Name: State of AI in Sales
Uploaded: 2025-05-15T18:31:23.778Z
Duration: 55 min 45 s
Description: State of AI in Sales

Transcript for "State of AI in Sales": Hello, everybody. Welcome to the session today. I'm Andy McCotter-Bicknell. We're super excited to get started. Alright. Let's dive in here. So welcome to the state of AI in sales. We're super excited to share what's going on, talk about what all the top teams in the space are doing, where we anticipate things going. But first, who am I? Alright. I'm Andy McCotter-Bicknell. I lead product marketing for AI here at Apollo.io, and I'll be your moderator today. I've been living and breathing this stuff, for over a year now at Apollo.io for, you know, from launching different features, collecting feedback, you know, from all of our customers to just helping shape the overall AI road map. So I'm really excited to dig in here, see, you know, share what kind of we're seeing with, sales orgs right now. we've got a quick plug here for something new that we just launched yesterday. If you're looking to really just get smarter on AI and outbound strategy, check out the new master class from Chris Orlob, one of today's speakers. Chris is a former sales leader at Gong who helped them scale into a company valued at 7,200,000,000 with a b, billion. He's an incredible educator when it comes to sales excellence and is now the founder and CEO of pclub.io, a skill building platform for sales and revenue teams. This master class, he worked on with us. He went really, really deep on just the exact talk tracks and strategies that he used to turn discovery calls into closed won deals. It's free, filled with examples, definitely worth checking out after the session. And then lastly, for anyone new to Apollo, we're an end to end AI sales platform with over 3,000,000 users across 500,000 companies. AI is baked into every part of the experience in the platform, whether you're prospecting, conducting prospect research, writing outbound messages that get replies, and even deal reporting. And a lot of what we'll share today is based on what we're seeing firsthand and both in how users engage with our AI tools and what they're telling us about, like, what's working and what's not. Okay? So here is the agenda. I'll spend about five minutes just kinda walking you through the latest trends that we're seeing across our user base and the broader sales landscape. And then we'll bring in a few great guests for a more, you know, just in general roundtable discussion on what's working, what's not when it comes to AI and sales, and then we'll finish with a q and a. So, again, please make sure that you drop questions as they come to you. Alright. So let's dive in to the current landscape. No surprise here. Sales is changing very quickly. Okay? It's harder than it's ever been. Sales cycles are longer. Budgets are tighter. Buyers are way more skeptical and distracted and demanding than they ever have been before. But while their expectations have really gone up, we noticed that seller capacity has remained pretty flat. Still have their hands full. They're still spending hours on admin work, whether that's logging manual data into their CRM, writing emails from scratch, spending hours researching different accounts that they should be going after, instead of actually doing the things that they are really trained to do, which is selling. And, this is the exact kind of environment where AI should shine, but the gap between potential and reality is still really wide. And what we've noticed this is this is something that really shocked me when I first start when when I first saw it. Anthropic conducted an economic index where they analyze their usage according to different professions. And sales represents 9% of The US workforce, yet only 2.3 of clog usage comes from sellers. So think about that. AI is arguably, you know, one of the most powerful force multipliers in decades, and reps just aren't using it that much. So this tells us that awareness really isn't the issue because everyone's aware of, you know, AI, but it's adoption. It's enablement. It's trust. That's where we really need to work to make things happen. But therein lies the opportunity. Right? So, the teams that do figure out how to operationalize AI, not just, like, test it here and there, but really embed it in their workflows are really going to pull ahead and fast. And we're already seeing that divide starting to form. So some teams, they're really scaling usage and seeing results, while others are kinda just sitting on the sidelines. And there's definitely more to it. So let's dig in here. Like, what specifically is causing those people that are on the sidelines to stall? According to our own survey data, 41% of teams say that their top issue is that AI outputs just aren't trustworthy or relevant. And that's really important because, yeah, why would you use AI if you don't trust its output? Another 27% say that the problem is a lack of expertise. They don't know how to apply the tools in the right way, even though they might be curious. You know? But aside from that curiosity, there's definitely intimidation, which keeps the it keeps keeps most sellers from actually just digging in further. And that makes sense. You know, we're creatures of habit. If we don't, you know, if we don't like a new thing, we'll just go back to what's familiar and what we know. Okay. But, like, what is actually working? Like, when AI does work, what's actually happening? And we found that, predominantly, it works by giving reps time back in their day. So we're seeing really strong traction in areas like writing outbound, summarizing research, prepping for calls, automating follow ups, all the really low leverage stuff that clogs a seller's day. That's really where AI can deliver consistent value without replacing a rep's judgment or more importantly their creativity. And that's the key. Right? Like, AI doesn't replace the rep. It really helps them focus. It lets them spend more time on high impact work, talking to buyers, building pipeline, closing deals, and it elevates them by clearing that path out of the way. And then here, the other piece too, I think that's really important to note, is that the best teams that we've seen, they really don't treat AI just kinda like a one off, like