Deep Dive: AI-Powered Marketing to Get More Leads and Customers with HyperGrowth Partners

Please welcome to the stage Guillaume Cabane, co-founder and GP at Hypergrowth Partners. Hello. Thank you. Usually I would like run and jump up here, but I, I broke, I tore my ACL, so I'm gonna I'm not gonna jump on the stage today, so I'm just gonna limp like around like this. Thank you for coming in this early in London. So happy to be here. My second time, I was here last year, had a great, great event and it's looking to be the same today. So today we're gonna talk about how we get more customers for any of you who are interested in making more business. That's what we're gonna talk about. I'm Guillaume Cabane. Everybody calls me G, so feel free to call me G. And we changed the, a lot of the deck over the weekend. Jason wanted us to go from something more theoretical to something more practical. So we pulled up a lot of examples from the advisory clients that I have to show you how it's really done leveraging AI tools in today's marketing teams. There might be some holes, point them out. We're not afraid. So yeah, I'm the co-founder and JP at Hypergrowth Partners. We are a collective that does advisory for high-growth companies. Previously, I built the growth teams over at Drift, Segment, and Gorgeous, and I've advised Ramp, Reddit, and many other companies. When I say an advisory collective, it means we are somewhat like a VC fund, but we don't give out capital. We invest our time, um, in exchange for common stock, uh, equity. Um, and we have about 30 senior leaders around the world, um, working, uh, together. Now that that's said, um, today we're gonna talk about a couple of things, but mostly around how do we as marketing leaders take in, uh, the AI platforms, the AI tools we have today to be more efficient. And a couple of wins that I've seen, uh, in a couple of channels. So I'm gonna cover outbound, SEO, and paid. But first, um, I wanna have a, a show of hands, uh, today to wake everyone up. Um, can you all tell me if your company has reduced headcount, uh, or decided not to hire because of, or thanks to AI recently? Can we have a show of hands? Okay, so I'd say we have like 10% of the people. Okay, I'll ask the same question one year from now. My bet is that we're gonna see at least 30% of the people where you have a reduced headcount or less hires, and I'm gonna show you why today. Okay, so if you, uh, you know, think about it, um, the real impact is that, you know, a year or two years ago, we were in, well, AI is good, but it's like less quality than humans. It's less for less, less cost, right? Less quality for less cost. And nowadays what I'm gonna show you is we have more, more quality, better quality and speed, but for less cost, right? So what does it mean? What if you look into the, into the cost angle first, right? Typically what you see is you have, different cost of acquisitions promotion, right? You have, you know, on the far right here you have like sales-led, which we used to, you know, which is fairly efficient in terms of closing but has a fairly high cost, right? And on the left side you have PLG, product-led growth, right? Which is fairly cheap, which is why when we were in the economic crisis of, you know, post-COVID, all the VCs were asking founders to do PLG, because they're like, free acquisition? Sure, that sounds great, right? What they don't see, or what they don't think, is that PLG generally drives, sure, lower CAC, but lower contract value, smaller customers, all right? And so you have kind of like this, this spectrum here from, let's say, low contract value, low ACV, to high ACV, which is correlated with low CAC, to high CAC. Right? And that's a pretty typical distribution, you know, like a, a 10 to 1. And with, let's say, those CACs, those ACVs, you have that typical like segmentation from micro to like enterprise. Okay? And with those, with that segmentation, you have the channels. So on the, let's say, cheap side, low CAC side, you can have like the SEO and the virality, all the PLG playbooks. In the middle you find like the outbound, right? And on the far end you can find the ABM and the, and the partnerships, right? But that, that's yesterday. The point of that strategy is what we used to do for the past 10 years, right? Because if you follow like the logical conclusion of the AI shift, like what's, what's changing, right? The fact that we're now capable of doing faster, and better quality. Example, we can scrape websites, we can scrape random content for extremely cheap and feed that back into AI prompts for outbound or for SEO. Drives an extremely low CAC on those channels, which means what used to be mid-market channels can now also be SMB channels because the CAC is dropping. You can have new audiences. If you look at this distribution, you can now do things like this, right? You can start saying, hey, I'm gonna acquire startups through outbound. I can acquire mid-market through SEO. How does that work? Well, it's easy. Like mid-market prospects with SEO is possible because you have better quality. By the diversity of content, you can create extremely high volume of high-quality content. And on the outbound, it's kind of the same thing, right? Because you're no longer employing SDRs, which have a high cost. You can, at the same quality, you can get lower contract value customers. So that's probably like the, the key learning we're gonna like hover around today is that the CAC to channel fit is changing and you gotta adapt to this new reality. Okay, so let's look at the, at the first one, the first channel, like the biggie. Outbound. Okay. If you think back to like 2015, like what were the best SDRs doing? Like I always learn from humans, you know, my wife is a sales, used to be a salesperson. So I learned