved also by Leo agents, but that's easy. We solved this like three years ago. Um we talk about like when you like you want to build an aircraft or robots or drones or even like we now like doing a podcast about um again like AI hype, but even AI needs to be built, right? So you need data centers. Um and like building means procuring, like someone needs to like if you build an aircraft, you need to procure thousands of suppliers. You need to build a factory to build this airplane. Um and like really small frictions can have like a crazy impact, right?

• 22:12Speaker 00

  So like there's a like if you're running a very large project of like building a data center or building an aircraft, if there's like one specific part which arrives two weeks later, this can have like a damage of like hundreds of millions of dollars and postpone and postpone the project. Um so like all of and all of those that that's why like all of those decisions have to be coordinated. Um and one part is like you you need to figure out like what you need with what suppliers you work. Um what are like what is like the best supplier to um to like to to get this part? But then, once you decided all of the stuff, there's like all this operational um back office um stuff behind it, which might sound like unnecessary and boring.

• 22:55Speaker 00

  But again, like operational means someone sends a confirmation of like, hey, sorry, like this part is going to arrive two weeks later. And this is like one of five hundred emails um in the Outlook or Gmail um of a procurement manager. And if they miss this email, hundreds of millions of damage done. And this had like this happens regularly because the only thing they store in their um system record is then just a date. Right? So like it will like not like not this Wednesday, next Wednesday. That's what you see in a system. But you don't see like like maybe that's okay, but maybe that's a one hundred million dollar damage. And someone has to decide that. And that's that's also what our agents are doing, right? They they are not only retrieving the information, they are then making the decisions.

• 23:39Speaker 00

  Um are like, does this have an impact? What kind of impact? And how can we resolve this?

• 23:44Speaker 02

  Yeah, um when when you are sort of so deeply embedded in the physical world, um what what kinds of, you know, physical world problems can you intervene with? Like some things I would think are just like unsolvable, you know? Like, let's say you have a shipment coming in and a bunch of stuff like falls off the ship or the street is closed or whatever. Like they're they're they're kind of they're things that like you can't do. And obviously, there are things that you can do. So so where where can you intervene and where does that really make a difference?

• 24:22Speaker 00

  But but that's like, actually, it's about like probability, right? So obviously, like you can't like you can't change if um if like some like if there's like a damage on the ship, like every every example that you manage, like you um you you can't um change that. But you can, if you have like all of the context, you can predict that.

• 24:42Speaker 02

  Mm hmm.

• 24:42Speaker 00

  Um because you can predict like how um how reliable is the supplier? Okay, so they're like there are ways on like protect the goods that you're shipping. Um and if you have like all the context, you have like one supplier where like twenty percent of the goods are missing and then one percent of the good is missing. Um and maybe like this one with twenty percent is like ten X cheaper. But for this use case, it's fine for you to pay ten X the amount um because you have like a higher probability that this thing actually arrives. Um so and this is the powerful thing because you have like context not only within one enterprise. Um so we like we talked about like multiple stakeholders, but there's also context on the outside world, right?

• 25:25Speaker 00

  So just the agent should like have context of all the news out there. Um maybe even having like context of like some bad sale like on polymarket of like, okay, those disruptions are going to happen. Um then like um information about like on the supplier side, on the seller side, on the demand side. And by combining all of those contexts, um I wouldn't say that there is a limitation um in the long in the long run. Obviously, um like today, we have like different sets of like probability, but we can we can help throughout the process. And this is what we are building um building at Leo, right? So it's much bigger than just procurement inter a company. It's more like intra company and like how um how are like businesses like how enterprises are doing business with each other?

• 26:11Speaker 00

  So like buyer and supplier side.

• 26:12Speaker 02

  You described Leo as a as a multi-agent system. So can you describe what the different agents are doing?

