Transcript for "How Perplexity runs AI Sales tools at an AI company. Here's what actually works.": Okay. We are live normally. Hello, everyone. We should, be live. right now. That's cool. Hello. We see the messages. Normally, we shift it to live. Can we can you tell us if we are live? If you see us if you if you hear us? Yes. I see. Yes. That's cool. I think we are 158 people attending this webinar. I think we can wait a bit. We had much more, like, registrants, and we are going to start in a few minutes. Be sure. Maybe this looks like live. And I see and I and I see you know how to use the chats, Chris, and the and the emoji and the GIF popcorn. The GIF is very nice. Yeah. Like like this part. Cool. We are at 183 in the session. We will start in one minute. If it's if it's okay for you, Nate and Pierre. Yeah. Cool. Perfect. Alright. Let's fucking go, Nate. Who, is saying. this? I think they fucking go the the way we we talk with each other in the I'm just like, LFG. When I when I when I arrived at Temis, I was like, what does it mean? Like, let's fucking go. Yeah. Size the man. Shout out to, Finn in Intercom over there. That's cool. Okay. I think we can start. We are at 218 people. I think people are, like, going to join minutes after minutes, so we we can start. So hello, everyone. Super happy to to be here today. Super happy to welcome everyone. When I say everyone, I'm obviously talking about Pierre and Nate, the two speakers that are with me today, I'm but also talking about the people joining. I said 200 and and 200 people. I think we are, like, 229 now, so we have, like, more and more people joining. To give you a bit of context about, like, the why behind this this webinar today, I think you have seen around a lot of discussion about AI AI in GTM. You have seen on LinkedIn, everyone commenting, like, perplexity AI to receive all these these playbooks telling you how to use AI in your in your day to day activities. I myself ask I'm asking myself the the the same question. How much should I use AI? How much should I leverage AI in my sales rep day to day? How much AI should I provide me? Pierre is asking himself the question also, like, how which agents to build, which agent makes sense, which pay address with AI. That's why we wanted to ask an AI company and the head of the team at an AI company. And we went to Nate and asked him, like, if he if he could kindly join the webinar with us. And I think he said yes because we have him with us today. So thank you, Nate, joining this this webinar today. Yeah. Remy, thanks for for hosting and having me. Excited to join. I'm a fan. That's cool. If I talk about the agenda, if I share with you what we will cover, I think there are three main parts. The first one is the shift. We we we we can we can all agree on the fact that that GTM just changed. Nate, well, that's ramped. He's now at Perplexity, so we'll see how much shifted and ask him a couple of questions. Then we want to be really hands on and show you, like, the live workflows, so the agency builds, what for, how the the reps are evolving with AI these days. And I think and Pierre will also participate with its cofounder at at CLAP. And the third part would be, like, about what AI should do but also should not do and what everyone should hold back talking talking about AI these days. So that's the agenda of today. At the end of the session, we really, really be glad if you go out and tell yourself, okay. I understood where AI can actually create a leverage, a difference in my system. That's the objective of the decision. The timing, we have the next forty five minutes together. Thirty five minutes of which is this agenda. So going through all these points I I just presented, and ten minutes is question and answer. So let's see. I think you understood the chats on the right part of the screen. Everyone is is asking, like, commenting and and using the g, the emoji. So I think I think so you know how to use it. The questions, there is a q and a part. So do not state you use this q and a part to ask questions because there, you can also upvote the question that's are of interest to you, and we can see them. By seeing them, I mean, like, we, Nate, Pia, and me, we can see them and answer them at the end of the session. But also the two moderators we have on the backstage, you cannot see them, but I see them, Dominica and Julia. They will be able from from Lemlist to answer them if they can. For example, like, GTM is short for go to go to market. Up. Okay. Cool. I think we can we can start. Everyone is ready? Up. Yeah. So first part, I think, Nate, it's going back to to you. The the question so I know you you have been at Ramp for three years and a half. You have built the GTM organization there. You have we have seen a phenomenal ramp up for this company. You have moved couple of months ago at Perplexity. From your perspective, from Ramp to Perplexity, what's the shift you observed in in GTM these days? That's that's my first question for you. You're on mute, Nate. Awesome. And you're back with, us. first of thanks for having me. Really excited about the agenda and being able to chat with with both yourself and Pierre. Yeah. I mean, the last, honestly, the last, like, two, three months have been wild. It's been almost as big of a shift I felt since, you know, the last year or so, maybe even more, just with the launch of these fully agentic. Let's call them, like, digital coworkers, essentially. So at Ramp, Ramp was great. They're very AI focused and still are, which I think is what sets them apart in the market and why they've grown so quickly. When I joined there, you know, it's core kind of blocking and tackling from the system's perspective, a pretty standard, tech stack. We had Austin Hughes who who built out a lot of their AI SDR before that was a real common thing, and now he's over at Unifi doing cold outbound there. And so it was great to work with that team, very future focused. We both had live SDRs and AI SDRs and would constantly just track feedback there. So, and and it was really interesting learnings. We kinda had two separate teams. And my role in the kinda surface area was a lot of the the tech stack for the sellers. So CRM management, managed a team of a handful of folks that that