AI, Sales + GTM in 2025/2026: This Changes Everything with Jason Lemkin and Owner CRO Kyle Norton

So, you just said two things that I think are super important, and the theme of the opener yesterday was tough love. So, I'm gonna make a tough love statement, because what you just said I think is so important is that you, as the CRO, are in the tools every day and you're super interested in it. You're genuinely curious. If your sales leader that you're thinking of hiring, especially that you have, is not in the tools and worse, not curious, are they going to make it today? You're cooked. You're cooked. You're cooked, right?

Hey, are you tired of listening to hours and hours of sales calls? Recording is yesterday's game, folks. Yesterday's game. Attention.com unleashes an army of AI sales agents that autoupdate your CRM, build custom sales decks, spot cross-sell signals, and score calls even before the coffee is cold. Even before the coffee is cold. Teams like Bamboo HR and Scale AI already automate their sales and revops using customer conversations. Step into the future at attention.com/saster.

Thank you to the early bird. Very kind, day three. Somehow, by day three, 9:00 becomes very early. Kyle, thanks for coming. Thanks for having me.

A little background, maybe a month or a half ago, we do a Saster Workshop Wednesday where we do a live meetup. I did a version of this AI and how sales is changing a month ago, and I went to redo it, and so much has changed in one month. So, I want to talk through where I think it's going to happen. And then Kyle is free. If you followed SAS, you've probably seen Kyle before. He is a CRO of Owner, which just raised at a billion-dollar valuation the other day. This week, Tuesday, you announced it, just for SAS. Tuesday. Tuesday. A billion-dollar valuation. So, congratulations to Kyle.

But even better, Kyle was close to the first meeting, and you were not the first person in sales, but you probably joined around a millionish of revenue, right? Two and two. Two. And in fact, he joined incredible founders. You can watch all the stuff on SAS, but maybe even while the company was in transition, right? Your first couple months might not have even been the easiest, the smoothest ones, right? And several stages. So, pretty impressive.

And I've watched Kyle do something. Two things that are germane to this conversation. One is they're doing SMB sales to tiny restaurants that do not read email, okay? That do not do this. The ACV is higher than you might think; it's approaching 10K. But this is a hard sale, okay? It's a full high four-figure sale to tiny businesses and vertical SAS that do not buy software, that are not looking for things to bolt on to Salesforce or HubSpot, that probably don't give a crap what CRM is, right? And Kyle has crushed it in general. So, we could do a whole section on the secrets, but then Kyle is just very curious about sales. He's a student of sales. And so whatever AI is today in sales, but to be more interesting where it's going to be the next, I couldn't think of someone better to join the conversation.

So is that a good bit of background?

Yeah, we're trying to put AI in everywhere we can, and I've mandated to my team that this has to be a mission-critical priority. So I'm thinking a lot about it everywhere, and I want—I've got a bunch of stuff I want to talk about, but when you said mandate for your team, because this is a question we got even to Yammen yesterday, what does that mean? How do you show the team the path forward for AI and sales?

It has to start with you as a leader. I'm in these tools constantly. I'm definitely the most well-versed in this area of the business. And I think I find a genuine curiosity. I'm interested in it. And also, like, I I'm I have such high conviction that this is going to be the most important thing that we can figure out in the next 12 to 24 months. And so anything that I'm doing, I'm putting into our leadership channels, my personal Clay workflows, or the ChatGPT projects like that I use individually for me as a leader. And I take a lot of demos. I talk to a lot of founders. I listen to a lot of content to try to stay on the cutting edge with it, because like you said, it's changed radically in 30, 60, 90 days. And the risk/reward is so big that it takes up a lot of my mind share.

So you just said two things that I think are super important, and the theme of the opener yesterday was tough love. So I'm going to make a tough love statement, because what you just said I think is so important is that you as the CRO are in the tools every day and you're super interested in it. You're genuinely curious. If your sales leader that you're thinking of hiring, especially that you have, is not in the tools and worse, not curious, are they going to make it today? You're cooked. You're cooked. You're cooked, right? Yeah, because the competitive pressure is so great; if you are not making this a mission-critical priority, your competitor is, and if you look at the gap in productivity between us and our main competitors, it's so drastic. I can spend so much more money on a per-person basis than anybody else we compete with in marketing, in post-sales, because our machine is anywhere between three and four times more efficient on a per-rep basis.

Yeah, I want to come back to that 1.5x, 3 to 4x is pretty big. I want to come back to that, but you're cooked, right? I genuinely think and that if anyone on your executive team, or honestly, I know this is tough love, anyone on your team that by June 30 is not deeply curious, like is not pretending to use ChatGPT once a week to get recipes, your team should be both scared and a kid in a candy store. A kid in a candy store. If your team is not—this is effing cool, right? Every week I use a new tool, right? And I was slow. I—for—and I want to talk about this—I was into AI as a student, but until we launched our SAS again, I said we had 130,000 chats in 8 weeks, then I've got it, and now I use—I mean, I was using, but now I use brand-new tools every day, like all the little videos I made, I use this app called Hicksfield AI, which is super cool, like a lot of the design, this was done on AI, our design—we have the best designer in the world—but this whole idea was done in Reeve.ai. So if you don't see these people, if they're not talking in the hallways and talking on the weekends and on Slack about this cool what you can do in AI and GDM, GTM, I think at least I give them till June 30. By June 30, if not—if July 1, I would like just start taking—I know it's mean, but start taking the blacking to the org chart because they had their chance, right? They had their chance. And I've already done this from a hiring perspective. Anybody in a leadership role. So in any leadership role, there's a component of the interview that's now—what are you doing with AI? What do you know about it? How are you using it today? And I really want to unpack it, and it's been a mandate. I don't care how good this person is; if they're not in the tools and what I consider to be AI-native, like we can't hire them. And I have a half-drafted Tobyesque letter for my team. Yeah. Toby from Shopify who wrote a letter to all of Shopify about how everybody needs to be AI-native. I've drafted the same thing for my management team and again, but it can't be superficial, like I'm—you might not have to get here, but like I'm in windsurf vibe coding my own apps for internal use as a CRO, yeah, as a non—I don't know, you told—engineer as a kid, no, okay, you're in windsurf. I took one Python course in early 2023 when ChatGPT was blowing up just to learn a little bit about it, but now with windsurf, I can legitimately—so I've got get my GitHub repos, and I've got it all set up so that I can actually do what I want to do. And so I'm pulling stuff out of the Lever API to run analysis on how we're hiring. Like I can do—you can do so much more if you put your mind to it. And that maybe you don't need to get to that bar right away, but it can't be—yeah, recipes and ChatGPT, that doesn't qualify.