• 26:21Speaker 00

  One level is that we um that Leo agents span across like all those four categories that that Cima mentioned um in her in her article. And um it again like depends on the um on the risk and the complexity. Um so we use like all of them, so like multiple agents. Um but the other thing is that like in in order to do a job end to end, those agents need to share information um with each other. They need to do this in like a very specific order. Um and when we talk about like a multi-agent system, this is essentially what we are doing. We are solving like the task end to end. And because the also like human level task involves eight people, eight stakeholders,

• 27:06Speaker 00

  um and maybe like three departments and five different um software tools, um we need to cover like all of those to do like the job end to end. And those agents need to then communicate with each other. And only with a multi-agent system you can do a job end to end. Um when we started out, we um like started off with like a retrieval, like more like a co-pilot, obviously, like three years ago. But then the next step was like a single agent. But then we very quickly discovered, okay, like that's like you can't solve and um you can't solve in like negotiation even um without having a contract um agent, without um maybe like having having an agent looking at the news and everything that I described before. Um so that's what yeah, that's what we define as a multi-agent system.

• 27:50Seema Amble?

  So if you have like a bolt, like an airline company needs to procure a bolt for, say, you know, Boeing needs to procure a bolt, what exactly is that process for procuring the bolt? And like where does Leo step in on that process?

• 28:05Speaker 00

  Yeah, so like this is like one of one of the um one of the purchases where that we can like run fully autonomously. And we can do this because of this multi-agent system. So like, first of all, someone has a demand, right? So they need to Somehow communicated. And even like this part is extremely complicated. So you need to call someone. Maybe you like, you like, you open up your laptop because you're like a construction worker. You open up your laptop only every second week. And now you are required to like work with SAP or any other like ERP system. So you can't even like issue the amount. So this is like how we make it very easy. So you check like you, you, you take a photo, like you, you, you upload a quote and Excel sheet.

• 28:49Speaker 00

  And it's actually everything that you, that you should know about procurement. Like no one cares outside of the procurement department about like categories, GL accounts, framework contracts, no one cares. We in the procurement will care, but no one outside that cares. And then our agents take off and they're like, they check the inventory. They find out, okay, they, they ask another plant, okay, can we, like, can we source those, those bolts internally? No, okay, then I'm calling the, I'm talking to the sourcing agent, finding out, do we have internal suppliers? Do we have external suppliers? Then some, like another agent has to draft the RFQ, send out the RFQ over, over email. Then a bunch of emails arrive. Some of them are like completely nonsense. Some of them are like just in the email, some of them are PDFs, some of them are Excel sheets.

• 29:34Speaker 00

  We retrieve those informations. Then we do like the next step. Oh, maybe like based on our price benchmarking, there's an opportunity to negotiate. And then, then we have like agents that essentially decide on the next step. So negotiation could mean strategic negotiation with a human loop. This could mean autonomous negotiation. This could mean auctions and e-auctions, calling then the specific agent, doing the negotiation. And then like doing this like end to end, think about like confirming the order, shipment tracking, invoices. And we, we are able to run this like fully autonomously, capture like all the context.

• 30:12Speaker 00

  And then obviously the next powerful thing is do this for like more complex parts where we talk about like direct procurement, where we, where we also operate.

• 30:25Speaker 02

  What, what's a direct procurement?

• 30:28Speaker 00

  Yeah, so like, essentially like everything I just described is, the main goal is here, automation, right? So you can like run this process fully autonomously. And then throughout the process, you can like generate even more savings, right? So it's not like, so we look at it's like, okay, what is like this, like end to end, like job to be done, how does it look like? So like, what are they doing like thousand times a day, but actually they want to do it like zero times a day? We've like run like fully autonomous agents. But there's like also opportunities of like, what are they doing zero times a day? But if a business would do this thousand times a day, that was, that would have a crazy P and L impact. Autonomous negotiations on spend, they never negotiated before.

• 31:10Speaker 00

  So, so this is, and this is like in the indirect procurement part, like think about MRO parts, building a factory, the bold example that we did. But also laptops and panels, marketing services, someone who needs to build up this podcast studio, those are like all indirect. And then we have like direct parts. This is like when you build an airplane, those like all these suppliers that actually, that you actually need to build the airplane or to build the drone or to build the robot. And then we don't talk about 50,000 suppliers, we talk about 100 suppliers or 2,000 suppliers maximum. And those are like extremely strategically important. And you have maybe on one supplier, like one billion of spend.

• 31:55Speaker 00

  So you don't want to run an autonomous negotiation. You want to run a negotiation which, which takes three months and where you're like crazy prepared and where you have engineers on your team analyzing, okay, what's the indices for aluminium? What's the indices for oil? How did the price change? So you like really take over like all of those drawings, you check the quality of this part. And this is what like, where it gets like really exciting deploying, deploying agents.

