Latest / The Chief Customer Officer Podcast / EP035: AI Leverage, the 24/7 Mindset, and the Value Era w/ Ziv Peled

Transcript

EP035: AI Leverage, the 24/7 Mindset, and the Value Era w/ Ziv Peled · The Chief Customer Officer Podcast

• speaker-1: Welcome welcome back everybody to another episode of the Chief Customer Officer Podcast. And today I have a I think it's maybe call you a former Chief Customer Officer, because that's not your your title anymore, right? I got my my good friend Ziv Pellet with me and We're just gonna catch up and we haven't caught up in a while and we're gonna chat about what's going on in each other's world and and figure out what it we're what we're both working on, what's working, what's not, what stories are you following in the industry and and share that with everybody. So Ziv, good to see you, man. Thanks for taking the time to chat. speaker-0: Good to see you too and thank you for having me. speaker-1: have things calmed down over in Israel right now, or you still got speaker-0: No for now yes, but i in general it's it was almost all the time chill, so like was very good. and again like I think that some things that I see over the world is that thing things are just exaggerated. I saw I was just this weekend I was off I was in Athens in Greece and I saw that there's huge rainfalls in in Thailand. So I was I was calling my friend just to see that it's okay and he shoot me a picture that not there was nothing there. Obviously there was some in some areas, but it was very exaggerated to to say that it's all over Thailand, all over Bangkok or whatever. Same speaker-1: Yeah. speaker-0: here for for Israel with the news and everything is fine. And I think that you know if before we go customer success if you take it tech wise that's one of the best years for Israel the exits, the investments, the we we can speak later. I was almost five weeks in the US back in July in August and I envy the the AI gold rush we can speak about that but We are doing significant things here in in many areas. a lot by the way in defense and and cyber. speaker-1: Yeah. speaker-0: One of the biggest outcomes of Israel, not today, but a couple of years ago, the Wiz acquisition billion speaker-1: Free. speaker-0: dollars to Google, the biggest acquisition by Google ever. And that's getting accelerated. I I don't I don't remember the numbers by heart, but like this year. one of the highest investment in Israel all over and we s we see that we and I think we can speak about AI also later. What does AI do for us? But I think that we are in a a amazing time for Israel and the technology in general. speaker-1: Yeah, and and the culture, the Israeli culture, I think has bred the ability to succeed despite, you know, challenges like the fact that you know most everybody serves, right? In the military. You have great training. You've got defense technology being developed at a incredible pace. So, I think it's a really it's a really interesting point that you're making. And and so you're you're benefiting Just as much as the Silicon Valley folks are benefiting from all this innovation. And I think, you know, the innovation is actually spread well beyond Silicon Valley with AI. If you're really paying attention, anybody can succeed now. where y y you might have had to be in Silicon Valley to do certain things in the past. I don't necessarily think that's true anymore. And I guess Tel Aviv is proof of that. speaker-0: Yeah. I I think I think we we we have our advantages, but I think that I I felt something really different. I was not in the US in the internet era of two thousand, two thousand and one and I w or ninety nine, it started probably ninety nine. And then I wasn't by the way, I was born by then. But like speaker-1: I was speaker-0: but and I wasn't in the US in the then after I don't know when was the l the next jump, probably somewhere twenty ten. speaker-1: To the financial crisis. Yeah. speaker-0: Yeah. And I I can I can tell you that I feel that there's some ingredients to Silicon Valley that they're not in the in Tel Aviv. They're not they're not in Israel. And then I think that it's for some companies, I think that this is what makes the a good great connection of either a company that grows over the the two hubs or sometimes acquisition or mergers, etc. And sometimes it's super successful. I think that I I mentioned the Wheezy case, but we can speak about so many different cases. I think that it's significant to the success of the company. In one point, you know, we can speak also about Upstar, super successful story of a small company from the next t next to Tel Aviv that became the leading in a market that and still the opportunities. So huge. So we we succeeded to the our achievement so far. Over 500 million in the RR. And just now the investment deal. that again, I I feel like I speak to people the when when I meet them. Some knows that we had an investment deal, but they don't know the details. Some knows it some some doesn't even know that we had an investment deal. But in in real speaker-1: When you say investment deal though, you're you're saying there's a there's been some additional investment in Apps Flyer? speaker-0: Yeah. Bottom line, there's a a strategic investment by four players. it's not a private equity, it's not a V C, it's four players. That's what's interesting in in them is that they are from the market. So it's Meta, it's Google, it's Molocoin Unity. three of these companies are are public and they invested significant amount in AppStar. But Well, that's less important for for this talk, but it's a it's an amazing opportunity for the importance of