a lot from her. Well, they do information gathering and research, email content composing and outreach on LinkedIn and whatnot. And they're being persistent. Okay. This works 2015, 2018. It's slow and it's high cost. You look in the middle, what happened is that everybody started automating those sequences on like Apollo and Apple Markets. We started having like huge volumes but low quality. And with that, well, obviously the conversion rate dropped. What's happening since last year is that we're capable of doing, as I said, scaled scraping data acquisition and generating high value emails. And that is, and I'm seeing it, it's driving the conversion rate back up at a high volume, but the CAC is even lower. So we can have better performance than humans at a lower cost. This is the shift that we're seeing, right? And if you look at something like this, like all those data sources, there are a ton of sources we can get great data from. But let's, let's be honest here, like which SDR is going to like do the job of going through at least one of the logos per data source to understand that prospect, that customer, and compose a great sequence. Who's gonna do that? No one. They don't have the time. They've got minutes per email. A good AI can do that. All right? And I'm not here selling anyone AI. I'm just saying like, I'm gonna show you how it's done. So let's jump into an example that I worked on recently, and this is my latest advisory client, Reddit. Fairly well-known social platform, and they gave me a goal. Get 10% of the US advertisers on to advertise on Reddit by the end of the year, starting from pretty much zero, which is a pretty big goal, okay? The first idea I came up with is, well, let's scrape the Facebook and, you know, LinkedIn and Google ad libraries, which are like public repos. We can find all of the ads, understand the ad, to make sense of it, understand the company, aggregate third-party data, and inject that into an email? Well, for that example, we went with, uh, Clay, uh, just to prove out the feasibility of the concept, right? It's a, it's a good, uh, CDP and workflow platform to start. Not that I would recommend building something at scale there, but to explore, it's great. What you can see here is we have a prompt. We're trying to figure out the, Facebook page so that we can find the ads. And then we're querying an endpoint to find the ad data. What you're not seeing behind the scenes is that we're making sense of the ad and injecting that into an email. This is an example, all right? And so here it says like, hey, I saw your ad on Facebook. Here's a summary of the ad. We even inject the ad image that we saw. And then because we made sense of the ad and we made sense of the company, we can start recommending the right subreddits where they should advertise. This email is super personalized. It makes a ton of sense, and we build that entirely automatically, and it works for all the US advertisers, which is the entire market. Once it's done, scaling that is extremely cheap. It worked great. We have done this in less than 2 months, And as of yesterday, I looked up the numbers, we've beaten the SDR team that's been working on that for like a couple years and there's 200 people. So yeah, you can, you can like, you can question the future of the SDR role when you see things like that, right? And when I'm talking about the headcount, I saw 2 days ago, 3 days ago, I saw this, this post from Maxime and he's saying, hey, we're having an outbound AI growth lead. Somebody who does outbound, who's good at growth, sales ops, tech, marketing, copywriting. Now that's interesting. Means founders are reacting and are thinking instead of having a sales team, SDR team, I'm just gonna have like one super skilled person who's gonna run the workflows. Okay, so what does it mean for like this, like 10-year, 20-year, like sales marketing battle? It means, Marketing has won. That's what it means, right? It means that outbound is now owned by growth and marketing teams. Okay? Just embrace it like that. That war is over. Okay? Because we can, with those platforms, scaled scraping, TAM building, intent signals. Like the key learning out of this is sequences are dead. If you think that the best way to engage customers is to have a series of pre-written, templated emails that gonna fire until somebody responses, like you are in like the 2015, 2018 era. You're like two, like eras, like removed from reality. Okay? Every new email is built purposefully based on the situation, the data we aggregate, and the existing conversation automatically. So that was for outbound and now SEO. So because we have some of the best people at Hypergrowth, I've pulled actually advice from Kevin Indig, who was the worldwide head of SEO at Shopify and Atlassian, and Marcel, who was the CMO at Deepgram. And 'cause they're better than me on SEO. And the summary is we are coming off of a world, a place of pretty bad SEO, to be honest. All right. We all, like, I am the main culprit here for sure, but like we all like outsourced the creation of like a lot of SEO content to fairly, you know, low quality freelancer writers to get, you know, on the search engine. And which was mostly like either like low cost generation, low value content or tokenized and spinning content, which, also produces low value content. Okay? So no one wants to read that. It kind of worked until it didn't. The playbooks that do work, the winning strategies in, you know, the previous era was like what we saw doing from like Zapier, G2. If you think about it, like, um, Notion and G2, they have this amazing, you know, product-led growth and user-generated content, UGC, in B2B, right? Which is very rare. It's very hard to pull off. Like G2 has this, treasure trove of like SaaS reviews that no one else has.