• 32:22Speaker 02

  Yeah, and, and for something like that, presumably like you'd have the expert engineers and the other procurement people kind of more as the front of house and like the agent is more back of house. Is that the idea? Or is the agent like, actually, it's like you, you sit across the table and you're shaking hands and it's like the robot instead of the human who's like negotiating like, is it, so yeah, is it, is it more back of house or is it like still front of house?

• 32:49Speaker 00

  It's, it's, it's obviously more back of house. Because you like, because you like, like these complex multimillion dollar negotiations. And, and that's, that's again like a beautiful example. 90% of the work is preparation. The end result that you see in your system of record is like, oh, instead of like one billion, I paid 900 million dollars. Okay, but like, there's like three months of preparation and like 10 people working full time on that. And obviously this is like happening in the, in the back. But actually we have some use cases where it's also like, like helping in real time. So think about, let's assume we would now have a negotiation and I have like a perfect preparation, the same as like with like those, those, those notes, those notes that we are having here.

• 33:36Speaker 00

  Imagine like while we are negotiating, I would have like real time insights on my screen popping up, where you tell me the indices for oil changed 10%. So like it's increased by 10%. So that's why we need to increase the prices by 10%. And I would have like an initial, like an immediate pop up of like, that's true, like oil increased by 10%, but the product has only 30% of oil contained. So like you, you can, you could, you shouldn't increase the price by 10%, but maybe only by 4%.

• 34:08Speaker 00

  So yeah, there are like exciting use cases also like in, in the real life part.

• 34:13Speaker 02

  Yeah, we, we know that, you know, companies like Harvey and Dakagon are really fine tuning models now.

• 34:21Speaker 02

  What, what kind of underlying models do you use and, and how do you approach things like fine tuning or, or harnessing?

• 34:28Speaker 00

  Yeah, so like, we believe like you can, so like we use multiple models from like all, all providers and we really see this as a, like obviously like as a commodity, right? So they are like, have really good like general business purpose. So like reading, creating a PDF and like creating, creating Excel sheet, like all, all of this stuff. But we also believe that like for some use cases, you, you, you can get extremely far with like combining the foundation model with a, with a harness. And you can maybe reach like 100% of like the job to be done. But there are also some use cases where you can have like the best foundation model, the best harness, whatever that means, but like the, and the best harness, but you still

• 35:13Speaker 00

  can get only to 80%. And like, it would say like, good examples for that is like, for example, what, when we talk about negotiations, what like cost engineers doing, right? So they're like analyzing drawings and then they like defining, okay, what should this,

• 35:34Speaker 00

  this part actually cost? Like, and that's why it's called like should cost modeling. And there's, there is definitely like an opportunity where we like thinking and already started, fine tuning the model to get then to, to get then to 100% in this part. Another example is like price benchmarking, where think about like the, the, like you would have a quote and in a perfect world, you would just drag, drag and drop the quote somewhere and you would get the perfect price. But it's like, and all those, like all those informations, they are like not publicly available, right? So there's all like, those are like all preparatory data based on like one, one enterprise or like across multiple enterprises.

• 36:19Speaker 00

  So like general purpose models can't train their models on that.

• 36:24Speaker 00

  So what we are, like what we are thinking about is like maybe like not training just an LLM, but I think like what we see now with models like Jeff or so popping up where you have like an, you train it on like text data, but the out, like the, the output is actually like an outcome or just like the perfect price. And you can't do this with harness because like, if you would give me like an, a quote from BCG and a quote from McKinsey, they, they could do like the exact same work, but this could be like a 10x different price. And I would have like no idea like what is better. But if you give this to a procurement manager, he would like initially have a gut feeling. Or like, okay, like this quote makes sense.

• 37:09Speaker 00

  Or like, I think, I think like good examples again, like content creation, like always like, I don't know like how much I should pay someone for creating a video, but there's like a gut feeling behind it. If I asked like another video creator of how to, how to do that. But if you ask them to write down the rules, they can't do this because it's like, just like gut feeling and instinct. So, and that's where we see a lot of opportunity actually, training an agent, but not maybe like a classic LLM, but more like, again, like what, what companies like, or like models like, like Jeff are now doing on the outcome based. So we have like price benchmarking, should cost modeling.