measurement and a AI in marketing in general. speaker-1: Absolutely. and you've got a so last time we talked, you were chief customer officer and you were transitioning into chief AI officer at this half a billion dollar revenue company. That's not valuation, that's a revenue. So just to be clear. how's that going? speaker-0: That that's probably the most interesting part of my career since I was VP product and chief customer officer in twenty seventeen until end of twenty nineteen. but and I'm still chief customer officer. but I'm focusing on more of the strategy and engaging with mainly enterprise customers in the regions. But for the main role, which I assume most of the chief operating officer roles in the company. it's it's it's probably the the most interesting thing I did I did in my career in one of the most you know I I don't remember an era that we had news every week Monday, Tuesday Wednesday Sunday even even until speaker-1: Multiple times a day. Yeah. Yep. speaker-0: We had a new model last week, Opus five point five. We had a new model yesterday. And that's without even speaking. Like I got to the level that I'm now going to the office with two laptops. I have one private, one for work, because not all the features are enabled for in the corporate laptop. So I have all the features enabled on the my personal computer. And then that's even before speaking about type safe. instinct use we can speak about so many different things but like yeah I think that today you you said something before that caught my hair saying that now everyone can succeed now and and I I fully agree with that it's like it's up up to the individual now to decide both professionally and personally how much of their time they invest in order to succeed A big portion of that is how much you invest in AI and to learn mainly to train yourself, to to use, to to get to know the new things as as fast as possible and to take it towards your career. speaker-1: Or yeah, it and I don't think it's it's not l learning what's happening from a news perspective in AI. It's actually hands on. Right? It's hands on. Even if you're chief this or chief that, chief customer officer, chief revenue officer, my my feeling chief executive officer, my feeling is you actually need to be hands on with this stuff yourself, no matter what. You need to be a builder. Everybody should be a builder right now. speaker-0: Yeah, 100%. I think that the builder is a big word because then people think that they need to write software. Yeah it's not. It's a workflow, it's a skill, it's automation. By the way, I you know I I take a lot of vision, a lot of inspiration from so many different things. The last thing I took this weekend was from the CEO, I don't remember I don't remember his name, of TypeSafe of Jev. And then E Very clear, like this 18 minutes video that I saw of him two months ago in early July. He spoke about the world today of AI's split to two things assistance and automation. And then most of the tools that are shipping value today are the assistants, the copilots, and there is AI in automation, but there's no success. And that's why he built and released TypeSafe, which enables to almost succeed in automation in everything you do. And I I think that you know, even to connect to what you said now about the chief, I think it a chief needs to be able to take this kind of a new model, run it, use it for some use cases, and understand the technology, understand why. your team needs or doesn't need it tomorrow. And to be able to very to ship it very fast to the team, either push the team to do it or get the team up and running with with that I think it it's it's so crucial because by the way the the as we said before that the these technologies launch every day with speaker-1: Yeah. speaker-0: it as something new and I feel like even B2C wise, I had a discussion with a few friends. we said I think three or four couples of friends of mine a week ago and we spoke about instinct. I converted I think 50% of them to use instinct. And they use instinct in the last week. And then I think one of the biggest things in this era, doesn't matter professionally or personally, is that FOMO plus adoption is at rates that we couldn't imagine. I see adoption of eighty to ninety percent of things that was used to be ten to fifteen percent. So it's like it's a it's the it's the total other side of of things and it's it's tremendous. And and the amount of value that you get from these, like I don't know if you use instinct or muse or speaker-1: Do town and yeah. speaker-0: It's crazy. By the way, I I used town before and I think that like the the instinct is a jump to you need to work less, even even less. speaker-1: Yeah. And for those who don't know, Instinct Town Muse, which came from Meadow, which was released, what, last week? these are personal assistants. And so what I found interesting about them and where they inspired me is that you know they interact with you first in the place where you live the most. So email, Slack, Teams, your iMessage on your on your iPhone or your your messages app on your Android. That's where Instinct That's its its first place to interact with you. I get a text from Instinct every day that's, you know, asking me about, you know, my schedule and whether I need to book travel and you know book a hotel room and that kind of thing. And so that's all coming. So when you say assistance, there are personal assistants. And then there are, you know, what we're we're trying to figure out and what we're doing more and more of is trying to figure out, okay, those are personal assistants for us individually. And then I've