magic trick, like a checkbox type of thing that they have to deal with. They treat it like a legitimate system. They build internal AI playbooks. They train reps on prompt strategies. They refine what works. They document it, share it across their teams. And then most importantly, they keep sellers in the in the loop. It's a it's a human in the loop process. They review, tweak, personalize. And because they built that foundation, they're confident in what comes next. And we've noticed that high performing teams are twice as likely to increase their AI investment. But that doesn't mean just, like, buying more tools. It means just, like, doubling down on, again, enablement, process, and system level thinking. And this is probably my favorite line in the deck. AI is a capability to build, not a shortcut to buy. You can't just layer in AI and expect a bunch of results. You really need a structure, that enables reps to use it consistently, confidently, creatively. Otherwise, it really just ends up being another underused feature tool. We all we've all experienced that at our companies. Right? The those tools that we buy then kinda just sits collecting dust. It might sound good in theory when you purchase them, but it doesn't actually, you know, increase pipeline. Alright. So that's what we'll be unpacking in the roundtable today. So where are teams seeing traction? What's still a struggle? And how do we separate the noise from the workflows that actually help reps hit quota. We've got an awesome crew joining us today, folks who've been building AI native processes, educating thousands of sellers, and helping companies move past all the AI theatrics into real impact. And with the roundtable, we're going to get a mix of big picture strategy and hands on tactics. So let's get into it. We got some folks joining up here on the stage first. We have Chris Orlob. So Chris is a revenue leader, entrepreneur, the number one sales trainer, and revenue leadership coach. I mentioned before he helped Gong grow from 200 k to 200,000,000 ARR in five years as a sales leader, now as CEO of pclub.io. He's helped salespeople and sales organizations grow revenue at breakneck speed. So, Chris, thanks for joining us. Yeah. I'm happy to be here. Thanks, Andy. Of course. Then next, we have Sarah McKenna. So Sarah is the CEO of Sequentum, a leading web data extraction company delivering high quality custom web data for corporate, financial, and government clients. Sequentum solves clients' needs for the most precise data for mission critical analysis and informed decision making. Sarah, thank you so much for joining us. Thank you, Andy. Thanks for having me. And then last but not least, we have our very own Eric Quanstrom. Eric is the head of Apollo io Labs where he's building the AI influence pipeline of tomorrow. What's up, Eric? What's up, Andy? I think you need to switch slides. Oh, whoops. There we go. Sarah, there's your slide. And then Eric, there's your slide. There we go. Okay. I'm gonna stop sharing now. Oh, cool. So let's kick this off with a couple of questions. Like I said, if if folks have other questions, feel free to throw them into the q and a. But first question, I wanna just kind of keep it light, everyone warmed up a little bit. What is your favorite AI tool right now? Chris, I'll start with you, then Samantha will go to you afterwards. You're putting me on the spot. You know, I know this is not exactly mind blowing, but I have a lot of very deep conversations with ChatGPT on a daily basis. Mhmm. Absolutely. It's honestly it's it is one of I mean, it it's ChatGPT. So, you know, like, we don't have to get creative with it. It is so powerful. Sarah, what do you think? So yeah. No. I use ChatGPT all the time for requirements and for research, both for, you know, proposals that we're putting forth, but also in, you know, who's attending a conference that we're going to, you know, what are the types of, you know, problems they have, etcetera. It's quite quite useful. We also use Cursor a lot, on our development team. And, you know, as a CEO, I can see the value that we get from it. We're able to prototype, new solutions very quickly and get feedback from our internal stakeholders and external stakeholders. It's a big value driver. Absolutely. Eric, what about you? Yeah. Well, besides Apollo, of course, one of my Yeah. Favorite, AI tools is ChatHub. And largely because you can not only roll up chat GPT, but you get access to DeepSeq, Grok, Gemini, Claude, all in one interface. And so any query and chat that you would carry on, you can basically extend that up to six at a time. It's fascinating fascinating to see the different ways that the tools, these generative AI tools interpret any thread of discussion and the direction that they take your queries. Absolutely. Very cool. Alright. Let's dive in. So the first question, I wanna I wanna kinda pose this to everybody. Okay? So whoever makes the most sense, if you have somebody that jumps out to you, feel free to jump in and answer. So I wanna talk about, like, just the pitfalls of AI adoption, and just overall, like, how it might affect prospects or audiences at large. So what are the potential roadblocks that you see in the challenges of adopting AI as it relates especially to to sales in in general? I mean, I I could take that, you know, I mean, I think I think that the danger is and this is this is why, you know, we favor Apollo.io. I think if you just try to set it and forget it, you know, you're gonna get really bad results, and you can see it in your own inbox. Right? I have the title CEO. My inbox is absolutely littered with terrible inbound, that has nothing to do with me, things that I'm not interested in, you know, my you know, each page of my inbox, you know, if I'm not applying the right filters, you know, the subject says quick question, Sarah. Quick question. Quick question. You know, I mean, it's just it's it's really poorly done, and and those are, you know, they're they're they're not adding anything. They're actually taking away from, you know, my they're they're reducing my sort of feeling about that company. You know, so I think I think it's if you don't do it the right way, it's potentially, gonna hurt you, your reputation. Mhmm. I think another one is, kind of foundational. If you