And so when you're interviewing an AE or an SDR today and remind me the ratio of like general outbound to inbound owner, but when you're interviewing for either role, what's a good answer to that question? What's a good answer to the interview?

I'm not actually as concerned for ICs. Yeah, because we are running an SMB motion, and so most of the automation and AI is centralized. I'm—it's in my Bisops and RevOps team. I'm now hiring for a GTM AI lead. So, somebody who will do nothing but this within the RevOps organization. And so, I don't actually need the reps in a bunch of tools messing around. Okay. But, do they need to be AI-curious? I think long-term they should, but for what I need them to do, I don't actually care because we're do. So, if you were an enterprise rep, absolutely. But for me, I'm going to do the AI call prep and sell between work and the CRM enrichment. I'm going to do that en masse for everybody on their behalf. Okay? I'm hiring one-to-three-year-tenure reps on the upslope of their career. And I don't need them to be recreating the system. They need to operate within my system and execute the methodology, but for the betterment of their career prospects, yeah, they should be trying to learn it themselves.

It's just funny. I think I—that makes sense to me, and I want to ask you then, what stack do they walk into today? What's this? And are they more efficient? But it's interesting if you're hiring folks with two to three years of experience, right? They could be any age, but let's say they're 24 on average, right? I wonder if the ones you hire in two to three, that's going to be okay. Two to three more years. The kids today that are graduating high school or college, if they're not faking fluent. Yeah. Even in sales. Yeah. I don't know, man. Yeah. And we talk—I talk about it a lot across the team. So, I do hope to seed that curiosity; it's not as much of a mission-critical thing. But if you're a manager and you don't have a high degree of proficiency to just default to AI automatically for any problem you're solving, then it's really difficult to think that person is going to scale with the chance.

All right. Okay. So, help us then. Let's be tactical about it. I walk into Owner. I'm a new AE with three years of experience. What do I walk into? And does that mean I can do more meetings, more calls, more outreach, more texts? Like how does that—what do I walk into and how does that 3 to—you said 3 to 4x, I believe you—you're very—you're not an exaggerator—that's a big 3 to 4x—so what do I walk into and how does—how does that output happen? It's a very low bar for some of our competitors, I know, but so it's—everything is on Salesforce is the foundation, then on top of that we use Momentum for automated CRM enrichment, and you can basically do whatever you want with Momentum. It can look at call transcripts, run custom prompts, fill in the Salesforce field, run an automation, ping somebody in Slack, you said the wrong thing on a deal, something that we know is the wrong expectation that'll go to the manager and the rep and be like, "Jason said this thing, yeah, we don't do that," so that they can correct it before the deal closes. So Momentum is the foundation for a bunch of our—a bunch of our automation, and that's been a big focus. Automate everything, like what is in the rep's way of being on sales calls and following up and generating revenue every single minute of the day. Yeah. And so we've gone from—we've—we're probably up 25 to 30% in terms of the revenue-generating activity time because they don't have to fill out CRM after. They don't have to update the notes for the launch team. It's all done. It's all done for them. That's—is that a—it's a classic KPI. Is that a top KPI, getting that above 30% time in deals?

I would like the RGA time to be like 70, 80%. 70, 80%. We've gotten them 25 to 30% more revenue-generating activity time. And how many—I remember back in the day when Brex was taking off and Sam was really into efficiency pre-AI. The goal was at least four—four appointments on their calendar for them. What is—what's the version of that today? So for us it's four demos. Four demos a day. Still similar number. Three. He—I think Sam talks about it—four meetings of any type. We're trying to get to four demos. Three held, cuz it's SMB. So there's a nutrition—nutrition rate. And then probably two follow-ups and then a healthy amount of white space where they're calling open opportunities. They're doing a little bit of pipeline generation, but actually not much on the AE side. Yep. And then the call prep, call followup, CRM, handoff stuff is at a minimum.

So, let's talk about AI and followup. You probably didn't see it, but I—I tweeted this, and I told this story of this, and I wrote a Saster post on great AI outreach, terrible follow-up.

I did see this. Did you see this one? I don't know if folks saw it. I don't want to repeat it, but like a week ago, I literally got the best outbound email ever. It's like for our podcast and video, like this session. They said, "I'll take your session with Kyle. I'll make the world's best clips. I'll automate all the editing. I'll put it on iTunes and Shopify and YouTube for you. I'll do it all for you. It'll all be—I guarantee it will be better than what you're doing today." And it was like—and this is exact. And here are my clients just like Saster that will do it. And this is a week ago. So I'm literally like—I'm in. He's—you want to do a demo? I'm like, no. What is it? I'm like, I didn't care what it cost, as they said, as long as it's fair. I have no time. I'll buy today. So I was in, and then he sends a follow-up email. What is Saster doing and how many podcasts have you done? Yeah. Rough. Okay. Yeah. And then yesterday, I told the story. I'll tell the follow-up, but I want to hear your learn. Yesterday, I told the story before, but I didn't tell the follow-up. Yesterday morning, I got an email from a well-known B2B company. "Hey, Jason, saw you'll be at Saster next week. Would you like to meet next week?" Yeah. "Hey, Jason, I know you might be busy." So many flaws in that. Yes, I might be busy. Two, this is the worst week to meet me of all the 50 weeks of the year, but next week. Okay. So, then I email him back. Worst outbound ever. I wrote in all caps. I really can't believe, because I just want to see. I'm not upset. I just want to see. And then the guy writes back, "This is what happens with bad AI automation."

Yeah, I got one of these recently, too. Did you get one of them, too? Maybe the same company. So, it's poor orchestration. I think this is when you have people trying to apply AI for the sake of it, but they're not thinking through customer journey, the entire thing. But how do you do that, Owner? So, Owner—like restaurants do not suffer fools. I have no time. No. And your product's expensive. Yeah. So, how does that—how do you automate the followup?