got agents like Claude and open AI that are are almost like my personal assistant at work, right? I can set up cowork tasks to go do things. But as soon as my account is gone, that task is gone, right? It has to be rebuilt by somebody else unless I've left it behind for somebody. But what we're trying to do is is figure out, okay, what's the right, what is the right tool for multiplayer AI across a team of people? So you've got a team, I assume you still manage the customer success team at Ab Slaughter, which is what how big now? 200, 300? speaker-0: it's one hundred seventy. speaker-1: One hundred seventy, okay, two hundred. So the question is, you know, like how do you build agents that serve that team and scale that team and take busy work off that team in a standard way so that they can spend more time doing what it is that they need to be doing with customers interacting. speaker-0: I I think the most significant thing is that that we've done and it's a journey because we started the AI journey very, very early. we used OpenAI until February of this year. I I decided to move to Claude June of last year. And then I was confident enough to move the entire company to Clode by early this year, so between January and February. But the the the thing and a a lot of companies did that. we we invested significantly in building tools, internal tools, that enabled mostly go to market. So engineering product they've a lot of things that they've built, I focus on mainly go to market, but it's all also serving product on the outbound side is that We took very advanced developers and architects. They looked at how the business is being managed, how we execute, and they took and built proprietary tools over cloud that includes an MCP, that includes a semantic layer, and includes a data lake. and very fast they they found ways of how to use Some of the things that we had in API or MCP, how to use it even better. For example, I always give that as an example. The Gong transcripts that we got via API. we do it now much, much, much better in processing it ourselves. speaker-1: Yeah, we didn't. speaker-0: very, very, very cheap model, amazing outcomes. And over the last year, I think that now we have 29. Different data sets of different things, including email, AI chatbot, Zendesk, Gong, even third parties. And then now when we now prep for a call or build an opportunity or strategize this renewal, then we do it with this tool and and it and it's it's amazing. The next step that we are doing right now, that we're working on right now, is how to build a copilot out of this so that Out of the individuals, the around 800 people that works with that technology, they don't need to build their own second brain and their own agents, they will get the agents inherently from a central place. So it's kind of like our other companies much more inspiring than what we do, like ramp and what they've done with glass and in other tools that they have for agents. and I I I feel we are doing a very good job. I feel that we can do it we can execute better, we can do it faster. I think we are not we are not yet excellent. I speaker-1: When you say speaker-0: have a lot a lot of internal critique to myself and my delivery. speaker-1: Well, it's easy. It's easy. Like everybody's an internal critic right now because you read so much sensational stuff, you know, wherever you do your reading, right? About what this person or that person or this company or that company is doing. But I think, you know, y you mentioned ramp. My my my sense on ramp is that they are very focused, their product and engineering team is very focused on agents for that kind of workflow. So the fact that they are taking support cases and running that through agents that are automatically, you know, pinpointing the solution to a broke fix issue and creating a pull request, writing the code to fix it, giving it to a human, maybe to do the final review, but maybe for some class of them, just like running straight through. I think the go-to-market side, there there's, there's a lot of complexity to that because there's a lot of fuzziness and a lot of go-to-market flows. But I wanted to To go back to something. So and then and then I want to tell you about a conversation I had with CTO at a bank last week, which is very interesting. And we're gonna have him on one of our podcasts or live sessions soon. But what how much time you've got this AI chief AI officer role, and then you've got the CCO role. What is your time split between the two of those? speaker-0: Yeah. I I think b because I did that in the past also, where I invested majority of my time, around eighty percent of my time in product between twenty seventeen and twenty twenty. so I I know how to do eighty percent. speaker-1: Eighty percent? Okay. speaker-0: So I I always like my my team I can I can invest on the CCO side in strategy. obviously I need to be up to date with all the latest things around our product and our and our go to market but in in the end it's I'm not managing customers and I'm not managing teams directly so I have a global VP of customer success and she runs everything. And the leadership I I spoke about it in many other cases in many other podcasts. majority of my leadership in customer success in App Slier. They are between seven to ten years in App Slier. The global VP, she's eleven years in App Slier. So they have all the, you know, the organizational memory and the context and the DNA and and the culture. and and and that's super helpful in order to even when they are hiring now new people b whether it's ICs or leaders, it's really easy for them to get people out by