if you look at most AI solutions out there, they are marketed as a solution with no problem. Right? It's just kind of like AI for fill in the blank on, you know, whatever category we have because that's the hot new thing to attach your marketing message to. And regardless of how you're using AI, I mean, this actually speaks to the low adoption rate for for salespeople using AI. I I don't think they they know what problem it solves. Right? You might buy some, like, shiny AI product for your sales team or your go to market team to use, but because it's just kind of marketed as, like, as AI and and what's the use case and what what problem does this solve, I don't think salespeople pick it up at a very high rate just like any other sales tool. Right? There are so many sales tools that if you close a deal with a customer and you're, you know, you're selling one of those, you're selling and marketing is not done. You've got to market and sell the living hell out of it just to get them to adopt it because there are so many tools in their kit, in their sales, you know, tech stack. And so that, right, not being clear on, like, what's the use case, what what's the problem statement this solves for you as the end user? I think that's a pretty foundational one. I agree with Sarah too where it it could be very easy to misuse this kind of stuff. The first one is, you know, how do we get salespeople to use it? And I don't think the vendors are very clear on what this actually does for them beyond kind of, like, high level hand wavy AI statements. Mhmm. Totally. Totally. I've seen that too. I've noticed that, like, AI and I think this is just because there's so much possibility when it comes to, implementing AI as part of our day to day. We get really excited about what's possible, but we don't really stop and think about, hey. What what problems does this actually solve? Like and how can we actually make sure that people are aware that these pain points exist before we start sharing the potential solution? You know what I mean? And this is, like, no offense to the marketers out there, but I think that's lazy marketing. Because when you have a tool that can do everything, it's your cognitive responsibility to cut through that clutter and make it clear, here's the one, two, or three things that can actually do. And I think you see a lot of marketers who just kind of like to rely on the, oh, we can do basically anything for you. But but if you do that, like, there's no concrete mental picture you're triggering in the mind of your audience, and therefore, there's no problem to solve. There's no use case. Absolutely. Next question I wanna pose over to you, Sarah. So this one specifically about trust, like, just as it relates to, again, data and automation in the in the age of AI. So how do you build that that level of trust, either with within your teams, like, to to ensure that they trust the outputs to help improve their processes or to to other customers that you're trying to reach out to to trust your AI or, like, the data that you are selling them. You know, Andy, our our tagline is is trust in data. So so just cut me off if I'm going on and on. This is really something I might talk about. Yeah. No. We, we are a seventeen year old bootstraps company. You know, we're over a hundred employees. You know, what what what we do creates enormous value, and it's incredibly sticky. And our customers, they don't like to tell people who they use. But when they go to a new company, they take us with them. Right? I mean, there's a huge difference between a company that is doing things in the right way and driving trust, and customer loyalty and and a company that's not. And how we do it, there's many different things, that we do. First of all, we, we sort of embrace the best practices of just plain old automation. Right? We wanna make sure that everything is defined in an atomic way, that you have checks for each of the steps in the automation, that you have ways to validate, and that and we have integrated into our platform, sort of deep awareness of, every step of the way, what what the data journey is, and whether it's meeting acceptance criteria and sort of the thresholds of, you know, what what is considered valid and and correct, and when it's not. Right? Then there's all kinds of ways to detect errors and automatically handle them. And if you can't automatically handle them, it raises a ticket. It assigns it to the right person or team of people, and it tracks the resolution all the way through. And then the platform tracks changes. Right? You have to, you know, download, point, click, fix, deploy, you know, kick off a run that may be outside of a schedule. And all of those changes are also tracked. So we have very, very clear transparency in audit trails, and we have views for all the different stakeholders. So, because it's point a point and click platform, makes it easier for us to generate documentation of every agent automatically, and, it also makes it easy for us to insert approval workflows at any point. So the governance is there. We're SOC 2, audit, accredited, you know, all of those good things. So we use this as a way to drive trust. If there's anything going wrong in web data, which is notoriously difficult to get right, web scraped data, then every stakeholder knows. They can see. They have different levels of permission depending on, where you wanna drive accountability, and then every change has an owner. Right? So that's that's how we do it in a nutshell. And what we have found with Apollo is that you guys have a lot of this transparency and good practice, in place as well. And you never sort of set it and forget it, you know, in the in the broad ways that AI can sort of be set to autonomously run. We also run at intelligent agents on our platform for customers. We've been in production with intelligent agents since q one of twenty twenty three, delighting our customers. And we know that it's tricky. Right? You can do a really bad job, or you can do a really good job. And the difference is really how much human, cognitive effort you put into standing up that process and making sure that you get it right so it's adding value. Plus one to all of that. And I feel like Eric Quanstrom, you probably have a I I can feel your your gears churning