So it's great. I have a great RevOps leader who is very fluent in sales and the actual go-to-market fundamentals. I think I see this mistake where you have RevOps people investigating tools and then applying them, but then there's no thought of, okay, how does this get to a rep? How do they follow up? How does the manager then look at what's going on? And so I'm in the process of building what I call the management intelligence layer, which is—what is the always-on layer for the manager to understand what's happening on a rep-by-rep basis and get prompted when things are either off-base or off-track, and so that they can nudge the rep back into track, back onto what they should be doing. And so you have to think through, okay, there's AI tools, but then there's also how does the—how does the calendar of the manager look and how does that integrate with what they're trying to achieve? And so we—we spent 8 weeks at the beginning of the year redeploying what I call my like core operating methodology. And so we wiped every single meeting off of the calendar, started from a zero base and said, "Okay, what are the most important activities? Coaching is number one. Okay, let's plant the coaching into the manager's calendars." And then it's—there's like some audit mechanisms, and then how do you make the audit mechanisms as efficient and effective as possible. And so we're starting from—what does the business outcome need to look like and then layering in AI to accelerate all those things. Not saying, okay, I want to use AI for this thing and then trying to fit your business into it. And so I think it's really about being thoughtful about how you're orchestrating the entire business and then using AI as a leverage point as opposed to, "Oh, this AI thing's cool, so I'm going to slap it in."

Okay, let's go to the future a little bit before I—I need—I should know this—the answer to this question. I don't—what's our typical sales cycle at Owner and how many interactions on average does it take until you close a restaurant?

A lot of one-call closes, one closes, sub-7-day sales cycle. Okay. And so the first touch point will be a BDR or an SDR following up. They book a meeting. They do a confirmation call, and then the demo happens. Lots of calls closed on that. If you're not closing a call in the demo, it's usually on that follow-up call. And if you're not getting them on that follow-up call, the win rate falls precipitously. And so it's almost not worth talking about what happens after that. So the demo may be a good challenge to this little tweet I have. But I still want to—this is still what I think is going to happen a lot. I wrote this April 2024, "AI can't beat a good sales rep. No way. It's dumb. It hallucinates." April 2025, maybe an okay rep, right? Yeah. And then April 2026, "Our AI is better than 80% of reps." I want to hear what you think about that. And then I pulled this. Clary just published this data. All—maybe you saw it too, right? I always knew this, but man, this is stark. Clary's—Clary I think was like a million opportunities or something. The top 10% of reps close 65% of deals. Yeah. I know you have an—a world-class team, but for these one-call closes, especially when the customer is more sophisticated, yours often isn't, but if I've tried Owner, maybe in a year when everyone has used Owner, do I really need a mediocre sales rep or can AI close that deal in a year?

We are augmenting reps with AI and not trying to replace as much. But we are building a PLG motion that will be heavily AI. You're going deep on PLG, right? Yeah. And we're in the process of launching our OneMind superhuman. So, it's Amanda Ko's new business, founder of Six Sense, and that will deliver a lot of the sales experience. So, they'll answer questions on the website, build problem awareness, book meetings, answer a bunch of questions. And me as a buyer, I had a lot of conviction about this OneMind investment because as a buyer, I was going to their superhuman and asking a bunch of questions. How does it integrate? How does it do this? And I was like, okay, this will work. Yeah. And it's a question mark if our SMB restaurant buyer is going to be as interested in that experience as I was. But that will do a lot of the sales funnel. And so what I'm hoping to have is a part of the sales team that is really just order-taking, and they are doing 15–25-minute—"Hey, I'm a human. Let me make you feel special. Answer a couple questions. Swipe a card." Swipe a card. I do believe that the superhuman will take some portion of people all the way through the finish line, but I—we're also building a customer journey that is not AI and then…

Human, over here. I'm trying to build what I call semi-self-serve, which is, yep, a consistent customer journey that has different channels, different experience channels. You can come onto the website, talk to the superhuman; you can talk to a rep; you can go through the product-led journey in the product where you've got your walkthrough and asynchronous demos.

And what we want to build—because where I see people go wrong sometimes—is a way for a buyer to go through that journey, interact in the ways that they want and not get locked into this attribution problem or these silos where you're shipping the org chart, and if you go through PLG and you don't buy, like there's no follow-up. No, I want a sales. That's the worst. It should be—it should be an automated spectrum, right? You should get whatever version, from PLG to handheld sales, that I want—that the buyer wants. Shouldn't I? Yeah.

And and people look at that, they go, "Oh, I'm going to pay a sales commission on this like PLG buyer," and I think it's the wrong way to look at it. I'm really trying to think holistically about the business, holistically about that customer journey, and let people go back and forth, channel to channel, experience to experience, and we're just going to close them wherever we want. And so you could have a rep that is a recent BDR promote that is closing 50 deals in a month at a 910k ACV because people are coming to them very warm. Awesome economics.

I don't need to squabble over—definitely you don't need to.

Yeah. A related topic, but you worst case you got to double pay for a while. The AI, you got to pay the human for the AI. Great. And anyone that's done a hybrid premium self-serve, this is not—this is pre-AI. You got to do it. If you get cheap on this, it creates too many conflicts. But it was interesting. Even though at first you're like, "I don't—I don't know that an AI is going to close these deals 100%." But then you have PLG. So you already have a continuum.

In a year or two, Owner will be deep into nine figures in revenue, and you're going to get your second and third-generation customers. They've already used Owner before. Yeah. And but they went to a—they went—they—that restaurant shut down, or they got a new job. Yeah, they might not want to talk to a person. Especially if the product has a flexible contract, especially if it's a good deal. I already know. I got a couple questions, Kyle.

Okay. I've moved to a New Mexican restaurant in Orange County. New point of sale. Do you—Yeah. Do you integrate with it? Do you integrate with—I got three questions. I'm in—I've already used it before, and I don't—that AI might be better than even on your team might be better than a human. So that—that it could be a material number of deals closing a year, couldn't it? With AI.

It will for sure. Yeah, it will. Right. Yeah. And it gives your team leverage to do higher and higher value activities, go after bigger, more complex customers. The same thing with the onboarding team. We're really trying to onboard and leverage AI to automate a ton of what they do, but it's not to necessarily replace a whole ton of them. Although we will definitely slow the headcount growth in that organization, but there's so many high-lever things that we could do with a customer in onboarding: teach them how to use the product better, make sure that they're like really well enabled when they launch Owner, to celebrate it with their customers and do a marketing campaign. We—we don't have the capacity to do some of that stuff now, but once you start to—to automate a ton of the more rote tasks that we do with people now, then you can move people to a higher-order activity.