the first to get the right people in and then to get them up and running. I work with CSMs, even new CSMs in different regions and the they they are so ready, so plug and play, so mature for the job. And then I see in many of them also the advantage of AI because like when they use things like the the tools, the internal tools that I spoke about before, it enables them one to have amazing strategy on how to work with the customer. Second, it enables them to have more time. So I won't say that they have more time for every customer, they still need to prioritize, and obviously we are pushing them pushing their boundaries and more and more and more to more and more accounts to be able to manage more accounts. But the accounts that they are managing and are prioritizing, they for sure a hundred percent they have the right strategy and the right time to invest in them. And I I think that this is the significant thing in in AI is that I don't think that we are even halfway into this technology and breaking of this what's going to happen. And I think that we we don't even understand, we don't even cope that in the future we will have six, seven free hours in the day to invest in the right things. We will need to strategize priority with the tool. We need to strategize I I believe that managers will be able to manage twelve people very easily. But by the way, I always say it better than you are managing today seven. Because you know everything they can do and you know their pros and cons and you know what they need to in improve. and and I I take it a little bit now to a lot of my AI strategy, but but I think in my AI vision, one of the things that I'm almost sure of And it's people call it loop, people call it loop engineering or now loop everything. And I think that I call it a little bit different, but it's the same thing. I believe that many people doesn't matter if you work nine to six or nine to nine, I think that we are still in a mindset that at some point we drop the pen and we continue tomorrow. And that's We are not we are in an in a new era. In the new era, we need to change our mindset to work 24-7. Doesn't mean that you work 24-7. It means that you need to split your work to what you need to do and what AI needs to do. And you need to make sure that AI has enough work to work continuously 24 7. And then it generates new work for you that you need now you need to evaluate. three four times a day that work that I had a very interesting conversation with a friend of mine is a CTO co-founder of different startups and he said like what we are now looking for for people to do is that it doesn't I I don't need you to work at nine pm but if you need to eval for five minutes something that ran now for four hours and not to wait for the evil to be done next tomorrow at nine a.m So just do that five minutes. You will have more time to be with your kids. You will have more that you can maybe come to work later, maybe you can w go off work earlier, whatever. But just be there where where you need to be where are you where the person where the human in the loop is critical, be there and make sure that AI can run. And I connect it all also to the AI budget to say like who said that we need million or four million or six million dollars a year. Maybe we need twenty million dollars a year, LLM, but that can generate hundred million dollars in new in new booking. I it's it's a matter of that equation and I think that a lot of people they are really stuck in the equation of how do I make the team more efficient? How can I run linear? And they don't look at the upside. speaker-1: That's right. How can I do the same thing that I've been doing for the past five, ten, fifteen years and automate some of those tasks? But I love the I love the the mindset shift from nine to five to twenty-four by seven. Like I mean absolutely true, but I think the only people achieving that right now are engineers, are true engineers that literally have multiple agents running. on multiple machines throughout the day, mul multiple, you know, with minimal touch points per day. I I have you I'm sure you've listened to DH H maybe over the past couple of months talk about AI. speaker-0: Yeah, a couple of times. speaker-1: I mean, he he's probably one of the more inspiring stories because even less than a year ago, he was on Lex Friedman and he was like, you know, not not quite sure about all this agent coding stuff. I like to code, it's still a craft. And now he's completely opposite, right? He's he's figured out how to engineer systems that literally run twenty four by seven and build useful, valuable things. He's created an operating system on his own, virtually, like him and maybe one or two other people. Right. It's open source now. that is never possible before, but you have to figure out how to engineer a system like that. And not, I mean, I hate to say it, but all the like most of the people we employ are not those kind of engineers, right? Because part of this is is not even are they not engineers, I would say most people aren't design thinkers in that way. Right. Like, let me think about the work that has to be done so I can instruct something or someone else to do it, an agent. So are you finding you're having to teach people that skill or are they learning it naturally? speaker-0: I think funny enough, I will say it maybe in a high level title and then we can go deeper. I think we need we need to I find myself teaching people to think. And maybe now you said even better design because like I feel I get a lot of great outcomes with AI, mainly cloud but not only lot limited. I'm also using OpenAI and other