a little bit here. You wanna follow-up with your thoughts on this too? Yeah. I think the the brutally honest answer of how you build trust in AI is is a hell of a lot of trial and error. Mhmm. You literally have to get deep and work whether it's prompts, whether it's the workflows that you're setting up. There's endless amounts of if you wanna call it QA on the back end of getting to expected results and deliverables that you want to see. This is actually where I I think we have kind of, like, almost a philosophical debate happening where and to do your job well, and we'll get into this in just a second, particularly with the role that's very big in in our organization called the go to market engineer, you have to know what good looks like. So, like, if you take a a discipline like outbound, like sales development, you have to understand what you're trying to get AI to do. Right? Whether it's focusing on research that's very signal and trigger based, you know, in addition to kind of, like, all the filters that exist in Apollo.io, or whether it's on messaging and really having kind of, like, a cogent persona based down to the atomic unit of one, in other words, the person you are communicating with, you have to understand, like, well, what does good look like? And can the AI systems that I'm leveraging get you there? And that's really where the trust kind of comes in or comes out to the point that you were making earlier, Andy, on on the slide around, you know, populations of users getting to that point. Absolutely. Yeah. I've noticed that just people, I think, have an expectation, you know, that you can just, like, punch in a prompt and, like, boom, you have a perfect output or it should be that way. But the reality, at least from what I've noticed secondhand from seeing all of the work that Apollo has done, is just there's so much more to it when it comes to actually building legitimate prompts that get you the outputs that you're looking for. It is not just a simple, like, type something in and boom, you have what you need. It's it it is a constant iteration of, like you said, trial and error. And and also oh, go ahead. Go ahead. Concepts that I would build on to that statement are we live in a a world that's becoming more sophisticated and complex by the day, especially with AI. And so, like, the great mistake is that people come in and they they do I'll I'll use a word that seems not as pejorative as as I'll make it sound, but, like, they they prompt in a very shallow manner and expect that the AI will kind of infer what's in their head or what, you know, the deliverable should look like. And that is absolutely a a key mistake that a lot of people make. And so, you know, the the the the running joke, especially in Apollo, is that you you actually can't get too much information into a prompt. Right? Like, AI is really good at consuming copious amounts of information. That that is kind of a known known. And so, like, prompts that tend and trend towards more sophistication as opposed to less tend to do better from the deliverable or output side. Absolutely. Yeah. We find a switch. Yeah. I I I I feel like somebody mentioned GTM engineers too, which is actually a great, a great, triage into this next point that I wanted to make, where we're talking a little bit about GTM engineers and, this new this new profession, that's pretty much, like, the the consolidation of ops, sales, growth marketing. We pulled this, go to market engineer job description here as well. And I'd love to kind of just pitch it. I wanna start with you first, Chris Orlob, because I feel like you would have the the such an interesting opinion on this is, like, how like, where do GTM engineers sit within, like, just the the universe of the traditional sales rep? Is this something that other sellers should try to learn skills of being GTM engineers? Or or, like, where do you see this role in just the overall state of things right now? I think it's an open question where they report to just based on, like, the remit and the mission of the specific job. I think it's a very fundamentally different skill set than what most salespeople have. But there's this guy I mean, maybe you guys know him. His name is Jordan Crawford, and he is somebody that I would consider to be a go to market engineer. And I've sat in on some of his trainings on, like, teaching salespeople and founders, like, how to use, you know, a variety of AI tools to get in front of people who are in market. And the level of knowledge and and skill that it took to do that just based on being in the audience was, like, way up here. So I think there's, like, a pretty big skill divide in using some of these tools in a way that helps you, maximize value from AI. But but I think the question poses something that I believe, which is to me, like, the number one use case or at least potential in AI for go to market is not just, like, eliminating manual tasks. I think that one's a given. I think it's getting in front of hot buyers. Right? Because, like, for for decades, people have talked about, you know, when you look at a % of your market, only 3% of it or 5% of it or 10% of it is in market or whatever. And the solution to that before AI was and therefore, you need to develop marketing messages, sales motions, etcetera, that convert people that are not in market to being in market. And that's still there, but three, five, or 10% of a market being in market is massive for most markets. And if you can tap like, use AI and some of these tools to continually tap into that low hanging fruit and continue to get in front of those buyers and max out that part of your 10 before you worry about, like, digging in and reaching into the colder parts of your market. To me, I think that's one of the major but in my opinion, is probably the number one use case of AI in in a go to market context. Yeah. I wanna build on that point because, Chris, I couldn't agree more. In in market buyers and targeting was was the great kind of, like, promise of maybe the first wave of of where we saw at least machine learning in intent data kind of, like, coming on the scene from a pure go to market or outbound perspective. One of the keys to anything that we do over in labs is really having a philosophy, a perspective, if you will, in the world where