I think—yeah, let me ask a specific question. Brian Halligan, who's better than me at every level, co-founder of Hubspot, but he said many times what he thinks is going to happen is AI is going to lead to more—less specialization in sales, more generalization. I think the opposite, and I—but I mean this specific because you're giving a really interesting background. I think what may happen with AI sales teams is you may just have a demo team. You can call them sales, but if the one thing you need a human and an Owner is that S-tier demo, maybe the AI does everything. It seamlessly grabs a person not next week when their calendar is free, but right now while they're on the phone. A human does whatever demo needs and then digital Jason can take over the rest if they want, or I can, right? But you might have these interesting—you might have this human part of onboarding where you need a little bit of human—a human demo and then this human—help me. I need more help than the AI. But then those might be like hyper-specialized rather than hyper-generalized. I—I certainly lean here. This is—we get into this big debate about the full-cycle rep. I'm not into the full-cycle rep because I believe much more in specialization. A—don't want to prospect. Cold calling is hard. Unless it's their only job. It's—there's no chance they'll prospect, is it?

Exactly. And so I am a big believer in specialization. And similar to you, how do we think about where we apply the most expensive human capital to specific parts of the customer journey? Yeah. But we'll see. One thing I just—and I want to ask you a follow-up on this before this one. One learning I had from you—I just hadn't thought about it this way—I think AI in sales—I have a slide later we probably won't get to—is going to lead to a big renaissance in self-serve and PLG because if you want an AI to close a deal, it can't be a three-year contract arguing over discounts and terms. The AI has to be able to close it with zero friction, no games. Yeah. So, I think that's going to happen. People argue—I don't think—but your point I think is even more thoughtful, which is that it may blur all the lines—the lines between self-serve, commercial, mid-market—maybe not deal size—but all the ways right now we have different teams interacting with different types of buyers and it's shitty for the buyer—I—whether you're business—whether I'm a tiniest restaurant, mid-size—I've got five restaurants—I'm bigger—it's just as Owner—just as important to me—it's a big deal—so why do I have to be segregated to some weird part—it's the game. Why shouldn't I have—why shouldn't all the experience be fluent whether I self-serve, little help, a lot of help—and so it may blur all these lines between the different categories.

Right, I think so. Yeah, I'm—I'm very anti-attribution in general, like I think we get into problems where we say, "Oh, this is a marketing-attributed, sourced deal versus a sales-attributed one," end up making the local maxima decisions as opposed to saying, "All right, how do we maximize this entire system?" and the credit. Let's not worry as much about the credit. And I think we're going to run into this because you'll have a growth team that reports into product that's doing all the AI stuff that's taking sales as deals. I—a CRO said this to me like a couple days ago. They were frustrated by—like, "What's your job here? Is it to hit your sales number or is it to build enterprise value?" You should only be here if you're really doing this for a significant equity outcome. And so trying to get people's headspace around this like more holistic understanding of building a business is—I think—going to be really important because you're going to see this—you're going to see these silly silos where it's broken and things don't—aren't communicated across, and we're trying to avoid that. Yeah. I think it's a niche—we'll see.

I don't have any passion on this, but I think by the end of next year, a point that everyone from Yam Ranga and everyone's agreed is—all through the dawn of time—we've added better and better tooling. We added SalesLoft and Outreach and Gong, but sales reps still close the same 3 to 5. They're closing 3 to 5 times their OTE. They're closing 500k, 4, 800, whatever. I think AI is not only going to boost that up, like your point. I actually think what's going to happen because of attribution is commissions are going to have—by the end of next year. Yeah. 'Cause this is how the math's going to work. You're going to get credit for the—for—it's not just PLG. You're going to get credit for Kyle's AI closing the deal without you. Yeah. Okay. But if that doubles the output of your org inherently—the—instead of whatever you take home—10% overall—15—it's going to be 5, but you get everything—whatever your patch is—team—but I think commissions will have to have—as part of this—definitely. I think where there will be some pressure the other way is for the best sellers—if you have an elite mid-market and enterprise seller—they will have so much more leverage with all of these tools that the competitive dynamic will say—it was only that seller was worth 350K to me before, but now look at what they're going to produce. Like, "I'll pay 500 grand for that person to pull them from somewhere else." And the best might make more, right? One way or the other, the best might make more, right?

I think so. And then what—And then let's go on the next point. But then let me—Okay, so you're developing an elite team. You work so hard to recruit the best. That's why you're here. You're here to help the team, but you're also here so people see you. You're constantly, as we've written since the beginning, like most of the best, you're constantly recruiting one way or the other, right? What about the guys we just talked about? The one that says, "Sasters next week," the—do they honestly—between us—do they have a job next year in this new world? Do they have a job? That mediocre rep working 15 hours a week from home with three side hustles, mostly working on their ADU and their Airbnb. This was everyone in 2021. Do they have a job?

I don't think they have a job next year. I think most—I think they're going to quickly lose all—and they're not going—and I don't think they're going to step up and become AI experts. I don't think they're going to go to your class at Pavilion and the next week wake up and be like, "Okay, I want to work 50 hours a week on this AI-enabled sales life." Yeah. What's going to happen to these—to the others?

I think you're right. Yeah. Would I make a $100,000 bet on it? Like you—I'll give you—I'll give you four to one. So, you just closed a billion-dollar fundraising. I won't give you 10 to one. I might give you 10 to one if you want. So, I agree with you very much. I think AI will contribute much more to elite performers. And there's some research—I can't remember where it came from—but there's a bunch of research that showed that AI for call center teams raised the floor significantly, and most of the gains were attributed to the lowest performing individuals. It brought low performers up to a better standard. But in the world of PhD research, AI was actually making the best researchers even better. So the gains were going to the most sophisticated people in that craft. Yeah, I think in many parts of sales you'll see the best get better, and there—it become an even bigger bifurcation, and if you're doing this side hustle thing—you want to work 20, 30 hours a week—yeah, it's going to be really tough—yeah—to survive. One of—one of the—I've known him for a long time—one of the best sales in IC sales—IC—I've known for years—right—emailed me this week while I was here—my middle—hey Jason, I'm looking for my next role—got any good ideas—and I emailed him back—"Have you gone to the fund website? Do you even know what companies like Owner I'm working with is now?" I haven't done that—if it was two years ago—I'd find him a job today—I—I just think it's hopeless—I just think it's—and this is a guy who's had multiple roles where he was in the top—top 10%, right—but he's running that playbook—he's always worked from home in shorts—okay—since the very beginning—now it's a Harry Stebings vibe—okay—but back then—I was like—I remember when he was hired—back then—I'm like—"We can't hire a guy that works from home in shorts"—but he crushed it—he crushed it—but man—that wasn't enough—effort—was emailing me in the middle of—in the middle of SaaStr. Didn't even know what's going on and just said, "Hey, dude. I'm looking for a job." Didn't—didn't even know what I'm involved with. Yeah, that's tough. He ain't going to—This guy is—But my point is he was great. I don't think he's going to make the jump. I don't think he's going to make the jump—without knowing him—I—I definitely lean in your direction. It's—it is going to be a very difficult market to survive in if you aren't still in a career stage where you are out there trying to figure out the next thing.