tools. base forty four for example. and then I I I I think that It was in the past, it was how to write the prompt, which context to give, how to sometimes it's also how to from which angle to to get into the problem or the things that you are you're trying to to build. And I I I feel that like I I lack that a lot in in in many people. And then I'm l I'm looking in how to achieve that. And you know, to the point what you said before, and I want to take it to go to market, for example. To give a very easy example, is it like let's say that you want to expend your time and then you want someone to research, someone to work on marketing, to do some experiments, for SDRs to do some experiments. A lot of that, a lot of that work you can build with agents that agents run over the night, qualify, and this is even before. Speaking about TypeSafe, but now with TypeSafe it's crazy. You can run claude code and typesafe together and then you evaluate and you score it's a it's it's ridiculous what you can do in one night. But speaker-1: Yeah, let me let me let me add just for those who don't know what TypeSafe is, Jev is a is a model that a company called TypeSafe just released. And what it is is a very is it's a very simple model and it's very low cost, but it's a what it does really, really well is classifies things. So it puts things in categories and buckets and gives you back a specific value. So it doesn't spit out paragraphs when you ask it a question about, hey, is this is this a qualified candidate? It will not go through a qualified lead, it would not go through and explain to you why that's not a qualified lead or it is a qualified lead. It would give you true or false. And that is why it's type safe. And a confidence score. Right. And then you can teach it basically to do a lot of different things. You can give it values that it could return, but it's only going to give you those one of those specific values back as a classifier. So it's very efficient model for putting things in buckets and and categorizing and scoring things. speaker-0: Yeah, and and it's super fast. speaker-1: Super fast. speaker-0: I've I've I've run probably now six thousand requests to the API. None of those were above one second. All are under one second. Sure. speaker-1: And it probably cost you probably cost you a quarter forty five US dollar cents to do that, right? Yeah. speaker-0: Yeah, super cheap and I and I did and I did evaluate very interesting things. I'm now thinking about like trying to do some go to market things, but like I did on my personal thing, classified all my credit cards, expenses, etc. So so many different interesting things. and I think that these these are the beginning of the technologies that will change the world because this will this is a technology for agents. This is why it's super fast. it gives the confidence, it gives, enables you also to take decisions with agents and to complete tasks. best example for go to market, by the way, is you get a request from a customer and then now you need to decide which department is going to answer it. This is going to work 100%. speaker-1: Absolutely. speaker-0: And and and and that's crazy because like before that with Cloud or Chat GPT, you will wait, you would have waited forty to forty seconds to a few minutes and not necessarily you will get the right answer. speaker-1: That's right. And you have to decipher it from all the words that it returns for all the tokens. Right. All right. So I I cut you off there and and took us down a different path, but g you were you were saying speaker-0: I I I think it TAM is a great example. By the way we can take more examples of go to market, but I think you know, no one needs to wait for the weekend to get by for the TAM to continue working and then then you can s you can do very cheap, you can run a few exercises of TAM and in the end you can still send it to a human in the loop to decide which of these exercises were were the best. And which one was and and and and then to continue and then to s decide and then you can start working on the new time and see what are the results and according to that so I think it's endless and then this is time. You can we can go the entire funnel. You can classify which is the contact that we need to engage with, which is the account. Now we all of a sudden we have CSMs that they run a book of fifty fifty accounts. So now it's easier To decide which of the accounts do I need to work with tomorrow. And then some of these accounts are mid market or enterprise. They have 30 people that we can engage. Who is the person that I need to engage with? Over time, I think these these models can be so powerful that you in the end you are you are you are becoming an operating machine that makes less and less mistakes. speaker-1: But a hundred percent. So but the the hard part of all this is the harness around the models. Right. And and that's what Claude is. It's a harness around anthropics models, Opus five five, for example, right? Or mistr what whatever the model is that you're using. So is is that well, let me ask you this. Is this a question that I've been sort of keeping in the back pocket here, but What tools have you bought over the past year versus what have you built? Are you building your own harnesses or are you using harnesses in some other agent platform or within Anthropic or OpenAI? Like, how are you organizing all of this work? And I guess it we're getting in like that question takes us into like, what's the technology stack behind all of this? speaker-0: Yeah, so we obviously we moved from ChatGPT to Cloud, so we bought Cloud and then we bought Cloud for