would, own different parts of the funnel. And so great place to learn and work. And in the last year or so there, I found myself using Perplexity almost every day. I was probably one of the highest highest users, of Perplexity over there. And and, yeah, the shift here at Perplexity is just the surface area, and workers that I deal with are now digital. So I'm I'm working with essentially to do a lot of the work that I had to do manually or my team would be responsible for. And, yeah, the role is a little bit more operational at Perplexity. It's it's go to market, a good amount of enablement, good amount of just kinda core rev ops, comp planning, things like that, as well as outbound execution and inbound execution, as well as, like, a Salesforce admin, which I think there's kind of, like, joke that Ethropic was hiring a a Salesforce admin. It's still real work, and it still has to get done. So, so the integrations with Recado, Hightouch, etcetera, it's both systems and operational, which is why I was so excited to join. I think I think maybe I can ask the question to you as well, Pierre. Think it was not part of the plan, but you have seen the GTM evolving the last months, the last years. How much AI is affecting the evolution of GTM these days? GTM for go to market, obviously. I think, like, for me, the biggest shift is really about to what AI can do now compared to what it was able to do, like, maybe a year or, like, two years ago. I think a year ago, we had, like, a lot of, like, limitations in term of performance, but also, like, interaction with, like, different tools. And and, you know, we are, like, big believers in, like, MCP's at CLAP. But the fact that you have, like, really agents being able to interact very quickly with, like, different tools. For me, this is really, like, the big, big, big shift that I'm seeing, and that is really unlocking, like, tons of crazy use cases. And that's the stuff I'm the most excited about, honestly. Think every. day is a is a new shift, Nate. Yeah. Fun anecdote. I think back in 2023, I had set up, like, a first ChatGPT Slack channel at Ramp, and I just connected a a simple Zapier over to my own API credentials for ChatGPT, and it was, like, mind blowing. Hey. We can have this shared channel. We could ask it to write an email for us, and it would write an email. And he had to really prompt it well, and it was just cool because we people adding. Eventually, I had to, like, shut down that channel because it was I was paying for it out of pocket. But that was kind of the first. And now that the leap that we've taken now to what, you know, Perplexity Computer can do end to end next few across systems has been just a a leap that, you know, that felt like a huge leap at the time, having a full AI assistant in Slack to now this last week, we released a Perplexity Computer directly in Slack, which is which has been kind of wild. Maybe one question, Nathan, because I'm seeing that Robert has a question about the size of the team, but, like, how big is the sales team at that place cities? And also in term of, like, shifts, do you also see a a big impact in term of, like, quota that you need to basically manage as a as a sales rep. Yes. I mean so so my kind of selfish goal is to have the most efficient sales team on the planet, and I I don't think it's it's out of reach. We have the tools at our disposal to do that. And the sales team right now, we're we're growing really quickly. So if you know anybody, great. Or if you're looking, please reach out. We have I think we're go to market team is still under under 20 directly, and and we really bring in a lot of product folks and a lot of engineers to certain customers. But it's really a testament to how simple our onboarding and tool is. So the cost or the time to value using Perplexity is just very quick. So so there's usually a handful of onboarding calls and support resources, and we do develop kind of ways to adopt within companies that can be self serve. That said, we have we have a lot of interest, and we have a lot of larger orgs. We are looking for for more of enterprise growth team, presales and post sales, as well as, we're testing out more human led SDR SDR motions as well. And so so yeah. So it's it's a much smaller size ramp. I think we had over a 100 SDRs and a 100 AEs at the time of of leaving, and the shift there is really as you think about systems and process, when you add a lot of people, it becomes not a nightmare for systems, but a challenge for systems. Because a lot of the work that you do to try to help the customer or make it the most efficient for the sales rep actually becomes a hindrance because you have to keep everything equal, and a lot of the work goes into actually rules and engagement and making sure that that everybody's treated fairly and you can scale the sales org and everybody everybody will have enough accounts, and and there won't be overlap in rules and engagement or channel conflict. So so that's one kind of nice thing about having a small efficient org is you don't run it bump into those problems quite as much. That said, there's, you know, there's only so many hours in the day and hours that we wanna be talking to our customers and prospects. I think that's pretty interesting. I think we we will cover these these questions. This burning question everyone is asking is, like, how much how we should we could feel looking at AI as a sales rep. Should we feel that we are going to be replaced at some point? Should we feel that we're going to be assisted? Maybe I can ask you the question now. Like, if I'm a sales rep, so talking to you now, how should I feel about AI? How should I feel moving forward with my current position? I I would be most excited, honestly, of, like, all the roles. I I don't have data to back it up, but I think sales sales folks, go to market folks that adopt AI are in a great place, especially today where, you know, in past worlds, you might have secretary or assistant or offshore team to do a lot of work. Now you really have a digital team if you take advantage of it. And I think even last year, you had to be pretty technical. The tools only get you 80% of the way there. And if you're good at your job, you could figure out a way to do it faster yourself. And