• 37:44Speaker 02

  Seema, we've talked a little bit about how labs are really moving into industry-specific work or working with incumbents to do this. When you think about what a durable vertical AI company looks like, what are the qualities that you look for?

• 38:02Seema Amble?

  One is around, you know, owning the end-to-end work that we're talking about, building up this data asset and being able to do something that hasn't been done for many cases. This is all said, I think when you talk a lot about moats, it's really, really hard to forecast your moat going forward. If you look back at all the best businesses, at the early stages, they were just thinking about, okay, I'm winning customer trust, I'm selling more to them, and there's a lot of opportunity versus, okay, I'm going to do these six steps and then get to the seventh step and then we'll have a moat. And so I think we talk a lot about defensibility and durability, and I think part of that is you're locking

• 38:45Seema Amble?

  in the customer, there's more dependencies, they find it valuable, and you're doing more of the work. And here, it's truly like, okay, you know, old CRM company was a log for all the deals, new sales AI agent is actually owning a lot of the sales prep process and the outbound process, fielding inbound and doing a bunch of the work. The overall customer is dependent on that product, and that's like a really important signal of getting to the moat. And everything we talk about in terms of stickiness and network effects and all that is sort of downstream of that initial like customer use and the value of the product.

• 39:30Speaker 02

  Vlad, have you had conversations with customers or potential customers who ask you, you know, why should I buy your product? Why can't I just, you know, plug into a model and do this myself? Or like use whatever existing system of record I have plus a model? Like, what do you tell them and how do you convince them to use LEO?

• 39:54Speaker 00

  Yeah, 100%. And that's a very fair question, right? And like, even if you look like internally at LEO, so like the first use case three years ago, which like kind of like went viral in the procurement world was like, just like having a quote and then getting this information into SAP. So like very, like, again, like technically like, but like tremendous business value. And so you have like this retrieval agent getting like all of the information, putting it into SAP. We like, this was our like first product. And we had an engineering team, like obviously small, just like the three of us, or maybe like four people building it and then selling it. But this is nowadays, a case study, if you're applying to work

• 40:38Speaker 00

  at LEO. So we give this to people to like to build this and they have like eight hours to do so. So what I want to say about that is like a product that we, like one of our first use cases, can now be somehow built by engineers within eight hours. So because it's very easy to build stuff nowadays. So obviously there's a question, okay, well, so someone can build this within eight hours. Okay, cool. But then couldn't like just procurement department also just build everything in two months? And the answer is like, yes, you can build this in eight hours and you can build this, but you will only reach 70%, let's say like, of the performance. And the problem is 70% of performance or currency or however

• 41:23Speaker 00

  you measure it, it depends really on the task, doesn't mean 70% automation, right? So this can be mean that you're like, have 70% of the performance, but you still need to do 100% of the work. Because 70% is not that much. So again, like all of the people have to have to check the data. So maybe you even created like more work, more work than than you did before.

• 41:51Seema Amble?

  Two things to layer on to what Vlad just said. One, overall, it's good if there's more adoption of the base models or just like GPT products, because it means that people are also going to trust vertical-specific products as well. So I think that increased familiarity, comfort, excitement about AI tools is generally just good for the market. The second thing is, I was chatting with the management team of a Fortune 500 company two or three weeks ago, and one of the things they mentioned was they had tried to build out their own cash collection product. This is a big enterprise business. And they, like

• 42:34Seema Amble?

  after, I don't know, three or four months of work at a minimum, they had found that there wasn't enough context, there was poor quality context. They had a lot of recordings, a lot of screen grabs, they tried to pull it all into one system, but they, there was no, there wasn't good enough. And then there was this giant question around, okay, like, you know, we've got now two different ERPs, and we were about to acquire another one. Who's going to update all the mappings, test out, okay, does it work? And then like, we keep talking about exception handling, you now need to map that onto a totally different system, a different way of doing things. And I think they quickly realized that the internal build didn't make sense. And so we keep hearing stories of this where people

• 43:18Seema Amble?

  are like, okay, I'm going to do the internal build. And they're like, wait a second, it's not different from what DIY has ever been in the past, which corporates have always tried. But I think enterprise companies generally realize that like, there's their core competency, and then there's building internal tools, and they should focus on the first camp.