Okay?

Or Zapier hired a team to build all of those like integration catalog. Okay? That's both of cases can be done, but like they're very expensive. They're very expensive. Okay? Now we can use LLMs for programmatic page creation. Um, and we can get the data without the workload. Uh, which in a summary can be like, not everybody can have a strong UGC motion, but everybody can be Zapier nowadays. All right. So let's look at how it's possible. In this case, uh, this is for Deepgram. Um, and so what they did is that they did some very typical SEO keyword research, like find the keywords that they think their audience is searching, the one they wanna rank for. Okay. Fairly easy. Then they're using Airtable to store that like in a, in a observable database, a command center, just like we were using Clay in the previous outbound example. And so now they're like, they have all these keywords, they have all of this data, okay? And they're now using AirOps, which is an LLM orchestration platform, which makes it possible to have different sub-agents go look for what is the best, uh, title, what's the best, uh, page, where is the best screenshot, what are the, uh, examples we have different. And then they bring it back together, uh, into one, uh, prompt with the data injection. That is a much better way to create content than have a huge prompt that tries to do everything. 'Cause it tends to fail when you do that. And so we try to just break it down. We store the data in Airtable and then we feed back the data from Airtable into the final into the final agents. Okay? Um, it's a fairly advanced workflow, but check out the results, right? The prior stage, like up to 2023, was like trying to do it manually with like low budgets. And this is when they're like, well, let's just like automate this. Okay? Let's just automate this. They got 3.5 million clicks. Like, how insane is that? So what are we seeing, seeing here? We're seeing they have one person to build a workflow. And Marcel told me it took him like about 1,000 hours. Sure. Okay. And then one human who's there ongoing for like control and validation. Okay. It replaces a team of 10 people. That's what it does. Okay. For now, it's possible to use LLMs to get organic traffic from SEO. Is it gonna unmask? I don't know. We'll find out. But there is a time window to exploit this right now. Scraping websites, building unique content, enriching it with third-party sources and doing the quality control. Okay? Critically, last line here on the slide, the workflows with AIs currently produce better content than autonomous agents and better content than humans alone, unless you're ready to pay like $10,000 per post, okay? So better strategy. Let's go into paid. So paid, we got the, oops, sorry, I skipped. Paid, we got the feedback from Rex. Rex is the world's best guy at paid. He handles all of paid at HubSpot worldwide. He's one of the very few people in the world that have spent over $100 million on Google Search alone over the past 5 years. That's very, very rare expertise. And so before, we used to have fairly, you know, fairly standard situation. You had either a paid team in-house or you had, you know, an agency or freelancers, okay? And the work that they would do is you would segment the ad campaigns by, you know, various customers, customer traits like genders and interests, geo-location and product, things like that. And then you break that down into the ad groups, the campaigns, the ad group, the keyword targeting, the ad copy that goes with it, and then you are fine-tuning the bids all the time. And Google tried to introduce AI support tools over the years, which didn't really work that much. The learning is, as of 6 months ago, that has changed and it is working absolutely. It's making the job seamless. Like, the LLMs now inside Google Ads are making targeting better. They're better at showing the right creative. They're better at showing it to the right audience, at setting the bids. They're going to get you more conversions than you could manually if you try to override the AI. What's working? The problems that are solved are audience targeting and bidding. What remains for us, uh, humans for now is the research part, as we saw, as well as the creative process. Um, but wait, the creative process, the teams are also reduced. Like we are seeing the first products where you can build ads, creative ads and design ads just based on your website. It will create the ads and the video ads for you at $0 cost because they're happy for you to spend they spend the money. Okay? So this is just the beginning. A practical example of this is, you know, the TL;DR is if you're still trying to fight AI by segmenting campaigns into a ton of different budgets, like, you know, Netherlands, Luxembourg, and Belgium, and setting like different budgets here and there, you're