and all the time spent on on trying to train AI was actually just time wasted. Now I think, actually, you can train agents very quickly and really have, like, a digital coworker to do some of the core asks that that aren't vital to building relationships with with customers and prospects. So we kind of all became manager in some points. We all have people, like not people anymore, but not headcount anymore, but, like, maybe assistance, AI assistant to manage and to level up and to be, able to accelerate. yeah, that that's the future that I'm seeing is just keeping on these agents, making sure they get stronger and stronger and better. And and instead of not instead of people necessarily, but people are there to to be accountable for a surface area of the company. So if you're a sales rep, you're accountable for making sure your customers are happy, and the people that are interested have all the the tools, you know, to to understand if they're a good fit for your your product. So you have to maintain that relationship, and I think the surface area of owning multiple things actually becomes much more reasonable with one person managing an agent team. Okay. And and maybe and then just to make sure I understand on my side, so the team is, like, 20 sales rep. Is it, like, full stack account executive, or do have, like, a mix of SDR and AE? And when you say, like, the AE is managing the team of AI agents, like, let's say, I'm joining tomorrow at Parplex City. Like, what's the process? Like, how many agents should I handle? Like, how does that work? Like, what's the playbook to kind of learn how to work with an agents? Like, how does that work concretely? Yep. So our yeah. So the, the fun part is we're evolving about every single week. So if you join tomorrow, which would be awesome, we'd love to have you have you join the team. I understand that that, you've got a pretty good gig going today. So but, it would be really a little bit of onboarding by fire, essentially. We have a very open model of shared channels. We have internal channels that we add our our Perplexity computer to, and we share new skills probably every every day or so. We're we're sharing new skills and involving new agents that people have found helpful. So I think we are still kind of start up in learning mode and and deliberately so because the space is evolving so quickly and customer needs are evolving so quickly. We want to be able to have folks at Perplexity that can also adapt and learn from the market. So with that, to answer your question, breakdown is we're actually just testing the the SDR work. Our, our growth team is responsible for both hunting, finding new deals, as well as, going through the full sales cycle for enterprise customers. And so a lot of our smaller, smaller enterprise orgs are self serve. So the nice part is you could sign up today for for an enterprise company. Lot of, sub 100 person companies or even one person start ups are getting a lot of value from just clicking and going and and starting to use, the Perplexity Enterprise Suite. And then we are starting to, hire BDRs just to to maintain that outbound flow. Since we've had a lot of inbound in the past, the shift is now how do we actually, generate more more outbound deals and go outbound a little bit more? And so we're we're testing SDRs. But right now, our team is actually doing a good amount of outbounding with with our tooling as well as closing deals. That's cool. I told you before I will be the timekeeper. I will be I will make sure that we we go from one part to another. For the one for the people joining, just to to you to have a a closer look to the agenda of today, we discussed the shift in the first parts and the fact that GTM just changed with AI. And so, like, also the feedback from Nate and and Pierre have the experience it in the day to day activity. Now we can move to this second part, which is, like, the the live workflows. And I think we can also, like, organize, and that's what we said we told we said, Nate and and Pierre, we said we can start with the seller and maybe the the meeting prep. So I stopped sharing my screen. Back to you guys, going to the agent, going to what you actually use. Can you show us maybe the behind the scene and what the sales rep can actually leverage in order to do his day to day activity? Yep. Yeah. Absolutely. I actually ran just a quick summary to to get a better understanding of of kind of the core things that that we use. And the main I'd say one of the the main ones that we've seen across the board is is meeting prep. And, one thing I've noticed at different companies is we might have engineers or or a centralized process for meeting prep delivery to different to different reps. And as the team grows, everybody has their own way they want to do it. So we would see, okay. You have you know, over in Gong, they have a meeting prep thing. And over in this tool, you have a meeting prep. And then we built this Workato flow that will send you a meeting prep, and it just didn't get widely adapted. People have their own way of doing it. They have their own things that they'd like to to see in a quick way. So with our skills, I talked to a few different sellers. They all have their own separate version, and we kinda have a centralized skill for meeting prep as well. And so I'll kinda show you what that that looks like in perplexity. I think you're going to answer, like, a question I'm asking myself all the time is, I'm as a sales rep, how can I use AI? How can I leverage AI? How can I make AI The. the agents capable of amplifying me and enabling me to eat my my quota faster? So that's meeting prep is one of the the pain point I have every day, so I'm super happy to see it. Yep. I'll kick off a flow here with our t our screen. Cool. We told you this was going to be a live interaction, so it's, like, super live. Like, Nate is currently, Yeah. like, showing its full screen. Let me go to speakers here. I'll I'll just take a screenshot of who's on stage here. Sometimes I just, honestly, just screenshot a meeting, and I'll say, I'm on the list pod Pod episodes drop on my calendar. So I I really like dropping just a placeholder on the calendar with the details. That way, if I'm jumping into a meeting, I can