• 43:37Speaker 02

  Yeah, yeah, makes sense.

• 43:39Speaker 00

  Yeah. So that's exactly what I've meant with like, obviously like they can get to 80%, but those last 20% really matter. And you can only get, like, they matter to get into production. So that's why you need like this harness, right? So you need all those like integrations, memory, work flows. And sometimes you need vertical data to do that. And like the Pareto, Pareto principle, so like those 20% can make like 80% of the effort. Or like, they are making 80% of the effort. So like to all like those Fortune 500, 150 companies, so you can do this. But then, let's say like procurement workforce of AI procurement

• 44:23Speaker 00

  agents should then become like one of your core competencies. And you should evaluate whether this makes sense or not for you to have this in your cost scale.

• 44:34Speaker 02

  Vlad, I'm curious if you're seeing suppliers start to use agents or AI at all and kind of like, what happens when both the buyers and the suppliers are fully AI-enabled?

• 44:49Speaker 00

  Yeah. So like, we 100% believe that like, in the future, there will be like agents on both sides. And this makes so much sense. But like surprisingly, what we see is like, so like when we look at the supplier side, this also kind of like equals the seller side, right? So what we see is like that the the sales side was like always ahead of the procurement side. But what we're now seeing with those suppliers for those Fortune 500 companies, this actually is not true. So they are like maybe advanced in like, let's say, video recordings and like using tools like granola and all this stuff, but like not like really having agents deploy, they're automating the work. And the cool thing is,

• 45:33Speaker 00

  as like procurement is like, is unsexy, right? So sales are sexy, procurement is unsexy, but it's like one process, and procurement is the counterpart. But now the good thing is, in those industrial companies, Fortune 5000 companies, procurement has the bigger power to the supplier. Because you as an like typical use case, like automotive supplier, you dictate to your suppliers what they should use, what the quality has to be, how they have to answer to a specific RFQ. So the opportunity is now, if we serve the procurement department and then they can dictate what a supplier should use, why are we not like,

• 46:17Speaker 00

  why aren't we like also pushing them to like LEO agents that are also helping them automate the work, owning then both sides of the transaction?

• 46:27Seema Amble?

  And Elaine, I know we were talking earlier today about, you know, how can you have two parties on the same platform? How does that work? I think it could even extend into like legal work, right? And these are two very adversarial parties, right? But like, if you have two law firms with clients with different interests, but both benefit from knowing, okay, here's the latest draft, here's where we are with the open issues, here are things that have been agreed upon, and just even tracking that, that doesn't really exist right now, right? That's all being human, that's being created by humans. And so that coordination effort, an agent could be doing.

• 47:04Speaker 02

  Yeah. Yeah. It's, well, it's super cool to think about how like, you know, both sides are kind of maybe evolving in tandem. One side might be going a little bit faster, as you're talking about Vlad. But over time, potentially, people are just like on the same platform. And actually, it's like, it's way better coordinated for everyone.

• 47:25Speaker 00

  100%. Because like, also like what we mentioned in the beginning, right? So like, Like the obvious, like the obvious question is like, okay, but like we, we also talked a lot about like negotiation agents. What if like both parties have negotiation agents? And price is a hundred percent a point where it's like a zero something or like they have like different interests. But we, but we also discuss is that like price is like the outcome of like 5,000 different, like other tasks that happened. And on those 5,000 other tasks, they have the same incentive. Sales wants to have as less, like little friction as possible. Buyers want to have a really fast time to market, right? Again, like coming back to

• 48:09Speaker 00

  building aircrafts, building data centers. You want this data center to be built as fast as possible. You don't want to be built at like six months later, just because it takes so much time to analyze all of the responses from suppliers. And you want to make the sale also fast. So like all of those 5,000 other tasks, the incentive is exactly the same. And that's why you can deploy, or Leo can deploy agents also on both sides, doing like automating the work of all those other tasks. This is the beautiful thing.

• 48:41Speaker 02

  I think that there used to be this logic of like, you shouldn't customize your software too much to, to one end user, a one end customer. But I think something about L L Ms and AI in general is like, it might increasingly be possible to customize without slowing yourself down as a business too much. So I'm curious if, if that is something that you're seeing, Vlad or Seema, and, and kind of what, what does that mean for, for end buyers of software?