signals and triggers prospect led actions or company led actions that would land them on your lead list for outreach is the right way to be doing any kind of go to market activities as opposed to the old way, you know, five, ten years ago where the vast majority of the world was viewed everything through a vendor centric lens. Right? Like, go grab a list. Maybe it's articulated by geo or industry or size of company and work that list. Oh, this is the height of strategy too. A to z. Literally, where you would pull up and and, you know, like, well, why are you getting a call? Well, because your company is called Acme, and you're the first on my list, which is a terrible reason for Outreach, in any in any way, shape, or form. What I love about where we're at in 2025, and this will only become more true, is that the bar is raising. And I say that with all due love and respect from the buy side. And so when you're on the buy side and and you're deemed a relevant prospect to reach out to, an executive who holds budget, someone who would be, you know, kind of like the target, one of the first things that that usually comes to most people's stage gate and filters for responding to any kind of cold outreach is a almost a mnemonic. Why me? Why now? Why care? Right? Like, most prospects think about those thoughts in their head every time they're about to interact with anything, whether it's a cold call, a cold email, a social touch or outreach on LinkedIn, an ad, or anything that they see on the web. Right? The difference with the web is it's usually self directed. Right? Like, why me? Why now? Why care? I'm doing the search, and I control it. So the the the big takeaway though is how do you get into that slipstream to your point when organizations or buyers are throwing off in market signals? That is really the the architecture of a go to a successful sophisticated go to market company when they are putting themselves in the in the collision path, so to speak, of their buyers. And AI is the great enabler, the key that unlocks the door for exactly that motion. Yeah. And for an offering like ours, which is technical. Right? We offer software as a service, on prem software. We offer platform as a service in your browser. We offer data as a service if you want bespoke scraping services, and we also have a data catalog. Right? We basically have four different things that we're selling, which is typical of a bootstrap company. Right? The AI knows what all those things are. Right? So it's it's it's it's it's a completely different tool, you know, in terms of, you know, go to market. You have a you have someone who you have maybe it's not someone, but you have a capability that is all of a sudden very well trained in what you do. Right? AI is very familiar with the Sequentum offering. We have over 800 pages of support documentation online. AI knows all of it. Right? So it's it's you're getting this expert or even a mixture of experts, with any kind of messaging, any sort of reading of replies. You know, it's it's it it it you know, done the right way, it's it can be an enormous enabler. Absolutely. Something that something that's interesting that I feel like everyone is kind of touching on right now is how to maintain, like, authenticity with your outreach while leveraging AI, which I think is typically, categorized as something that is very natural, like, inhuman, nonhuman. Right? I think a lot of people are concerned that well, rightfully so, that by implementing too much AI, we're going to take the humanity out of our outreach, our authenticity. But I think if used the right way, it's going to bring more humanity to the process. It's going to allow us to reach out to more people that, we would naturally, connect with, that we would naturally, that would naturally find value in our solution. And so I wanna pitch this one to you first, Chris. So, like, specifically, like, when it comes to maintaining authentic relationships with clients, prospects, like, what's what are the most effective strategies for leveraging AI, to to build on that? Well, I I think it goes back to what Eric said earlier, which is most not just salespeople, but people, use shallow prompts. And I think the way to inject authenticity versus something that's obviously crafted or written by AI is, are you using the AI for the substance or are you using it for the cosmetics? And if you want to maintain authenticity, you wanna use it for the cosmetics. Right? So if I'll give you an example. If I'm prospecting somebody and I'm using ChatGPT to write a cold email, I'm gonna write, like, a bad version of this is, I am targeting Shopify. What problems are they dealing with, and can you write me a cold email to their VP of sales? Right? That's asking AI to come up with the entirety of the substance, the polish, the cosmetics, everything. And it's gonna be easy for a sophisticated buyer to sniff out the fact that there is no authenticity behind this. Now if, on the other hand, you are not shallow with AI and your prompting is deep and you're bringing your own point of view to the prompting, and you're saying something like, here's what I know about their challenges, here's the angle that I wanna try for, here here's the value proposition and the pain that I think they're feeling based on what I know, you know, something like that, and you're having a multilayered threaded conversation. And now you say, based on the above conversation, can you write me a two paragraph called email that that will resonate? At that point, you as the human, you've provided the sum substance, and now you're asking AI to do what it does best, which is put together the right string of words that aren't gonna be total a total cluster if I if I try my hand at this. So that's one use case. Right? We're talking about, like, the the one the use case everybody knows, which is using AI for full email. But I think it's a good rule of thumb is are you relying on the AI to provide the substance, or are you providing the substance in the point of view and now asking the AI to simply clean it up and make it more polished for you? Yeah. You can even take that a step further. To me, that that's the exact point that, pre Gen AI days, the Challenger sale was making. Right? Like, paying attention to a company's business to develop