see this kind of blank slate. Some some people like like that. In the skills section, most will save, like, an actual full skill for for meeting prep. So you can have you know, if this goes well, I'll run that skill here. And I'm just sending this Slack DM. Sorry. What do you mean by what do you mean by, saving a a full skill for your meeting prep? Yes. So this would be a skill would just be, hey. Run my meeting prep skill for today. Okay. I'll run a an example here and show you kinda how computer thinks, and and I I can come back to this on mobile. I can, you know, reference this and see if if something's going to take a while, jump back and and, be offline, and this will continue to run. In the connector section, it kinda show what's what's connected and available. It can reference my calendar. If I have questions about the Teams calendar, it connects to Salesforce. So I can say, let's look at all the meetings that are on my Teams, the go to market Teams calendar for this week, and then, you know, send me a Slack overview, give me a slide deck, etcetera. And so, yeah, over 400 here. I believe Lemlist is on there as well. So Jumping back over. Let's see. Oops. what you're basically saying is, like, me as a sales rep, I can connect from the morning. I just I can ask for complexity to run a check on all the the the slots I have in my calendar for a day and to and to prep the meeting for me. That's what I can. do. Okay. The the the question I have for you, Nate, is more about the skill management part. Is it, like, the kind of collaborative skills, meaning, like, it's accessible for everyone in the sales team. And, basically, you are a bit in charge of managing all of that as a good market leader, or do you have, like, a unique skills for each sales rep? Yep. Great question. So both is is the short answer. So we have org skills. We also have some example skills. So sales call prep. Probably should've just jumped to this one right away, but, this is this is a great one. It has kinda defaults default skill. You can try it directly in the chat as well as as well as you can just modify these these skill files and upload. So I've uploaded a few. We use a few different tools that I want to always call. So if I ask for technographics, it knows which tool to go reference for technographics. If I ask for ROI calculator, I believe I have a skill that that will give a good ROI calculator. That was actually one that I wanted to show, that I did with Lemlist, and I'll pull that example out as well. But yeah, so we'll go through. Oh, I've actually also asked. I'm gonna do this podcast today, draft me a a sequence. We could do this with Lemlist. Apollo is our example. It will auto draft, and then we can actually ask it to connect and and push this sequence directly into into Apollo. And then if I had the attendees here, we could just add them directly to that to that sequence and fire it off right away. So these are kind of the end to end workflows that it can do. And let's see. I haven't let's see. DM on Slack. And I will pull up, actually, for for sales reps, we have one other that that I thought was really helpful for for proving out ROI. Let me pull up our doc here. one you the one you presented. Yeah. So I understand, like, meeting prep, launch sequence automatically. That that's what you said, and they are right. That the three ones you see and you use among your teams. Do you? Yep. Yeah. Meeting prep, everybody has their their skill example for meeting prep, and these are also repeatable. So I I run these every morning, and people run them every you know, at a relative frequency there. So you can say run at 8AM. I also have one just for learning new skills. So I say scan, LinkedIn every day for the the most recent kinda go to market, AI tools and and give me an update in Slack of what people are saying. So I have, let's see, a link to the ROI calculator that I wanted to to showcase here. Here's a meeting, like, I guess, a quick overview of that. Lost my tab here for the LEM list. ROI. Coming back. Yep. Let me change. I think the the the meeting prep at the at the big one, the one you you presented the first, like I'm I'm I'm speaking like the in the meantime, when you you're showing ROI. I think it just myself, when you have, like like, five meetings in the same day, just be be capable of just knowing who you talk to, the context, and be capable of jumping in a meeting from back to back meeting ready to just have a context and and and have the size breaker ready for the beginning of the call is really a game changer, I think. Yep. Agreed. That one, and then even just pulling for ROI calculator. Sorry if they delay here. Yeah. Here we go. So once you have that kinda meeting prep, we can take it a step farther. So in a skill, if it's a follow-up meeting, for example, I can go to this this thread that I that kicked off on our last call, but basically said, build an ROI calculator. Gonna meet with Lemlist, and would like to showcase just a a full application that's usable. So what it did was it it referenced your products, your pricing, current tech stack, and kind of what customers might be using in their their existing tech stack, and built full end to end configurable ROI calculator that you could share with a customer. So half a million ROI, that's awesome. Different customers kinda have different makeup, so it could actually do the research on their company, what they're using, learn from transcripts through through Clap or Gong or Apollo, and say, alright. I'm gonna pull the transcripts. I'm gonna really understand this customer, then I'm gonna build an ROI calculator, but based on what my company is offering. And this is something that they can share internally, so you can make this private. Anyone with the link, which is more public, or anybody at your organization. So the ability to quickly spin up web apps for sales, I think, is a total game changer, and it really helps in a really fast changing environment. And the fact that this could be built on top of just a normal check my daily meetings flow as well as slide decks, which I'm sure that most folks have seen before have some skills behind. We we've gotten a lot of value from from quick ROI slide decks as well as end to end kind of full web