We haven't gone through this big of a dislocation since the beginning of the AI acceleration tool era, and I was super early. I was like one of the first Sales Cadence customers. Kyle Porter onboarded me himself, and we got like a 5x leverage on a per-BDR basis. The BDR team I inherited was like a dumpster fire. But all we had to do was some better training, and we gave them this tool that nobody else was using, and we still used that like predictable revenue playbook. Hey, I'm looking for the person that does this, and people would respond to that email and tell you the person, and then you forward it. That obviously doesn't work anymore, but the—the competitive advantage we got then was so significant that it changed the trajectory of companies. And the same thing is with AI. You're going to win or you're going to adapt and build a competitive advantage, or it's going to be really painful. And some of this will be competed away 'cause everybody's going to do it. And so thinking about how to do it most thoughtfully and be ahead of the curve—things have never been more difficult, I would think.

Yeah. So this one—I think we hit this slide—you—but feel free to add on top of it. This is—this is actually a slide from Gorgeous. We talked about it. They've got 20,000 more and more—more AI-driven contact centers and e-commerce in the Shopify ecosystem that Kyle knows. And this is them and their competitors. I've locked out the names, but they're—you don't know who it is. There's Zendesk, Intercom, others in there. And this is—this was—this is from a couple months ago, so it's already changed, but I'm color blind, but the big—the bar charts are what everyone says AI does. All their marketing sites say 60%. And these were the actual numbers based on a third-party study they paid for. Okay. So on the one end you can see that the—there's a delta between the marketing claims and reality, but also this—but it—but none of them are zero. And so what we're already seeing in the contact center is—yes, some of the claims are a little aggressive, right?—but folks are laying off 25%, 30% of their teams—not to—just to save money—but because the AI is better, and the AI is always in the top 10% of—it's never number one—it's never the best—and we can mock this—we say that's because it's doing easy issues, right?—but who cares—who cares if digital Kyle solves my freaking problem because I got the wrong watch when I ordered it online—who cares—but this is disruptive, and I think in the end of next year—maybe when we're here next year—this is going to start to happen to sales, and we're going to see—but Owner—it's only 10% because everything's a demo—but at Gorgeous, where you don't need a demo because most folks self-serve—half of it is done by AI—I think we're going to see charts like this next year for sales—100%.

Yeah. Then I think where you see people are going wrong right now, they go, "Oh, it's not good enough to do X and Y today," so they're like not investing. So we use too much. We hear that. So we use Crescendo for support automation, and the first couple months. Yeah. We're like, you know, you—you expect AI is going to be this like miraculous thing. And it was like, okay, but our knowledge base was crappy at the time. Yeah. And it would be really easy to be like, "Oh, I don't know. It's—it's not worth it." But then every week we produce 4 to 10 new knowledge-based articles, and that AI containment—that bot containment rate—jumps 3%, 4%, 6%, and now like almost 50% of our tier-one support tickets are AI-contained—%—and it wasn't easy, and AI made us realize that some deficiencies in how we were managing support—and so you just have to be betting on the future because that will—so it's 50% of tier one—it'll be 80% and tier one soon—and it's going to eat a bunch of tier two—and then what we use real support people for will be like very sophisticated stuff that—that you need more highly trained folks for. So I think you're going to see that everywhere—just so you started—and then I'll tie this to like what I learned from RAI—and then Dar from Delighted app is going to explain how it really works—but so you started with tool—you didn't upload enough of your documentation—you thought we uploaded everything—we realized that wasn't enough—it wasn't enough—right—that it didn't just magically work out of the box—so you added more—did you do QA—did you see mistakes and update and interact—yeah—every week there's a QA meeting with Crescendo and their team, and we go through what—what needed to be escalated to a person and why—and should that have been bot-contained. We go, "Oh, there's nothing in the knowledge base that explains how this is done," even though our support team might know how to do it. So then—but then they just go add to training every week. Yeah. And then we're using AI to create that information. It just gets better, right? Yeah. Through Guru. So we looked at a bunch of like knowledge-based AI stuff and ended up using Guru's AI. That's cool. Yeah. Rick's been doing this since the beginning. Super smart, the CEO here. Yeah. And that's a great example of a company who could have been massively disrupted, but they went all in really fast. And we looked at a bunch of comparative things and we're like, "Oh, the best tool is the one we have. We just need to turn it on." And so now when there's answers in Slack, it pulls that into a knowledge-based suggestion. We do a quick—the knowledge team does a quick edit, clicks publish, it's in the knowledge base, it gets picked up on the Crescendo side. And so there's this flywheel now of issue identification, content creation, deployment, and that loop is fast.

Yeah. For us and Kyle's stories, but just to tie it together, if you go to try RAI, it's on saastr.com in the chat bubbler in the middle of the homepage. And you can see it here. It's trained. It took about a day to train it on 17.8 million words of content. And it was pretty good. And already was great because maybe it had more content than your knowledge base. It has 12 years of SaaS content, very specific. And then once in a while—and then I read as many of them as I can—and once in a while it—and I tell you what I learned specifically—the mistakes it makes. So it's not that many. It makes a mistake. So I—for—and I enjoy it. I like it like you did. I like this stuff. Right. So when I'm having my second cup of coffee in the morning, I see what the handful of mistakes it makes—a couple percent—and I just go—and I change it. It literally takes—I don't even need a team. I do it in two minutes. The only real hallucination it did, and it is interesting, is for SaaStr Europe, it kept telling everybody the wrong date for this year. And why did it do it? We have—and maybe Darra can explain to us how the ragging and vagging and vectoring works—but we've been doing this since 2017, but we didn't publish the date online. So, it was guessing. Yeah. And it kept guess—and it didn't guess the same, which is interesting. Sometimes it would say, "We'll see you in London in September." Sometimes it would say June because

We've had it at different dates. And I saw it a couple of times. So, I go, and there's literally a train thing. It took me 5 seconds, and I said, the question is, when's SAS Europa? I said, we're having Christmas in London the first week of December, and the next day, next time. So, if you just keep doing it, improving it. It's this is the thing people don't get if you haven't done it. If your AI is great out of the box, great, okay. If it does 20% of things, that's a lot better than zero, but just every just improve it.