commercial insights to go in and basically provide value around how some business could be doing something better, especially in the most common obvious cases of things that they're doing that they shouldn't. Right? Like or things that with your tool or solution or product or service that they could be enjoying a better tomorrow than they have today. The essence of all sales. Right? Like, I always like to say that sales is a change management exercise and, you know, because if there's no change, there's no sale. And so how do you work your way back from that change? How do you work your way back from, again, a better tomorrow in that company than they have currently, their current status quo? And so having the substance to really have that perspective of of why you would need to have a sales conversation in the first place is the essence of what I think you're talking about, Chris. Well, I I think what one of the ways this shows up is your posture toward using, like, a prompting tool should be one it's almost like coaching a little kid. Right? You tell it what to do. It gives you a response and you're like, okay. Good try. I wasn't specific enough in this area. And so now I want you to go deeper over here. And then it gets you to the next level. And it's like, good. That that now we have an improvement. I I I lost sight of this piece over here though. And so can you now do this? Because AI is gonna do basically exactly what you tell it, which is both good and bad. Right? It's it meaning if you're not specific enough, it's not gonna, you know, it's not gonna give you the level of specificity you need. And so your role in a lot of these situations is almost think of yourself like a coach with the AI and your prompts are the coaching feedback and the questions and the directives. Yeah. I think that the best label I've ever heard for this is what, some have referred to as the eager intern. So, like, if you think of AI as, like, an eager intern that really wants to do a good job, one of the places that we get ourselves in trouble, in addition to shallow prompting, but also as a and byproduct product of shallow prompting, is if you ask AI to do something where there's nothing it can draw upon or no data to go source, so it'll make shit up. And so, ultimately, like, that eager intern wants to fulfill the job, give you an answer, and oftentimes, the answer is completely, you know, fabricated unless you have an even bigger problem on your hands. Those are a couple of, I think those are two of my favorite, things to think about. What Chris mentioned with cosmetics versus substance, that's a great frame, and then the eager intern. That is such a great I've experienced that so many times too, like, where, like, I I'm reading the the response from from ChatGPT or whatever. I'm like, that's just not true. That's that's absolutely incorrect. I appreciate the enthusiasm that you are giving me this information, but it is just not true. You know? Okay. Listen. I have one more question that I wanna pose to everyone before we dive into q and a. And so I want to start with you first, Sarah. So why don't you tell me just what's working for you and your team as it relates to AI, and then where are you planning on investing more? Where are you gonna double down in the future? Yeah. So we're, we're actually so we've always integrated AI into the workflows that we run-in our platform, but we're actually using it now to generate, agents, which is super fun to see. It's also a little bit scary. We're, you know, working very hard to make sure that it works super well. So we, you know, we we're really focused on testing. But, you know, we've got we've got prototypes now that you where you can put in a URL and automatically get a schema and sample data and questions, you know. Do you want this or do you want that? So it's it's it's really fascinating what you're able to do. So in our product and our platform, we're integrating more AI. But then in the team at large, like, our UX team is is using AI for to mock up new pages and new interfaces. I already mentioned Cursor, for unit tests and code documentation, as well as just mocking up code itself and pointing out bugs. Our QA team has been exploring, AI for test automation. There's incredibly, valuable tools out there that will detect security bugs, you know, etcetera. I'm using AI in my email, to cut down on a lot of the junk that comes through. There's just so many ways that we're enhancing our productivity as a team. I've probably left out a bunch of them. But, yeah, it's it's, something that we are wholeheartedly embracing, and, obviously, we're, we we we evaluated a number of different, AISDR solutions and having a healthy dose of skepticism about, you know, what you can do if you let AI just kinda run wild. You know, we've we've come to Apollo and really enjoyed the consultative approach. What about you, Chris Orlob? Yeah. There there's a a couple of ways we're using it. UX and u or UX and UI design and mock ups is is a big one, kinda like Sarah McKenna just mentioned. Another one is we're pretty militant in our sales process about using what we've heard slides as well as, like, business cases. Right? Like, during the discovery call where we'll extract the information we need. And then before presenting a a demo, we frame the demo with the what we heard slide. Those used to take forever. Right? We would have to go back and listen to, like, the call recording because there are a couple slides we're summarizing. And instead, like, you know, it it's like, hey. Build me a what we heard slide, based on this call's transcript, and I want one slide that's, like, current state versus desired state and what's the financial impact. And then the next slide, I want to be the root causes of that. And that used to be a thirty to forty five minute exercise, and so we could only really do it on, like, deals that warranted that. And now we do it for every deal, and it takes us about five minutes. And I'm yet to use one of those or have my team use one of those, and they present it back where the customer is, like, doesn't respond with some form of, yes. That's dead on. Alright. Let's let's move on. And then the other one is, like, what we're doing in our own product. When it when I think of the use cases of AI, there's