fighting the wrong war. Just stop. Like, let the ad network figure it out for you. No, let's say, let's say that you expand into, into Benelux, right? And, and you have those, those ad budgets. Maybe Google's AI would've preferred to spend 7, 18, and 5 on a particular day or every day. And in the tests that we've done, the more you handcuff the system and restrict the AI, the fewer optimizations it can make on your behalf and the worse off you will be. So what this means for paid teams, right? Um, is, is that you have your job now kind of similar to what we saw in SEO, smaller teams, more technical teams. Your job is to build the infrastructure to enable the AI platforms to learn, okay? The data quality is critical. It has to be able to learn the good and bad outcomes, okay? Um, and you need enough budgets to ensure the volume conversion again so it can learn. You gotta create the conditions for that to work. And I'd say these days you're seeing different, depending on the platforms, the conversions being between 25 to 50 a week so that this kicks in and you can start using things like Performance Max. So a couple of takeaways that I wanna share. I have a lot of takeaways. So across the board, generating leads comes with like, less humans and more AI models. Okay, as we discussed, like this is a final takeover of SDR into technical marketing roles. Okay, more than ever, this is a battle of CAC. Okay, the AI change shift is shifting the CAC of many channels. It's opening channels to us marketers that we did not consider for those audiences. Okay. Uh, if you invest in technical people who can build the workflows that we saw, they can beat at least on SEO outbound channels, um, the preexisting teams at their game. Okay. Um, so invest in, in those technologies, in AI workflows, uh, and you will thank me later. Uh, just two deep thoughts. Before I leave you and we can get into the Q&A. Seemingly unrelated. One is the tech cycle is accelerating. I've been doing this for now 20 years. I've been buying, I bought millions of dollars of SaaS over the past 20 years and I used to be able to make a good evaluation, a strong recommendation and say, hey, this recommendation that I've done based on all the software I've seen to do outbound or whatever is good for the next couple years. This is the winner. This has, this is the best technology. These days I make a recommendation, 3 months later it doesn't make sense anymore, right? Because there's this new company, this new model or usage of a model that completely changed the game. It is really hard for me and for all marketing leaders to build conviction, uh, on which technology to build on top of. Probably this means you need a composable tech stack. You cannot invest all of your marbles in one big platform. You do want to be able to like compose with different technologies, have an independent database, and so you can, uh, switch and replace parts, uh, of your workflow. Second thought, uh, I've just given you like 20 minutes on how, uh, AI is better than humans, uh, at many things. I believe that's true. I believe also we're gonna see something opposite. We're gonna see like the certified real human which can be the new like Parmigiano-Reggiano of like sales. Well like, hey, like this is a stamp, this is like actually a meat bag that's talking to you. And some people are gonna like that. Some people will pay more to be able to like engage with a real human. And that's a reality. And we probably want to think of how do we offer those two paths, right? Maybe there's a VIP treatment for the enterprise prospects. Speaking of which, are we seeing the same shift in closing on the AE space. Not right now. For now, most people prefer humans. However, I'm seeing the first steps, um, in AEs becoming the puppets of, uh, AI models. If you think of Gong, Gong records and understands the entire like, uh, deal situation and will recommend to the AE what to say, what not to say, which competitors to bring up and how to price. That is the AE being the puppet of an AI model. Okay. I'm also seeing some industries where the, where there's some models being built, some chatbots to do the closing built on the conviction that some buyers will prefer buying from an AI model. You think that's crazy? I think that's crazy, but there's a good example. Look at tech buyers, engineers. Do you think engineers prefer talking to a sales rep or to an AI model? I got the answer. All right. So that shift will, will happen and, um, it's gonna be gradual, but we will see part of the market going to real humans, part of the market going to like entire like AI models, like the entire cycle. And most of the market being like the start is AI and the end is, um, human closing. And that's it. Thank you very much. And I think we have ample time for questions. Thank you, human, taking the risk. Go for it.