apps. Maybe, And that's, one question for you because, like, the I I guess, since you are, like, selling that complexity, you live in that complexity, and for your sales rep, it's, like, pretty easy to have, like, the reflex to go on there. But for, like, many people that I meet every day, one of the big issue is really about adoption with, like, the AI agents. And I saw some question in the chat about how to kind of fully automate the entire process because you know that if you don't automate stuff like the meeting prep, you can schedule it. You know that's not going to be fully adapt by everyone in the team. Do you have, like, some advice on how to handle that in term of, I like, features or maybe even the training or triggering and so on? Do you have, like because I guess, that, like, city, it's a bit different. But if you had to deploy it for a ramp for 200 sales rep, like, how would you manage that? Yeah. Training enablement. It's I mean, I think everyone kinda feels that there's kind of a a general anxiety of how fast everything's moving. So I think it's kinda get on board, and and we're of that mindset of we're gonna have shared channels. People are gonna drop ideas. Ramp was very similar. We have, this we we had an AI channel where people are testing new things, sharing AI. Then there's gotta be some top down structure to it of like, hey. This is really useful. There's the bottoms up learning from from AI forward go to market folks and even folks in in finance or other other parts that have found something useful in AI. They're sharing kinda their their edge with the rest of the company. And then it's really up to kind of operations teams to operationalize that if if the tooling isn't there. And I I think I heard you say that I'm I'm kinda living in perplexity. That's been honestly the biggest hurdle for me here is is it's kind of like asking, hey. How do you use perplexity? And I don't know if I was sharing the the ROI calculator or not. Looks like I stopped sharing screen. But but, essentially, the the the answer is I use it for almost everything. It's kind of like asking how how do you use Google back when when Google was there. It's it's kinda like it's where I ask questions. It's how I I execute with the the 400 connectors. It can actually execute across things like Salesforce and other tools. So it's it's kinda where I get a lot of work done before before asking other teammates for help. I think we covered the, when did I hear the the web app at all, or was that did I stop screen sharing at that point? I think it's good. I think it's it's it's back to me in this screen. sharing. I just wanted to to focus on the on the second point we wanted to to cover together, maybe, like, moving from the IC, from the individual contributor, the sellers themselves, to the manager. What are the so we we put it on screen, but what are the agents that you give them in order to execute their job and to amplify themself? Yep. I can kinda take this one first. So someone. that I found really helpful is, just overview of what's happening that week. I think sales managers have one on ones. They have general meetings, and they sometimes check, you know, Salesforce forecasting. But in in order to insert themselves in deals, they they really like to know, one, what are the red flags on these deals? And I'd I'd love to hear from you, Pierre, kinda how you handle things like deal risks. I think transcripts are kind of the gold mine there that that everyone really should lean into depending on for for every deal risk, commercial, product, etcetera. And that can be the flywheel to make product better. It can be the flywheel to make your pricing plans a little bit more more fit for the people you're selling to. One really quick one is is just I have a skill that does a weekly, meeting summary. So I will, let's see, share, once again, share my screen here briefly. And this this is a skill that just goes every Monday and says, hey. Oops. Let's go here. Show me the the breakdown of of who we're meeting with and and the key opportunities and how much they're for. So this will send just a quick update on, you know, all of the the companies we're meeting with. This is kinda fake fake data for for this, but links out to Salesforce, links out to the the LinkedIn of the folks, has information on the deal size. So that way, the full team can say, hey. I actually worked with somebody over there. Do you want me to multithread and backchannel? Hey. I'd, like, I'd love to see my meetings on here. It's cool that they actually have visibility to the the broader leadership team, and how can we multithread? I like that comment. I do need a desktop cleanup. That's absolutely right. But I just not didn't make the top list of priorities there, but but yeah. But that, yeah, that's, with less. I'm hearing your thoughts. I think that the gold mine has always been, like, Google meetings have been somewhat inaccessible or they'll pop on last minute, but this this daily thread of just pulling out pulling out important meetings that are happening that day, that week, and then surfacing to Slack has been really helpful. Yeah. I I I have a bit the same on my side for the weekly sales meetings and so on. I think, like, the idea was to really say, okay, as a manager, maybe you are going to to do your weekly pipeline review. And instead of, like, going to maybe, like, Salesforce, HubSpot, check all the call recordings that you don't have time to. Basically, what I have is I have a weekly agents that's going to scan, like, all the open deals in HubSpot. Then, like, check all the call recordings, and the structure I have is basically, like, the deals to be checked and basically an update on the forecast. Then I also want to know a bit about the top objection that we've been facing and how well the team is ending that. Then I have, like, a bit of the some competitor signals to make sure that there's, like, not new stuff that we need to jump on or maybe, like, new battle cards, risks, and so on. And then the last piece is more about, like, some products, know, maybe, like, feature requests, some confusion, and so on. And, yeah, basically, that being sent, like, every Monday at 9AM. And this is pretty useful because as a manager, this is one of the feature