efficiency, there's intel intelligence, and then there's signals. There's probably more than that, but I think of it through a go to market lens. And so one of the things we're doing, like, knowing we have the skill development platform, is now working on using AI to analyze rep skill gaps according to data. Right? Discovery skills, multi threading skills, negotiation skills, because skill assessments historically have been a very manual observational, process that that can take weeks, if not months, in some cases. And now if you have AI that's able to observe different levels of skills across many different roles, many different organizations, not just one, now you're introducing a level of intelligence that wasn't really possible before. I I would like to add one more thing that that we're doing a lot of. So we it used to be people would call us and say, you know, I understand you have these well heeled customers. I wanna do something similar to them. Why don't you tell me, like, what are all the data sources? And, of course, we can't do that. Like, you know, we couldn't help them at all. Right? But now someone calls and says, I'm trying to generally do this. How would I do it? And with smart prompting to chat GPT right there on the call, we can discuss data sources and schemas and schedules and, you know, how they might visualize it and how they might use the data. And by the end of an hour long call, we have pretty clear requirements to go back to the team and do, you know, some sort of a mock up or a POC. So it's it is incredibly helpful at, getting, you know, sort of distilling a set of requirements and clarifying those requirements and helping provide that depth, even in the conversation the initial conversation. Absolutely. Alright. So we are at the q and a portion of the session. So if you have any questions, please jump over to the q and a tab and, add them in. We'll filter through and ask them as we see them come in. Okay. Cool. So let's dive into some q and a. Let me stop sharing. The first question I have, this one this one's really interesting. I'm gonna start with you, Eric. So this is from Yurian Van Reyes. I I hope I'm pronouncing your name correctly. If market signals are the new ICP defining category for outbound work, not every company is very prolific with their accomplishments, strategic changes, wins, losses, all that kind of stuff. So how do you create a prompt that filters out the right signals or even picks them up in the first place? Yeah. That's a great question and something that we've been spending inordinate amount of times over in labs, figuring out. One of the ways that we've attacked that problem is actually building a GPT that we call signal craft to even get us to the starting line. And so when you think about it, like, using AI to help you prompt better AI is actually a really natural use case. And so inside the GPT are are essentially a bunch of methodologies and training frameworks around how you should be thinking about a given set of signals or triggers for, like, discovery because they're literally unlimited. Right? And Sarah can probably reemphasize this to the high heavens where, you know, if you think about the open web and what you could crawl and and maybe the start place would be the website of your target customer. Right? Like, in all the things that you could learn about their behaviors or their in market buying activity, just on their own website, it's staggering. Staggering. I would encourage you perspective wise to think first there. Right? Because then you're starting to walk a mile in the shoes of the people that you wanna do business with. Right? Like, you can learn a lot by what they're putting out publicly to the world and taking advantage of those signals by if you're using Apollo.io, rolling up to power ups and using the unstructured data that is harnessable with power ups in conjunction with any lead list that you might be building using all of our 65 kind of like standard filters. So it's kind of like combining, like, the best of go to market over the last few years with the best of go to market in the future and beyond. Pro tip. One of the other ways that I like to think about this is you're not just sourcing signals and triggers for lead prioritization, which should be your number one goal, right, at the research phase of being able to determine, hey, if I'm gonna reach out to a thousand companies this month, who's number one? Who's number nine nine nine? And have that in order so you can define your own kind of like outbound cadencing, resourcing, process orientation, etcetera. But then my favorite pro tip is once you've got that lead prioritization kind of like mindset down, the next way that you can really leverage key research in unstructured data is to pull it into messaging that you're then gonna use. Chris used the word earlier cosmetic, and I actually think that that's a pretty appropriate word. I never thought to use it before, but that's a lot of what you're doing in kind of the show me, you know me type messaging that you would wanna do to even get engagement. Right? Like, what you learned about a a person or a company from that unstructured data kind of like research effort should go into your messaging of why they should respond to you in the first place, why they should care. Every human that I've ever met that roams the earth, like, cares most about themselves more than any other person. Right? That's just natural. And so if you can kind of, like, dial into the other focus, make it all about them, make it, like, where they are the object of your kind of, like, affection, so to speak, in an outreach sequence. You will have much better performance than if you make it all about you as a vendor. You know, coming in cold, interrupting their day. Food for thought. Absolutely. And Chris, Sarah, any any additional thoughts on that? Go ahead, Sarah. Hugh, you nailed it. Yeah. You nailed it. I think that was really, really good. I think, you know, show me, you know me, and being able to harness the capability of AI to go out to the web to really understand any research reports, any messaging that's out there, even beyond just the website of the company. You know, these are activities that took an enormous amount of time, before this new new tool set and these new platforms