• 49:15Seema Amble?

  I think the overall principle, there's a lot of forward deployed work happening right now. And part of that is because the state of the customer data and understanding customer N is a lot harder than understanding N plus one. And so we're sort of in the early stages of deployment overall. And that's why there's still a lot of humans as part of this product. And by the way, that is something that is harder for the incumbents to do because they're also like not set up in, in a way to have even their, the way their product feedback cycle works where they have a, they have implementation teams. But that's very much an afterthought versus something that feeds into the product side. The beauty of AI is A, it learns over time. So that's why we're talking about learning loops. And you

• 49:58Seema Amble?

  have the right eval process. You can do more and more complicated jobs over time. And part of that is automating the deployment itself. And so you can, and I'd be curious to hear how Vlad is doing it, but a lot of our companies even are doing that at, you know, at a rapid clip where more of the customization, A, is being handled in an automated way, and B, the customer is able to turn the knobs and levers around customization via software versus, okay, I needed to bring in, A, the original, it was like, you know, you brought in Accenture to do your SAP customization. And now it's like, okay, I've got a forward deployed team that's going to build some, you know, help build and spec out. And ultimately,

• 50:43Seema Amble?

  it's going to be completely software driven.

• 50:45Speaker 00

  Yeah. I mean, that's the reason why, like, when you look at the org structure of Leo, like 85% of the people are engineers. And even if you look at like the people where you like, they don't have an engineering title, they have like mostly like an engineering background. And the reason for that is, we obviously, like, we don't want to be a consulting company, right? So we make sure that we have like overall the best agents in indirect and direct and finance in those parts. But then like, as you mentioned, like, there is a lot of like forward deployed work to do if you go to enterprises, because they have like different nuances in their, in their processes. But how we work is, we, as you mentioned,

• 51:27Speaker 00

  like building a product in a way where it's like, where we're reducing this customization, but also where it's a lot of like self-service. So the job of our like F D Es and forward deployed engineers is on the one hand, making it like self-service, but like internally is like automating their own job, right? So like their KPIs, like you're seeing this happening like multiple times, like you're, like you're, literally your job is like to automate yourself. And then if you automate yourself, you go to the next task. I think it's like also like an approach that Google was doing. So like, but yeah, a hundred percent agree with you. And that's exactly what we are, what we are building, how we're doing it.

• 52:09Speaker 02

  There was recently a very big system of record event. And, and we're not, we're not talking about Dreamforce. We're talking about the, the Bots and Buyers Summit that, that Leo hosted in, in New York. And, and just wanted to, to kind of hear, you know, stories from the ground and, and what, what you're seeing among buyers. What are people excited about? What are people looking ahead toward? What are people, you know, asking you for? Can you just kind of tell us some stories from, from that day and that event?

• 52:43Speaker 00

  So, I mean, this was the third time that we are doing, that we are doing this event. Now we did it in, in New York, just a few blocks from our, from our office here. And over 100 procurement leaders arrived, right? And what we, like, we made sure that we, like, when we do, when we do such events, that we only invite like high caliber people, right? So like C level, C P O, Vice President. And there are like two things very, like very different of, of how we do this, or like why they're so amazed. So the first thing is, when we look at like how procurement used to work in the last 26, 27 years, a lot of tools emerged, right? So you can like, they're like all those technology landscapes, and you can find them on LinkedIn. You will see like, there are like 500 procurement tools.

• 53:28Speaker 00

  But like, if you like, and this is also the reason, like, but when you talk to procurement people, like, very painful. And like, I challenge someone like to find someone who like loves to work with procurement, like no one does. Like you can really state, people hate working with procurement. And when talking about the requesters, and I'm talking about the suppliers, and even people in procurement hate working with procurement. So like, what's going on if there are like 1,000 tools? And the reason for that is like the, all of the tools, they just made the, the process more efficient. That's, that's all. But they never changed how those, how those people actually work. And it's like crazy to see that they

• 54:13Speaker 00

  work in emails and on Microsoft Teams, and in Excel sheets and PowerPoints. It's like their main channel where they work on. And it's like zero, like zero AI enabled, obviously, in, in this part. And the, the other thing is like, we, we give them like a very cross department perspective on procurement, right? So we are not talking about like, look at this crazy invoice feature that we developed, but we're more looking like, someone needs something. And in the end, you have it on your table. And this can be like across indirect, direct, logistics, finance. And you can see how we are doing it here. And we're also putting it into like more into like a physical world, because like AI agents, that's very

• 54:55Speaker 00

  abstract. So what we are doing is we building our booths, and we even have this in our offices, in, also in New York, where you can walk through the booths and experience like, all of those agents like really hands-on. And that's what, what's, what the people love. And it's like, I mean, the next event will be with around 700 people. So you can imagine how, how crazy this, this grows.