everyone and now we have budgets for consumption for everyone and we still have open AI and some users are using open AI, specifically a few individuals but also there's many people that now so we have cloud code is the main cloud main code tool, then we have cursor. And then we have now more and more people are wo going to Codex. But I don't know, maybe the new models will stop that. But I I know maybe I I I know for sure that I'm going to bet for multiple LLM for the next probably couple of couple of years we we must be multiple LLM. and and not by just API but by having also the tool having the the desktop app, having the in engagement with the company. so that's that's significant. and then I think that in in addition to that I think it's it's so strategic to understand the build and buy. I was always building and I think that it's important to build as now as the chief officer in the shoes of the chief operating officer. So I have all the go to market apps, I have all the Salesforce Netsuite, I have all the billing. So the things that we are the closest to the source or the destination, we must build. So Salesforce we must build, net suite we must build, billing. Nineteen ninety percent we must build, for sure. speaker-1: All right, hold on. T tell me what you mean by that. When you say Salesforce we must build. speaker-0: I I must have the developers for sales for I cannot put it on in a in system integrator and be be short. 'Cause it's in the end it's not one building block, it's many building blocks and like there's no there's no one task that is touching only Salesforce and it's only solved in the workflow, whatever, and it's easy. It's nothing is easy. It's getting easier, but it's always multiple. systems and it's always integration. So you have to do that. And then I think that when you go to new new areas, new technologies, vendor area, more application layer, then it's it goes more and more and more to buy and not to build. For example, AI gateway, I can build an AI gateway. speaker-1: But like open router? speaker-0: Yeah, but I I I but the best will be to be to buy. speaker-1: So what's the last thing you bought? speaker-0: So we we have many things. There's no last thing we we bought. We we bought a a lot of things for security. we invested a lot in make for automation. So the agents are not enough. The main tools are not enough. You also need automation. So make is a good investment that we we've made. and we we continue to look at that many, many different things. by the way, the cloud. Is something that from the product perspective, we are already invested in the cloud providers, but then now you invest even more. So we don't buy everything from Anthropic and OpenAI. Some of it comes with from BreadRock or Vertex or so you also need to do these investments. if I go a little bit to the go to market, I don't have a good thing that I can tell you I bought now. So we have clay and we have Salesforce. I just funny enough, I just renewed Salesforce for three years. I don't think Salesforce is going away. speaker-1: I was gonna ask you about Salesforce in in particular. Like yeah. I know from because I've known you a lot of years. Yeah. speaker-0: I don't think it I don't think it's going away. I think it's going to be harder and harder to get off of Salesforce, but they will need to make good choices because I I told them I I I already pay a lot for Salesforce and I'm not going to invest more. So they're just trying to sell and sell and sell and I'm not going to buy anything. I I don't need the agent whatever they call it or the revenue. speaker-1: This week. Yeah. speaker-0: Yeah, you don't need the Dream Force Week sales. speaker-1: Yeah, right. speaker-0: I I I I have read and write to Salesforce from Claude for the last six months. So like speaker-1: Well, I was gonna ask you about Salesforce and so my thesis on Salesforce, HubSpot, some of these other system of record, you could put workday from an HR standpoint, NetSuite from a ERP standpoint. My thesis is that if these guys want to stay relevant, that they should actually begin to figure out how to lower their prices, right? Make their own Jevons principle happen where the the technology becomes so inexpensive that more people choose to use it because it's valuable and stop necessary. I think actually Salesforce has taken a very and HubSpot has done this as well. I think they've taken the only stance they can take, which is I've got to be the the foundation layer, the data model, and I've got to let the agents be the agents, right? External to my platform. Yes, there might be some internal agents that operate on the data that's in that, in that platform, but for the most part, I have to I have to basically seed the agent layer, right? I mean if if worst case scenario for them is they get bought by anthropic, right? And become sort of the the foundation of anthropic CRM play, which ultimately they they might make, right? But speaker-0: I'm saying since July, I'm saying that I think hundred percent anthropic will buy Salesforce. I think that the the the the the the the challenge is Salesforce has is that most of the data is not in Salesforce. Right. Also the data points are not in Salesforce, they they try s strategically to get it. And then I think that they're they're they're Stand is still much better than Upspot or any other player, and I think that they've made a lot of great decisions in i the first and foremost is to invest in Anthropic. speaker-1: Fifty million turned into five billion? Yep. speaker-0: at least. speaker-1: I think it's at least. Yeah. It would be good. speaker-0: And then I think that the