that you already have. So you are, like, kind of used to that. But at the same time, you the feature is, like, super super time consuming, and I find that what you should need about bringing the data directly to Slack is very, very useful because you know that people are going to read the read the information. Awesome. Yeah. That's, and actually kind of similar. I have one other example that I spun up. By the way, Perplexity Computers only, broadly released for enterprise as of last week. So a lot of this is new, and I'd love, like, I love the comments and and learning learning from this, but the ability to to reference it and work out of it has been has been really cool. And the transcript stuff here, I'm learning new things I can do it, like, every single week. So I think it's such a cool place to be because that's where the actual customer relationship happens. Emails are usually concise if you're a good writer. They'll they don't have a a breadth of information, but the actual transcript is where where we get the most of our information. So I was gonna share just one more quickly use case that spun up this last week, and I I respun it up this morning just for dummy data. But here we can see it is with the voice of the customer dashboard, and it's pulling call transcripts from, you know, 80 conversations. And the left side for a manager to see, you know, what's happening in the field is is super useful. So you can see some customer love quotes by tag. At the top right here, I asked it to do full quotes versus just a summary. Like, hey. Just give us the overview of what they were saying because sometimes full transcripts can be a little bit verbose. Product feedback, key themes, competitive information, commercial feedback. So so this has been really helpful in just saying, like, I'm not gonna sit there and listen to all of our calls, not nor is necessarily a product team, but a quick hyperlink out to the LinkedIn of these people. The the actual call itself, wherever your call recording is done or the Salesforce account, has been super, super awesome for me to see and and as well as, like, pain points, use cases, give us new ideas on how to go to market and and pitch certain types of customers. This is one way that we've leveraged transcripts and and built kind of our own internal deck that that makes the most sense for how we wanna visualize the data. Oh, that's really cool. Yeah. We we have some, like, a a couple of application of how AI can help sellers, scale up managers. I think, yeah, that's something for us on the on the upsides for the slide deck creation. I think you wanted. to show. us something like on the sports. Yeah. Basically, like, the backstory of that is that we started, like, automating our slide deck about, like, two months ago. And at first, it's exactly as you described, Nate. It was really some people like me or Yan, basically, creating a slide deck or LLM. And the result was, like, really like, it was really, like, just amazing in terms of, like, personalization. But I think, like, the key question we had is was really how to scale that for a team of, like, 35 people given that. you don't have, like, the same level of, like, AI adoption and the kind of geekiness. And so, basically, we came up with, like, two different solutions. So I prepared something to to show what we have done. But, basically, the I don't know if you see my screen. But, basically, what we have done is that we I've decided to create, like, two types of agents. So one which is, like, basically, fully automated, and this is something that we do for, basically, smaller deals. On our side is, like, between, like, five to 50 seats. And in that case, the workflow we are going to have, everything is going to happen directly from HubSpot. As a sales rep, it's going to be pretty easy. Basically, the only thing that you need to do is just to from HubSpot, you are going to click on this button, and and, automatically, we are going to trigger an agent that is connected to CLAP for the context and to Google Slide. And, basically, in one click, you are going to receive your Google Slide that then as a sales rep, you can easily edit and so on. So this part, I I think, like, the the bigger pro of this approach was really about the scalability because you don't you don't need to go to any LLM as a as a as a sales rep. The second stuff, I think, for us, it was also a way to kind of control the structure of the design and so on. So we had, like, our designer at Lemlist, basically, creating the template on the Google slide and so on. And then, like, the addition, you can directly do it in the on the Google slide. I think the main cons of this approach, and this is the reason why we only do it for, like, smaller deals, is that in term of, like, personalization, we are a bit limiting in term of, like, a structure, content, and so on. And this is really the reason why we decided to also, like, enable a second type of agent for enterprise deal that really enables, like, for hyper personalization of the sales deck. In that case, the workflow, basically, is going to be a bit different. So as a sales rep, basically, you are going to chat in your LLM. So for example, in the dark, like, city or, like, other solutions, you are going to connect with your contact using. So in our case, a cloud MCP. Then what we have and the way we handle the design is that we have basically two custom scales, one for the design system and one for the deck building. And then automatically, it's going to generate the deck in the LLM. And then what we have done so this part was to also solve for the publishing aspect because if you publish an artifact, maybe on the complexity or cloud and so on, you always have, like, an issue with the link. So what we have done, we have actually created a solution where you can basically host your artifact. And if you publish your artifact, basically, you are going to have something super clear to share with your with your customers. And this is a type of deck that we're able to really super personalized for enterprise for enterprise deals, basically. I think. yeah. I think it generates a lot of value. I think we we have this run on production at Lemnis and Clap, and I think people the salespeople we have at