Pick a platform. It doesn't actually matter if it's Guru or Crescendo or Gore, or it doesn't matter. And just commit to making it better every week, and you'll be shocked where it is in 8 weeks.

Right. Yeah. 8 weeks, it'll be so much better. And I don't know if I blame the vendors for this or the buyers, but everybody expects it to just work out of the box and be this magical, miraculous thing. No, it's hard work. And we had the advantage at Owner of being a very disciplined documentation culture. And so, like, we use this company called Avara as an AI sales simulator. Now it's an AI sales coach, but I write a lot, and because I we're a remote company, and I want everybody to have written documentation of all the things that we want to teach you. And so, when we got stood up with it, it very quickly did a pretty great job of mimicking not a generic challenger style or sand. This is my methodology that I've created over the years because it's all written. But that's really hard work, and we have the advantage that it was already done. But I think the buyers of these tools need to understand that you need to go do a bunch of the foundation work to create the content and tweak it every single day, every single week. I'm in the OneMind sim like on a every other day just doing sims and filling up the spreadsheet. It's not easy, but the leverage is so significant. I know that every hour I spend there, the incremental leverage of now this thing getting in front of thousands of website visitors every single week. You got to you got to do the work, and it's none of the work. None of it's magic. This is Yeah. People some of it to the prior slide is vendors overpromising. Let's be clear. But last SAS Angel, the last one was not a year ago. It was only in September, and everyone laughed about AIs. Okay. Tomas Tongas came and said he did the survey, and no one could get them to work. I didn't have an opinion, but I did it on our LinkedIn poll. We had 3,000 folks. Only 2% of folks said they got the SDRs to work. And so I'm all in, but but I missed the point. Yeah, I missed the point that faster day SDRs do work. In fact, the stories I just told, Yam said she responded to two SDRs. It's just you got to know what the tool can do. And every month it's going to get better. But expecting that the lazy approach to AI is also death. Yeah. The lazy, oh, I'm going to hire an agency to do AI SDRs. I'm going to just hire a clay agency, and I'm done. That's where you're also going to get slaughtered. You got to be in it all the time. Maybe forever. We're going to be We have to be what how we train our AIs in two years is probably nothing like today, but we're still going to be training them. Yeah. And maybe our AI will train our AIs, but And you'll be at some point managing a bunch of people and a bunch of agents and but that's real work. It's a different type of work, but it needs to happen.

Okay. We won't get through all of this, but I want to hear you as a case study on a few of the points that I've made in the past. This one that folks can't see, but I said not everyone has the quality of team of Owner, but I said AI for us, this is what I learned from our AI is already better than almost any BDR I've talked to in terms of qualification. The average kid coming out of one year of school that doesn't know how to spell Owner, that has done no research. That doesn't happen in Owner, but man, it happens. It happens elsewhere that can't answer any of my questions. Like, it's better. So, what I did for us is we have a prospectus for the sponsors here. It's a public-facing, but then we have this document, 100 pages that Amelia has developed on everything. How can I get ice cream in my booth? How big's what's the artwork look like? What's the DPI? What is it? Right on our little team of three, they don't know the answer, right? Yeah. I stuck it in our AI through Deli, and the next day it could answer all the questions. Mhm. And it started, and then it answered all the related questions, and then it answered a question. What's the difference between a silver booth and a gold booth and a bronze? What's it what and then all of a sudden it was better than any human because there's too much detail. Yeah. That amount of detail, you cannot you can only expect so many humans to know what's the size of the ice cream cone I can have in a silver booth and what time do I have to show up and how many leads will I get, and it said, on average, you get 218 leads for your type of person, so I'm like, I've never met a BDR this good. And I think so my view and challenge this, I think the mediocre BDR in a year will be out of business. I don't think we'll need any of these BDRs that when you reach out, I think a website should be an AI should be able to qualify, and no one likes to be qualified. I hate being I can smell the qualification; you it smells smarmy, and so I just don't see the need at all for this qualifying, not outbound, but this qualifying BDR very much in the future. I think for a ton of markets that's very true. I think there's some niche cases where you have a prospect that really wants to talk to a person, and for us, it's going to be really interesting to see the pickup rate of our OneMind superhuman with a restaurant buyer that is not very technology-savvy. And I would I'm hoping it'd be proven wrong and people to love and cuz it's AI atom. That's the thing. We've trained it to I think they're I think they're going to love it. They love AI. Now I know it's tech-focused, but we've had we they love AI, Jason better than me. And so I think for Yeah. I think for at least by volume, at least by volume, you can't remember every single thing you've ever written. It's impossible. And so I think for most markets this is very true. I think the SE thing is a great use case. Yeah, you have these AIs that can join a Zoom call, answer questions which are when you're called upon. I think there's a bunch of that's going to be really compelling, but then there are the solutions engineers that are great at asking the really incisive business questions and have worked with a bunch of similar customers, and there will be value. But again, it's the best will still have a place, and anybody that's punching a clock Yeah. like why wouldn't this be better?