Me, me back here. So, um, you mentioned that you want to now have a composable tech stack in SaaS and like SaaS obviously always tries to sell annual contracts or multi-year contract, which is the opposite of what you recommend. How do you think the accelerating change, uh, velocity is impacting customer lifetime value for, for SaaS? Yeah. So basically, especially in sales and MarTech.

Okay. So if I got this right, you're asking like if the tech cycle is accelerating and G's recommending a composable stack, are we gonna see more churn? And if I'm a SaaS vendor, like how do I react to that? Great question. Great question. Um, well, I mean, on the product side, seems obvious that you wanna be able to like try and be the best all the time. Obviously. Sure. Okay. Um, I think what I haven't covered today is there's a huge value on proposing something unique, different. If you saw the Reddit example, the Reddit example was about scraping the Facebook ads library with like, you know, tools here and there. Why do I do that? Because no one's selling that data. No one is selling like a structured database of all the Facebook ads in the US, which sounds crazy because the data's public and there's an API, but no one's doing it. Um, if you're capable of offering something that is unique and that produces unique outcomes through the LLMs, then that is truly valuable. You have this niche market all for yourself. So I think we're gonna see, instead of having like huge platforms that try to do everything, for example, like all of the YC companies that are doing like AI agents that does the entire SDR for you, like the output of that is like, eh, it's not great. It's okay, but it's not great. Okay. What we are seeing working, however, is like very specialized LLMs at like a small thing with unique data that produces great output. So I don't know, if I had to put my money somewhere, I'd put it on like specialized LLMs for now. Thank you.

Hey, uh, I have a question about, I, I think what we're talking about artificial intelligence, particularly for marketing and sales, is companies becoming much more capital efficient. Yes. You're also an investor.

Yes, I am. Are you, for better or worse? Yeah.

But are, are you reducing the check sizes? Are you telling VC companies like, I won't give you more than a million bucks because you should be able to go all the way to Series A with that?

Yeah, so Christophe Janssens is somewhere around here, so that's a better question for him. But yes, like the huge rounds we were seeing, which were driven by huge capital expenditures for headcount, especially in California, do they make sense? I guess it depends what you're building. If you're investing in engineers building, sure. If you're investing in like salespeople, marketing people, Like maybe not. It also means that the moat is becoming harder to build in marketing. And so yes, like I've reduced, there's two things. One, I've reduced the number of checks 'cause I have a hard time building conviction on what is going to be the successful technology platform 2 years from now. It's becoming extremely hard to read. It's always been a difficult job, it's harder. And second, yes, we're looking at burn rates and we are questioning Like, why are you burning so much money? Why are you not building the outbound automation infrastructure, the SEO infrastructure at a fraction of the cost that what is— yes, that pressure will start to disseminate and go around all the boards around the world for sure. 100%. So I agree with that.

Absolutely, absolutely amazing talk. Thank you. Um, the, the one thing that you need is data and you said you were scraping. So in the terms of service, for instance, of LinkedIn, it says you're not allowed. Yes. And if you try, your, your account gets banned within a day and many other websites. How are you circumventing this?

Yeah, so, uh, there was, uh, I believe, uh, in 2019, 2020, there was like the Ninth Circuit Court of California. There was a case that happened against LinkedIn. And the, I remember if I try to remember the, where the case settled is that whatever LinkedIn lets Google scrape for search purposes, it has to allow anyone else to scrape. Okay? Otherwise that is anti-competitive and that's illegal. Okay? You can't just decide that you will get— let Google do something and then you're not gonna let somebody else do it. Okay? Um, so whatever is public, you can scrape. Okay? So that covers a lot of the signals we want cuz a lot of the content of LinkedIn is public. Um, that's one. Second, the risky part of that is if you are just re— like tokenizing and just like printing back the LinkedIn content you saw. Now if you're using LLMs to make sense of what you saw, summarizing and making a content that's interesting with information you have, well, first it's really hard to prove where you got that from, right? And even if that's the case, just think about it. If you had a salesperson go on LinkedIn, look at somebody's profile and write an email, is that illegal? No. Why would that be illegal if a large language model does the same thing? And so that is gonna be a legal question that I think is gonna be fascinating is like if we have AI impersonate human beings to do the exact same job at a fraction of the cost, why is the outcome different? And my bet is it's not gonna be different. We, we, you, and also we can't, you can't say like who did it and for what purpose.