Lemnis and Clap, they they gave us, like, super nice feedbacks about this text tech creation. I think we have seen today for sellers, IC, managers, and ops a few examples of what they could how they could use AI. That's what we wanted to do today. We wanted to cover the the end to end GM system and where AI can fit in and where where it shouldn't. Of course, there is a lot to see to to to show. There is a lot of questions also coming. I think if you if you want, guys, we can we can go to the q and a. What do you think? Yeah. Yeah. Yeah. It's got. I've got one other real like, thirty second use case, for. it. Go for it. Yeah. More more operational that I thought was really cool this last week is, like, we get asked about what our lead flow journey or diagram is. So was like, let me just ask perplexity, and it's something that with comment browser and and access to to browsing, it just went through and took screenshots of, you know, of our inbound lead flow, and I'll share my example here. But end to end documentation of our journey, which is something that, you know, isn't a ton of work to do, but having this in a in a two minute output was really helpful of, like, which fields are required, what is the landing page, what are our pricing pages look like. Let me zoom out a little bit. And and being able to do this even weekly if we'd like to to see what's changed. We could do this for competitors to understand, like, what's different between us and and potential competitors, what we should be doing differently, and then the onboarding flow. So it's really helpful in documenting. I've done the same with the comment browser and Workato, which systems folks on this call, I'm sure, can empathize. Sometimes in RevOps, it looks like you have Zapier, you have Salesforce flows, you have Workato, you have High Touch jobs. I just opened up those tabs and said document what's going on to update our Salesforce fields, and then hyperlink out to these Workato flows or High Touch flows so I can better understand and then write them in natural language, what what each they're doing and if they're active or inactive. And it's a really good way to just understand what's going on. You can still build quickly, and it won't take hours and hours to try to dig through what's actually updating which field. That's a cool one as well. That's why I was asking the question if we had, like, one more to to to present before we jump with the the q and a parts. I think we we had a lot of questions on the q and a part, also on the odd bond parts, the the personalized outreach, how we select ICP, what what we do for it. I think can you, Nate, maybe, like, cover one or two, like, workflows or examples of how you use AI to personalize outreach to to go for enterprise accounts on on the outbound parts. Is there any anything you do with Replixity? Yeah. I was actually chatting with with Jared on my team about this or on the the go to market side, and he had some some really good workflows. He just, you know, asks about we we have these target accounts, and so we we can ask of my target accounts, you know, which ones have shown some intent signal, do deep research on their account, understand what their technographics are, and then draft up emails for Outbox. I think there's we could have a whole another podcast on, you know, email deliverability and things like that. But with our enterprise, you know, outbound clients that are assigned to to folks, we want that to really be personalized in both with LinkedIn and Gmail. So he has a draft and and drop in the outbox with subject line that has something to do with their their, you know, something personal. So if it's alma mater, if it's something they have in common, if it's a LinkedIn connection, a really personal short subject line that will get a good open rate. And then send and then reviews these in the outbox and and sends directly. So that way, that they'll auto get logged to Salesforce, but this is just set up and running for him. And a lot of the outreach, you can fully automate, but I think there's I would say a good part of the job is being able to do ad hoc requests. Like, this is one example of, you know, an an awesome webinar that that will probably get some new new inbound interest related to. And so want to be able to quickly draft a policy sequence, create a campaign for the the people that have attended, and drop their Salesforce accounts in the campaigns. That that'll be one example more operational. But for sales rep and outbounding, we get examples like this too. Say, hey. Here's a list of people that made a Forbes list, for instance. Let's use Comet to scrape that list. If you can't, like, copy and paste into a spreadsheet, then Comet will just do that structured data for you. Then we could do the enrichment with computer and then drop the ICP folks into an email sequence. So instead of that being, hey. Could somebody offshore do this for us, or could an operations person do this for us? The sales team is actually very, very self sufficient in those type of tasks. That's yeah. And then we have another question. Like, I'm I'm reading the question as we speak, and it's in order to see, like, who is asking what. We have Patrick asking, like, where are we seeing the biggest gaps from human interaction versus AI? How do salespeople stay relevant? I think that's that's one of the question we want to cover. That's one of the questions we asked a bit before. But what can we say about human in the loop? How humans stay important? What is your take on this, Nate? Yeah. I mean, I think it's relationships. You you don't buy I I still am not a huge fan of talking with a a a chatbot, especially if I'm making a pretty important decision for my company or buying software, I'd like to have that kind of accountable surface area. Kind of like when you decide to hire an employee, you really have to have clear picture on what they're responsible for and make it clear to them what they're responsible for. Similarly, if I'm taking on a new software, I would like somebody responsible for how that software performs versus just a full PLG. So if it's a it's a really big investment, I think certain software's PLG just works. And, ideally, I shouldn't have to have a touchpoint at these