Yeah. The other thing what everyone again, and then let me I just want to pick one related thing to get your thoughts on. This is one of my top learnings from RAI from these conversations is that people there and for a lot of reasons will prefer to talk to a great AI if it's mediocre. No one wants to talk to a bot. I think bot people still use this term bot. I would banish it because bot is like IVR. It just has negative maybe Darl will disagree. I just think it for me it has negative connotations. The last thing I want to do when I'm ready to buy is be forced to talk to a crappy bot. Okay. But when RAI is so good, it's like, great. Thank you so much. You're awesome. Thanks for doing People. This is to an AI. You're awesome. Thanks for doing what you're doing. You don't say that to a bot, do you? No. You're awesome. Again, thank you. Great help. Thank you. I would rather, especially when I'm working at 6 p.m., especially when I don't want to wait to get scheduled. I would rather talk to this S-tier AI and it knows the product cold. So, I want your thoughts in general, but there's a specific bit of hubris that many folks in sales have. So many. I'm a people person, Kyle. And people like to talk to people. Now, people like to talk to the best solution architect in the world that also deployed my product, even if we don't even use it at Owner, that also helped me get going at my my buddy down the street. And this is David, and he did it. I want to talk to David. Yeah. But this idea that you can be a people person and not be a product savant in tech, I just don't care that you know who won the NBA finals. Yeah. Hey, buddy. Like that. People think they're people persons in sales, and I think they're lazy and they're smoozers, and I don't think it's going to work in the age of AI. We're going to find out. I'm Do you think being a people person and not knowing Owner helps you close a deal? No. No. Absolutely not. But most salespeople think that, don't they? They're great people persons. Many do. Many do. And a challenger would tell you, the challenger sales book will tell you that profile is actually one of the worst-performing profiles out there. And what is the best-performing profile? It's the challenger. Somebody who can teach, tailor, and take control. Teach means you know the market, you know the product. Tailor, I can really listen to you. So I can do both. Yeah. Going to take the third one's take control. You can do that too. Yeah. It can I think it can do the best first two better than 90% of humans going to that first point from if you build it the right way. I think you've had great help building this. I had almost no help. Amazing. I appreciate the team. That's what I mean. I had an hour of help. Really? This is the point people are missing. The Deli team didn't they I had Listen, I was a little frustrated at the time. In all fairness, I got about an hour of help just because I had a little trouble with the ingestion. I got no help after that. Maybe in the background I got a little I've done it. Wow. I talked to people, they're like, "Yeah, I hired this AI," I'm like, "I did it. Listen, I'm a little tech-savvy, but I haven't pulled up my Python coding. Okay, I'm not using I love Windsor. I'm not using it. I do it. Not no one on the team. What? Good for Deli. That's a great Yeah, but all the tools in Deli is cool, but all the tools should be this good. There's no excuse in 2025. There's no excuse. They're definitely not. The This is I think a lot of sales tools are not as good as Windsurf, and they should be. Yeah. When sales tool, this is my view. The only reason sales tools aren't as good as Windsurf is because Verun isn't building a lot of them. Yeah. Okay. I just think, and there's no criticism, I love everyone that's here. There aren't enough S-tier founders. If an if a bunch of S-tier like the best in the world said, you know what? We're going to build an AE. Not all the other stuff. Not discovery, analysis, sentiment. I'm actually going to go further, Kyle. We're going to build you and me are going to hire the three best engineers out of OpenAI and Windsurf. And we are going to build a complete AE, not a partial AE. Okay, I think we could do it, and I think it would blow. I think it's possible today cuz it's possible in Windsurf. GTM tech often times doesn't attract those founders cuz it's so competitive, and people don't want to invest in it, and it's a go be that one. Be the Windsurf of your category, you will run away. I think I agree. I very much agree with that. Maybe Owner's a little bit of that. Yeah. I think that in restaurants a little bit. Yeah. I think there's a one of the challenges that so many of my CRO friends have is trying to figure out who to choose, and there's so many okay products that are a great demo but a really bad implementation, and there's a thousand people in every single category, and so you got this sort of paralysis, and so you have to be quite good at picking these early-stage winners. If you want to be early, it means you're partnering with really young companies. And I've had the advantage of doing a bunch of angel and being involved in some venture funds and following it very closely. So, I've been able to take my founder radar and apply it to the vendors I'm choosing. And so, I'm choosing founders that have that like A+ upside that have a really sharp opinion on the market and have probably done it before. And so, I've made some pretty good bets. Like, momentum is winning in their category. Avara like they're ripping like shipping product at a ridiculous rate, and I think that's a big reason we've been successful is because we've chosen good tools, and I see a lot of people choosing this like the most random things coming out of it's complicated. Our AI gets this question a lot. What should I use? Yeah, cuz it is confusing. My meta advice is when the world's changing, copy someone you respect, whoever it is. It doesn't need to be those that stack. It could be it but talk to someone you respect. Copy it and then go in with five but measured expectations. Go in and do not expect magic but do not expect it to be magical immediately. You're this is a commitment whatever tool you pick, Guru, Deli, whatever tool Artisan, which is very cool, all of them Ampar, and then every week you got to invest. Yeah. You have to invest for real, and you and this is going to make SoSs and COs have to be much more technical, not engineers, but much more technical. You're going to and then every whatever platform you pick is still immature today. It's still immature, but invest in it, improve it each week, and then the magic is going to your point. If you have a high-velocity team, the product itself will get better every month. It's small, but with Deli, like when we launched this 2 months ago, it couldn't ingest PDFs, okay, for whatever reasons. Then it added it, and now founders have uploaded like hundreds of decks for VC. Give me feedback on my board slides. Give me feedback. I didn't have to write a lot, do anything. It just added it, right? So, if you invest in a platform and the team is high velocity, every month it will get better, too. But you got to pick one and expect like you'll get half of what you hoped, right? And half's a lot. And you have to be really committed to making it work. You got to put the resources behind it. And as a CRO, like you have to consistently push people on it. The amount of times I have said in Slack, use momentum for that. Somebody will be talking about, oh, we audited 20 calls to see if people are doing I'm like, what are you doing? Use momentum. You just write a prompt, run it on every single call that ever happens, stuff it into a Salesforce field, get a Slack notification. And so I've pushed it so hard on so many things now the muscle is there for people to be to have an idea and go oh like I'm just going to talk about RevOps of building that into momentum. It takes 20 minutes yeah to do, and so you have to put in the work to set it up right and tune it, but then also consistently push adoption down to your folks. The other thing I'll say about picking vendors is push for pilots. Push for somebody to show you that this thing actually works. It's not always possible. And why not? Why isn't it possible? I mean, yeah, there's training. I'm But I'm with you. I I agree with you. Push for pilots. I don't mean to be a grouch, but if you can't get a pilot, this is just me. I'm being a little But I mean it. I'm not being I wouldn't buy the product. Yeah. There's exceptions. There's exceptions, but it says something if you can't get a pilot, isn't it? That it's hard to deploy. We didn't do a pilot with OneMind because there is so much in the implementation. Yeah, we've put a lot of effort into it, and so I understand why a pilot wouldn't have made sense for that cuz there's it's trying to do so much. It's plugging into all of these tools we have. But basically everything else I said this is great. I'll I'm happy to pilot, but I'm not making any commitments up front, and 90% of the time you'll get that back channel hard. Tell me about some customers and then go talk to them and get the real feedback. You got to be really choiceful because you waste a lot of time with the wrong tool.