So great, great, great answer. Just some of the stuff you can only see once you're logged in. Yes. And, and then your account and the IP will be logged and so on.

So yeah, but for example, I'm allowed to download my data. And so if I download my data and I pass it to the LLM, then I'm allowed to do that. And so I think it's gonna be really hard eventually to prove, you know, whether, but I don't disagree with you. I think like LinkedIn, those, the bigger question is like, why is LinkedIn doing nothing? Like why is LinkedIn not selling the data or building LLM tools for salespeople? Because they are the ultimate source. You gotta wonder what they're doing, especially because they removed their intent data product 3 weeks ago, right? So you gotta wonder, is it 'cause they don't care or is it 'cause they're gonna announce something big? I think the latter.

You can pay for the API in order to scrape. So that's the model.

Thanks so much.

Hey, so yeah, I have a question I guess about roles and skills, which I think is interesting. I think there's like a key red thread through what you said of sort of like keep a paid person, but they need to become more technical. Technical. Keep an SEO person, but they need to become more technical. I'm interested in thinking how you think about first what skills you should be looking for in terms of hiring and training, but also what roles in a traditional marketing org do you see as like very defensible in this moment and which do you see as kind of—

Cool. Okay, thank you. So there's a couple of questions linked to that. One, which is like I heard a lot with what What are SDRs going to do? What are salespeople going to do, right? And well, a lot of the SDRs are gonna end up being AEs. Like SDR was always a training ground for AEs. And so the SDR role is going away. The people will become closers or will evolve into something else. What's that something else? Probably sales ops. They will handle the orchestration, the workflows. What we're doing right now, whether it's in sales, in, uh, marketing for SEO purposes, for outbound purposes, is what for now we're learning from the 10% best people and impersonating their behavior. We, we still need that 10%. If I don't know what makes a buyer tick, I can't write the prompt. So I need somebody that's creative, that understands reciprocity and all of the sales, the, all the, the sales things, um, to help me build those workflows. The same is true in content. Like, if I can't say— if I can't do the keyword research, if I— no, then I can't do half of the work. So we still need the expertise, we just don't need the execution anymore. That's the thing. The expertise is super important, which is hard because the expertise comes from having executed and failed. So eventually the people here who have the expertise, you're gonna be uniquely valuable because no one else is coming after you. Like that's, that's, it's fascinating when you think about it. No one else will be trained into spending money on Google Ads because no one's gonna have to do that anymore. And so you're like the last dinosaurs who have been able to spend money freely like crazy. Just think about like the people who come in 5 years and they're gonna ask you, what, you spent like $30,000 to figure out if it worked? Like how crazy is that? Yeah, you in Mad Men, right? That's gonna be true across the board. Like you, you hired random freelancers to write content on a keyword and figure out if it works. That sounds like archaic, right? And so we're gonna move to more technical people that are creative, have deep expertise in the channel for sure, uh, and can build the workflows to replicate what they wish the freelancers of that team would do. So that's, that's what I think is gonna happen.

Absolutely. Super interesting. Thank you.

Hi, great talk. Thank you. So for outbound, like your main channels these days are you've got email, you've got LinkedIn, and you've got phone. What do you think the future is of these kind of channels? How long will they remain viable channels? And do you see other channels that are emerging that people should be looking at as well from an outbound point of view?

So there's a— and I did a talk a couple of weeks ago. There was a question. Um, uh, the end on will people still believe what's in the email? Uh, is email a viable channel? And I think that's a tough one. I'm not too sure because right now if you think of like why email works, it's because if it— in some cases you know the person, in other cases you see value in the email or it elicits, you know, curiosity. Trust, reciprocity, like there's a reason, there's a forcing function forcing you to respond, okay? When you truly believe it's a human that has gone great lengths to engage with you, you feel compelled to respond, okay? And I always give up like my snail mail example about the grandma writing you a letter and you being curious and opening the handwritten letter. Now the question is, when that is truly done by an AI, and you no longer can tell the difference, will you A, distrust all form of email communication because this could be a bot, or will you B, look at the face value of the content and say, hey, this is actually helpful. And I think we're gonna see divide, a polarization where we see some people going one way and some people going the other way. Overall, email is at risk because we saw the charts with like the cost going down, All right, that only means one thing. Everyone here in this room and elsewhere is going after like outbound AI-based email. Volume's gonna go up. People will be desensitized for sure. There's a time window right now to exploit it. Right now it works. 12 months from now, here next year, I'm not so sure. All right, and you gotta exploit this time window, but eventually you gotta think, how do I stand out? Is that by doing phone calls? Is that by doing LinkedIn? Is that by doing like in-person things? Email is at risk for sure. There's one company that's doing something interesting that I saw. Um, they're using your actual LinkedIn connections. When you go on a website, it says like, hey, uh, uh, you know Matt that uses this product, do you wanna, do you wanna ask him? So leveraging trust, uh, and networks, uh, so that you say, hey, like, I know Matt. Like, I, I, I had coffee with Matt. Matt is a real human, and I will trust his advice. Going back to the, uh, uh, certified human Parmigiano example I had.