softwares. But, realistically, we live in real world with humans, and we have relationships. So if if we know there's somebody I can reach out to at that company and they actually are experts in the field, that's another thing that I've seen AI just it's mind blowing how fast people can become really great a sales folks with the help of AI to understand the landscape, understand pain points for customers, and actually help them go through procurement processes. So so I think AI sales, I think, will help get your foot in the door with conversations. And then actually being an expert and being a human and and understanding and empathizing how you can make the buyer, the best at their job as possible because, you know, they they also want the best things for their company. If you can partner with them and and be a person on the other side of that conversation, I think it's it's going to be around for a while, those relationships. That makes completely sense. I have a question, like, on on how much can you do with the transcripts with perplexity after a call. So imagine you have a call with someone. You have this relationship going on between, like, a a human and between two humans, let's say. Yeah. You The transcripts. How much can you do with this transcript after what did they I think this covers some use case, and I think Nate can answer this, but I think, Pierre, you can also also answer this with Clap. We'll start with Nate. Yeah. I mean, we're I'm always looking for new ideas. I I was a big fan of of Momentum. Early on, ramps, so we adapted that that tool for transcript analysis, for great things about CLAP. And so I think having those transcripts and having those accessible, you can kind of do whatever you want. I think one is personalized outreach, drafting the right follow-up, which, you know, I think a lot of AI tools do, but it is very important because the most time we can spend on on more proactive sales work and taking in time and meetings and then automate the rest is really important. But just cool things like, hey. I heard you love skiing. Let's ping the seller and say, hey. Why don't you send some ski socks to this person? You're go they're going skiing as a thank you and, looking forward to the the continued conversations about the partnership. So I think there's, like, kind of fun use cases too of, like, let's send some appreciation gift or some perplexity swag to this person because they they just had a newborn, and we have some really cool, newborn shirts. So there's there's, like, those examples that I think are kind of the the more personalized ones that that are kind of a small touch, but I think really help the the relationships and partnerships. And I think Pierre on your side as well, like, in in Nate's quoted you a bit Yeah. saying that you you do cool things with CLAP. What what can you say? Like, how how much do you use this transcript for? I think, like, I lost transcript because you can really be super creative with what you can do with that. And I think it's also a bit about the kind of the whole of the sales that is going to evolve. That when we start with with AI, we try to basically automate, like, tasks that we already we're doing, like, emails, updating the CRM, and so on. But at some point, I think, like, the keys more about if you want to win the deal, you need to create, like, a trust. You need to create, like, the most, like, personalized experience and so on. And, typically, like, the sales, like, something that works very well for that, but I think you can also be, like, super, super creative with, like, all the stuff you can do with the transcript to kind of, you know, increase the customization of the experience, whether it's about the crafting the best follow-up sequences with, like, club plus lemlist or even, like, creating, like, almost, like, personalized marketing collateral for them. We also had, like, some clients that you while using the transcript to understand the type of, like, help center articles to write, to send to clients, and so on. I think at the end of the day, the the questions now is more about it's not about what can you automate, but more what are, like, the stuff wish you could do that are, like, not possible before, but now you can do it at scale. And we increase the kind of the experience that you provide to to your prospect and and clients. Yep. Absolutely. And then and then one really tactical one on top of that, you mentioned the, like, updating Salesforce fields. With AI, it's actually very good at making these these unstructured data. I've noticed that if you have a large sales team, people have different ideas of what an SQL is. So it makes it a little bit tough to to have, really structured, pipeline and and forecasting if different reps are flagging different things as SQL, non SQL. So that's that's one example that's that's that's good. It's like, do we have band qualification here? Do we have SQL qualification based on this transcript in the past emails? That's another really tactical example to keep pipeline hygiene consistent across the board. I think, guys, we we are close to the end of the of the session. We are, like, five minutes past the announced timing. It was really cool to to have you together. Let me share my screen once again to thank everyone for joining the session, for covering what we what we said at the beginning, meaning, like, where AI where can we leverage AI into the organization from end to end, not only on outbound, but on anything a rep, a manager, or an ops can do in the in his daily activity. What we are going to do after the session is work on giveaways, like, just just give give you our slides, the agent we talk about, so see what we can we can share together with with Nate. Maybe, like, look at the ROI calculator. We we do have resources on Lendus and CLAP, and we do have also webinars coming more on the outbound parts. We couldn't cover all the topics today. Of course, we just have forty five minutes. So we do we do we'll deep dive, like, on on on on on the outbound parts with Roy in two weeks. Thank you, Nate and Pierre. Thank you everyone for for listening. I think I can wrap. up now. What, do you think? Timmy. Thanks. That's, cool. Ravi. Thanks, Pierre. Thank. you, everyone, and see you. Yep.