• 55:22Seema Amble?

  Procurement people gone wild. Yeah, yeah. It's gonna be, that one is in Munich?

• 55:29Speaker 00

  Yeah. So we're doing them like in, in Europe, Munich, and in New York, all the time. Yep.

• 55:36Seema Amble?

  Cool. If you're listening and in procurement, you know where to go. Maybe, I guess, one question, one question for me. What do you think it takes to get, to, to get people excited about procurement? Is it, is it the agents? Is it the people? Is it the time saved? Like what, or like something else? Like, you mentioned it, like procurement is one of these things that I, I remember, you know, people aren't, they don't like, like it's like a universally kind of disliked low N P S area. I remember talking to a guy who was at a procurement, like, I don't know, seven or eight years ago, as I was looking at this category. And he was like, oh, I hate talking about this product. Like I, I use, you know, this legacy system of record. I'm on Coupa and I don't, I don't want to

• 56:19Seema Amble?

  buy anything else. I don't want to talk about it. It's fine. It was like the most disgruntled customer call I've ever done out of like millions of them. But I'm, I'm curious, yeah, what it is that you think, you know, really gets people excited about this category?

• 56:33Speaker 00

  Yeah. So, I mean, it's like, it's like, and this is also why, why I like procurement is like, it's on one hand, like, so like, the reason like why we started in procurement are like, it's not essentially like what happened, but like how people react to it, right? So if you talk to the, to the people, they're like really frustrated. So this means it's a like highly emotional topic. But it's like, like, let's be honest, like, B to B Zs, okay? But it's like highly emotional. So that's a good thing. And then when, if you combine this with like something which is boring and niche, this is also an advantage because, again, like the, it's, it's like also easier, easy for us to amaze those people,

• 57:14Speaker 00

  right? Because like the really last revolution they have seen is like 20 years ago. And then maybe a nicer user interface 10 years ago, but nothing else happened. And so you have like boring, highly emotional, and then plus crazy business impact, right? So like, it feels like it's unnecessary, but like I told you, like some examples. So like, it has like obviously crazy P and L impact, but it has impact on like the whole economy, right? So we're like, we are talking about like how data centers are built, how aircrafts are built, how cars are built, how drones are built. So it's extremely important that

• 57:56Speaker 00

  you have a fixed procurement process, not only to like make it happen and build something, but then also when you talk about, like when we, when you look at like the competitive landscape. So to get like 1% margin increase, you need to make 10% more revenue, 10% more sales. So like, if you just manage to get like 1% savings, it's like equals like 10% of sales that you have to do to get the same outcome in your P and L. So like, it's tremendously important. And like, you combine all of those three things, and then you have like a trillion dollar business opportunity. That's, that's my opinion. Like for procurement, but there are probably also like other things that like emotional, boring, and have a crazy business impact.

• 58:40Speaker 02

  Well, Vlad, thank you so much for joining us. This was a ton of fun.

• 58:47Erik Torenberg?

  Thanks for listening to this episode of the A16Z Podcast. If you liked this episode, be sure to like, comment, subscribe, leave us a rating or review, and share it with your friends and family. For more episodes, go to YouTube, Apple Podcasts, and Spotify. Follow us on X at A16Z and subscribe to our Substack at A16Z.substack.com. Thanks again for listening, and I'll see you in the next episode.

• 59:12Erik Torenberg?

  As a reminder, the content here is for informational purposes only, should not be taken as legal, business, tax, or investment advice, or be used to evaluate any investment or security, and is not directed at any investors or potential investors in any A16Z fund. Please note that A16Z and its affiliates may also maintain investments in the companies discussed in this podcast. For more details, including a link to our investments, please see A16Z.com/disclosures.