biggest asset that Salesforce has right now is their contracts. speaker-1: Hundred percent. The contracts and the trust that they have with those clients. speaker-0: Yeah, and and I I'll tell you even something that I said to a few executives at Salesforce. And I said like I think that they are late a little bit to the game. They need to be there where where Bedrock is, Amazon Bedrock. Is that I just I I I will even say the the the the the sounds, but like I pay around one million dollars to Salesforce, I will pay I I I bought a minimum of one million dollars in Anthropic, but I will pay probably two or three million. Sure. But Salesforce should have came to me and said to me, you are going to do three million in Anthropic. I will sell you that if for two and a half million. And then hence my contract will with Salesforce will be three point one, th three point five, sorry. that's the str the the best strategy they could have speaker-1: So you think they should resell Claude? speaker-0: Yeah. They the by the way by the way they are anyhow because anything that I do in Slackbot is claude. And that's free. By the way, anyone that wants free tokens, they're available in Slackbot. speaker-1: Yeah. Interesting. I mean, d it will certainly erode their margins to do that, right? Because most of that money is going back out the door to anthropic, especially if they do it at a discount. But that's what everybody's doing right now. Everybody speaker-0: It's a GR. speaker-1: it's a GRR play, that's right. And it's a top line revenue play. It's not the margin play or the the net income play, I think, is sort of dead for the moment. I was actually thinking about this, you know, I use granola. It's not our primary call recording tool. What you're looking at here is Fathom, which is a fine tool. But I use granola too, because sometimes I'm in person for meetings and I use it on my phone. I have never once they haven't billed me for anything yet. Granola. So everybody is trying to capture market share right now. So I think I think you're right. Like that the relationship with the customers is what's going to matter in the future. And so to sacrifice margin is a smart smart play. speaker-0: I I think for us customer success people, I think that it's a g b by far the the best era ever, that everyone are are focusing on value. And you said granola, they don't even build. I use plod for the last three months, both both physically and on the laptop and it it's crazy good. And then I think that, you know, everything that we spoke about now is that with between chat gpt and cloud and insting and muse and grokbot and even plot granola in the end we are mixing everything professionally and personally and we just we we buy value and this is the value era and i think that it's a great era for customer success i'm I'm very excited for customer success people doesn't matter if it's going to be a few more CSNs or a few less CSS or if CS if customer success is going to sit here or there. it doesn't matter because it's all going to be a value play. And the the by the way, I f I see it over the last seven years. I see more and more customer success excellent customer success professionals are becoming account managers. speaker-1: Yeah. speaker-0: People that are able to sell and upsell customers. speaker-1: Yep. Okay. I have a an interesting story that I wanted to run by you. I think it's interesting. I I read it I can't remember what book I was reading. I was reading a book called Scientific Advertising, actually, by a guy named Claude Hopkins. And it was talking in there about how the how companies used to sell the first motorized sewing machines. Okay, this is back in late, probably late 1800s or early 1900s. What they used to do is they used to go door to door, first of all, to find the wives who are home and say, Hey, like we will give you this sewing, this motorized show sewing machine or this motor for your sewing machine for you know a month. And for the first two weeks, you're also going to have access to this person who is going to be your on call help. He can, he or she can probably he can help you use the motorized sewing machine, teach you how to do it, teach you how to adopt this new. technology. What does that sound like? That sounds to me like what used to be a customer success manager, right? Somebody who would help drive adoption of a new technology. Well First of all, SaaS technology is twenty five, thirty years old now, right? Do we really need that anymore? We've automated a lot of that. We've digitized it. We've put a lot of the adoption into our products. Like if customers aren't adopting your product, like your product sucks, right? Your product's too hard to use. It's too clunky, or it's not a good fit for the market, right? That that that could be a reality. So I think now more and more. I think we get back to this idea of what you just said, which is a customer success manager needs to be somebody who's helping drive the growth in the account, not just somebody who's driving adoption. And yes, by virtue of driving growth in the account, you have to make sure that adoption is happening. Right. And do what you can to support that element of it. Yeah. speaker-0: I would I would reverse it. I would reverse what you just said. I don't care about the growth of the account. I think that the most successful companies, and it's including I'm speaking about Upslier, but I'm also speaking about many of the different technologies that we spok just spoke about. including, by the way, there's a big play about about it in the Anthropic and OpenAI. Is it how do I grow? The customers, how do I drive outcomes