Yep. All right. Let me I want to dig into one or two things for run of time. We talked about this one to two call closes. You thought it might be 10%, but now you're starting to agree with me. It might be 50%. As we go on, depends on the time horizon. It will be 50%, is that in 6 months, 12 months, 24 months? It it is an eventuality. Yeah. Let me ask you about this point. I actually didn't realize it on the slide. I think it's actually the third one. I'm color-blind in yellow or green or whatever it is. Maybe I said CRO and CRO, but I also mean RevOps. I think this year to be a RevOps professional imminently in a CRO by the end of this year you will have to know how to manage a 50/50 team, 50% AI, 50% human, not just tools, but you will have to learn how to manage agents, agents, AIs, whether that's training, whether that's whatever your team your extended team, this is PLG 2.0 on steroids, right? Your team may be half AI, half humans, or it doesn't matter if it's 20/80 or 30/70, right? Your team will as soon as it's double digits, yeah, you better know how to manage this hybrid team of humans and AIs. It sounded crazy a couple of months ago. I don't I think it's going to you're going to have to know this year. Will the CRO need to do that? I'm not sure how like how much of this can you have in your RevOps team? And it's okay, we have these three strategies. This is what's happening, and the agent management is going to happen somewhere. Yeah, this is why we're hiring this GTM AI lead cuz there's a bunch of things that I want to do that I'm tinkering with and we're doing in RevOps, but I want somebody to own this. And so is all of the agent management happen is can we put that in one function cuz they will do it so well that there will be so much more leverage. We're going to find I'm actively trying to answer this question. I mean, I hope so. I worry though is if the CRO or VP of sales job is to hire Janette that is my AI RevOps person. That to me feels like a CRO that's never done outbound hiring, Mr. Outbound. I've written about like the 100% failure rate. Yeah. New VP of sales, CRO comes in, only done inbound, only done whenever. I'm like, oh my god, I got to do outbound. I'm going to go find a director of outbound that then hires 11 SDRs off the street in one week, and it never works. If it's not your core competence, super hard to do it out of the gate. So if it's I hear you, you're the best of the best. But the next CO I interview, I'm gonna ask, how are you going to manage a team that's 25% AI and 75%, and if his answer is I'm going to go I'm going to find an AI guru. I think I'm going to tell a CO not to hire him. That's a fair point. Yeah. And maybe it's semantics. There's a bunch of my personal workflows that are all and we're saying anything AI is an agent. Now I I have a different definition of agent we don't need to get into, but yeah, a lot of my job is done through AI. So am I managing agents today? Sure, you could call it that, but I think the point remains a lot of your job is going to be through AI workflows and agentic services and managing all of these tools. A lot of my job is managing the

system and infrastructure of our business anyways. And so the same thing called by another name. I think the point is you need to be a highly systems engineering focused CRO to survive over the next two to five to 10 years. If you don't have the acumen to be able to understand how customer journey connects to systems and infrastructure and connects to the rep process and how this whole orchestration works, you will greatly struggle because you can't win on oh I'm a great sales guy and I'm a good manager and people like me that I'm a good manager, people like me. That version of a CRO will not survive in this era. You need to be able to think in systems.

Maybe maybe the polyiana version is may maybe I'm wrong. Maybe what happens is the another possible outcome in some cases is the CRO's role just shrinks. Okay. So imagine for your revenue in a year 40% sold closed by an AI salesperson and 60% is by the sales team. Maybe instead of the CR having to be an expert like we're like we want maybe the role just shrinks right and and certainly like in the B TOC side you've seen some of it at Shopify or at the low end of owner in some ways that happened because marketing does much more and more of if in the B TOC side if marketing can just run the campaign and close the deal there's no salesperson involved right so maybe they don't become dinosaurs it's just they become but even that that's a hopeful view right is that a dated skill set just gets shrunk but I think you're going to have to manage your AI team and your human team.

Yeah, I agree. Which is different than a tool. It's harder than a tool to manage an AI team and a human team, right? There will be a bifurcation of outcomes. I think if you don't become AI native really quickly, your role will shrink and then this will get done elsewhere. If you can become AI native, then your role as a CRO will contain that. I spend a lot of time on PLG. I spend a lot of time on on this whole thing. So, my role has grown. If anything, there's more things that's it's grown, right? But I'm super early on it and I'm making it a big focus.

Okay, last one. And we touched on all this other stuff. It was great. Th this is my view that people think is kooky. I don't think it's cookookie anymore. That and maybe you're even going to pilot this. Maybe you hit it with Amanda's product which is in the next X months. It's not everyone's got a freaking noteaker like Granola, Janola, Otter. We do these workshop Wednesdays this year. There's 200 noteakers. Hooray. I give up. Right. I don't view that as and they're sales focused notetakers and it helps. Right? I think a digital SE like this is Brian Halligan's digital one on Deli is going to come to the zooms. Yeah, why wouldn't it? The human it's person is going to be there and I brought this up and Darra said give me another release and this will be full duplex like my AI and Brian's can talk together but this person is going to be there should be there on all the Zooms at least as an SE.

Yeah. Like as good as you like you're going to have digital Adam the CEO digital digital Kyle should be here. that is you but better because you know everything in the product. This humanlike version should show up. I think this is soon and will change the whole way we do Zooms in sales. This is very much on my road map. And you're doing this right. Not a noteaker, not someone to send a chat, a literally a humanlike person that we're not pretending isn't an AI, but is better than a human as an SE, right? Yeah. And can be activated when called upon and also proactively jump in. What I would like is when a sales rep says something that's incorrect. The AI goes actually the integration isn't like this. It's like that. I That's what I can think is going to happen. But for what it's worth, no break. I don't think it's going to be proactive. I think this only works if they're on every call.

Yeah. Oh, they'll be on every call, but every call and they listen. And then you ask Kyle, actually, I'm confused how our POS integration works. And digital Brian jumps in. This is exactly how it works. And it's a it's a little digital, but it's better than a human, but not I don't think that rep will even get to choose. It will be on everyone like a notetaker. Absolutely. The will not be choosing it owner. It will be there. You're the best. We'll dig on this more in the coming months. Thanks, my friend. Thanks, everybody.