Do you see any new channels emerging or things that we should be looking at from an outbound point of view, or is that not really yet?

Yeah, not really yet. Uh, like video didn't really pick up. Uh, phone call is coming back because of, uh, technologies like, uh, Oram. Phone calls. There's nothing coming back, but nothing new.

Yeah, I mean, I run SDR agencies and phone is where we drive pipeline, not email or LinkedIn. Cool, thank you.

Yes, I had a question about the AE meatbags. Ah, yes, at the enterprise level. Yes, if we're getting rid of SDRs because we're automating them largely out of existence, where do the AEs gain their skills? Where do they learn their craft? How do we— where do we stand in 10, 15 years' time? Yes. Yes. Um, absolutely.

I, I hope my wife won't listen to this recording when I, when she hears the AI meatbag reference. Uh, um, yeah, that's, that's a tough one. Um, that's true. As I said, in, you know, marketing as it is in sales, uh, the, the people who like SDR is a great job out of college for people who wanna go in sales. That job is going away and you can't just go out of college and start building like AI workflows. You just don't have the experience. You don't know how to write, right? And so either we're gonna have more training in one form or another that brings them up to speed to what is required. I think what we're gonna see is gonna be an expert's role to build and maintain those workflows with the sales tactics that we know for such a small fraction of the workforce that those people can be trained. We're talking like in the US, if I recall correctly, I think there's 750,000 1,000 salespeople, like SDRs and AE combined. Okay. How many of those are SDRs? I don't know, a third, you know, that's a good like 150, 200,000 people. Like that's a big workforce that, and a job function that lasts only like 18, 24 months. So yes, that is a real challenge of how do CROs train, their future closers without having them, you know, being SDRs initially. I don't have the perfect answer though.

Yes. Hi, based on your background with Drift and all the other companies you've had exposure to, are you seeing AI take on discovery in a very advisory role? Not just, you know, if then and some recommendations, but real advisory.

And when you mean discovery, you mean in the sales process, the understanding of the customer? Part of the sales process. Yes. Yeah. So Gong is kind of doing that and all the AI recorders. If you think of discovery, there's like, there's multiple parts of discovery. There's the SDR/BDR part of the discovery trying to understand the business and, you know, doing the information gathering, like BANT style gathering. And then there's the on-call discovery where they're trying to like create, some rapport with the, with the person and then try to, you know, get the information they want to, to help them close.

Um, but that deep discovery.

Yeah. No, I'm not seeing anyone do that for now. Uh, there's no system that is trying to, uh, make it click with the buyer to understand. For example, um, a good way to look at that is are we seeing AI be efficient in enterprise deals? And the answer is not yet. Okay. Enterprise deals are complex, they're long, there's lots of potential buyers, lots of information, uh, and AI just isn't efficient there for now. There's just too much context, too much risk. If you have any AI involved, this is still left to humans and will be left to humans for a while, uh, to be honest. Um, what we, what we are seeing, however, it's tools that help teams, uh, summarize data when they hand over. That is pretty cool. So we're seeing fantastic tools that help sales team AEs understand all the marketing context of the acquisition and the SDR context and tools for CSMs to understand, hey, this company that was just closed, what did we promise? What did we agree to? Uh, who's this customer? Who's going to be the user? And summarizing that so the onboarding goes a lot better. That I think is fantastic. But the— what you said about— no, I don't— I have not seen anyone do that. Let me know if you do. We're working on it. You're working on it. Cool. Awesome. I think we're at time. Thank you very much for coming. Have a good day.