for the customer? For App Slier, it's how do you drive growth marketing? But I think that in a lot of a lot of things, and I think this is why now everyone are running the FDEs, they think they do FDE. I I think a lot of companies are not doing FD of what FDE means really, because FD is not professional services, FDE is net dollar retention. And then but the outcome must be for the customer, and it must be it it has to be judged by the customer. So like you go to the customer, you are the customer success manager, you drive both adoption value and outcomes, and you get the judge, the customer, to say what it means for their bottom line, what it did for them. for their business. And that's the best case. Like I think the best case, the best like the for me, the Eureka QBR moment is where you hear the customer say what was operationally being done, what it means strategically to to the customer, and then the the number. By the by the end the number. Because and and then if you are a mission critical product like App Slier or Claude or Salesforce, then you are able to say what it means. Because in the end, I'm I'm I'm now doing a very interesting task. Part of my vision is to say you can take any company. for me it's upstarry in a 1250 employee company, and just to say like for every department, for every profession, for every job, on one side What is the investment that I I gave the employee as the budget for ChatGPT Cloud and the other tool? Some employees also cursor. And what is the output that I'm getting? And it if it's an SDR, I by the way, I have a very interesting, I have a nice slide about it if you want to see, but so if it's an SDR compared to three years ago before AI. How many calls, how many meetings for salespeople you created three years ago versus what you get now? for a salesperson, how many opportunities are you managing and progressing over a month? For CSMs, CSQLs and engagement with customers. What I'm trying to do by the way is like to make sure that it's it's end of the funnel speaker-1: The outcome. speaker-0: outcomes. I don't care about the the way it has to be end of the funnel and it to has to be quality KPI. So let's see if I find this fast. speaker-1: The out the outcome before the role, the reason the role actually exists. Yeah. speaker-0: But what it what it drives for the company. speaker-1: Right. The the business impact it has. speaker-0: Yeah. speaker-1: Yeah, I'd love to see the slide. speaker-0: Pois é. speaker-1: so I know I know we're running out of time here. When are you coming back to the US again? We we have yet to meet in person. speaker-0: week first week of November I'm in San Francisco area. This is the slide. So speaker-1: this is great. speaker-0: can we share this? speaker-1: Can I share this with everybody? Okay. I'm gonna take a screenshot of it. speaker-0: Yeah, I can send it to you also later. what what what this came from so many different discussions with consultants and people that helped me get into the role. And so in in the end the lagging indicators or the company outcomes, it's what interesting for the boardroom. speaker-1: Yeah. That's right. speaker-0: And and you have to be very careful about this because some of it is a great trend that you want to measure for the entire company. And some of it in this era, specifically now with more AI, more AI tools, more AI products to sell, in some areas you want the cost per employee to increase. But not not in general. And then revenue from new products. And for some companies, obviously the revenue from new native AI products. And in the end, obviously the revenue growth year over year. And then you have functions, and then you have operational measures. And I think one of the biggest ones that I you know, I spoke to many vendors and many people in the industry, in the AI industry. And in the end, a big measure is AI leverage. Is to say if I took the example before of the N SDR, before AI they produce five meetings a month. And either now they can produce 10, so their AI leverage is 100%. And and hence that that team can now ship 2x. If you get the entire company to an AI leverage of 100%, you have double the outcome. You have the same number of employees, but double the outcome. And it has to be in the end measured by every everything, as we said, end of the funnel, qualitative win rate, wins. more products, monetize faster, reclaiming time. But here this is a this is a a bad thing that it's you have to be strategic about the time that you reclaim because you need to invest it in something else that will produce something an outcome. And in the end the AI leverage goes to the work without with that go connects to what I said before about working 24/7. If you know which agents to run, then you will be able to create more of the work without me, which means AI leverage, which means the multipliers that we're looking for. speaker-1: Yeah. This is awesome. I love how you said you have to be careful too. It brought brought to mind this idea of Goodhardt's law, which is a metric that becomes a target ceases to become a good metric to measure the business, right? So very careful with with those kind of things. Well, hey, we're thank thanks for taking the time to do this, especially on short notice. This is a really good conversation and let's do it again. I loved I love checking in, as you know, and I always learn a lot when I talk to you. So selfishly, I'd like to talk to you, talk to you more and someday we'll meet in person. So speaker-0: Absolutely, absolutely, for sure. speaker-1: All right, Ziv. Good to see you, brother. speaker-0: Same here, same year. Thank you so much.