and, you know, used in the right way, you're providing tremendous value by showing up, in front of that person who is sending out that signal that that they need what you have. %. Okay. Next question we have is coming in from Scott Knowles. So with so many layers of a sales and messaging process, what's the best way to manage AI hallucinations as prospects go through the pipeline? We kind of talked through this a little bit about, how much context, AI needs in order to be, in order to not hallucinate. But would love to hear, folks' thoughts. Anyone anyone have something that jumps immediately to their to their minds? I mean, for us yeah. I mean, go go ahead. Go ahead. No. No. No. Mine's a blatant product pitch. Go ahead. Yeah. Eric Quanstrom. Someone mute Eric Quanstrom. That's right. Well, I'll try to make it short. So one of the the the coolest unsung features that Apollo.io has released in the last few weeks is what we call content center two point o. And it's really like your AI brain inside of Apollo.io where you get to load up all the information about the personas that you're reaching out to. And so you can think of it as like a pre prompt because AI will consume or can consume in any of your future prompting in Apollo, the content center and take advantage of that so they can pattern match. Oh, I'm reaching out to CFOs? Great. Here's everything that I know in this brand's perspective about CFOs. So I can throw a lens and a harness on top of that, and it'll affect everything that then is output from a messaging perspective, which is why we call it the content center, in Apollo. And so I'm I'm in love with this feature. Shout out to the product team. And and I I couldn't help or resist giving a little airtime. Yeah. I've been and that's the exact right approach. Right? For intelligent agents, every step of the way to mitigate hallucinations, you need to have, multiple different ways to check and confirm and reconfirm that you're getting what you expect. You're setting up those acceptance criteria. And the way that you do it in traditional automation is setting up some sort of oracle or some heuristic test. Right? And that's exactly what what Eric has described. And that's certainly what we've seen going through the process, with Apollo is that, you know, they just sort of reflect back to us. We believe that this is right. Let's go through and see what what has happened here. And we're able to course correct and see examples and and edge cases and whatnot. And that's the exact same approach that we take. Absolutely. Next question. I think we have time for one more. So this one is let's see here. From Shashwat Yadav. What is I'm gonna I'm gonna kind of, paraphrase this a little bit. So what is the best way to refine your ideal customer profile or your ICP in the age of AI? Does does AI change anything as it as it relates to how we're building our ICP and how we're taking action on them, or is that generally staying the same? Chris Orlob, I'm gonna I'm gonna lob this one over to you. Yeah. I I can be conceptual about this one, but the entire process of pinning down your ICP comes down to pattern matching. Right? So, like, if you're like a founding AE or or if you're a founder who's doing selling, you you usually accelerate who's in your ICP and who's not simply by having a lot of sales calls, closing those sales calls, seeing seeing how those customers become successful or not, as as they're onboarded. And the more data you're collecting, the easier it is for AI to start making it clear what those patterns are for you. Right? So if you've acquired a hundred customers, for example, and 15 ended up churning and there's a cohort of those customers who stayed for a long period of time, the sales cycle was short, the customer acquisition costs were low relative to others, then that's patterns that you might not be able to see that AI can help you detect. Right? So I I think it's a tool like like I said, the way I think of AI, at least in the go to market use cases, efficiency of intelligence and then signal. And I think this is one of those intelligence use cases where, yes, you might be able to see those patterns yourself. You might be missing some of them though. And why not have a tool that's very good at this exact skill, pattern matching, help you pinpoint other attributes that make up your ideal customers and also attributes that make up your non ideal customers. Right? I I actually think that's an entire concept that's really overlooked, in go to market is like, yeah, we wanna hone in on our ICP, but we also wanna be definitive of who are the Phantom ICPs. They might look like ICPs if, you know, just kind of on the surface, but there are certain attributes that would make them inherently unsuccessful as a customer. Right? And can we divide define both of those worlds using AI as a tool to help us do it? Absolutely. I, I previously, was, leading competitive intel at Apollo.io, and that was one of those things too. It's very similar philosophy. It's great to understand who you're competing against definitively, but then also who are you not competing against. Right? It's like that same kind of, like, defining exactly who it is that you're trying to target or, like, go after. But then also the people that you you don't wanna spend as much, like, brainpower thinking about, people that, maybe that would take you down the wrong path, a market that's not big enough. So absolutely plus one to that. Anyone have anything else to add on to what Chris said? Okay. I'm resisting your answer Answer, Dilshad Ahmad's question in chat with another shameless plug. Go ahead. Well, you can log it into the chat, and then, I'll I'll go ahead and and, do the goodbyes really quick because I wanna make sure that folks here have, can plug in all of their stuff. And so what we'll do right now is I wanna just, number one, thank everyone for joining. This has been awesome. Really, really great conversation. Thanks for all the guests. And, also, thank you very much to Chris and Sarah for for joining and spending time out of your busy days to hang out and talk with us. I would love, if we could kinda go around. And, Chris, where can folks follow you if they wanna learn more about you or PClub.io? Yeah. Just, you know, feel free to connect with me on LinkedIn or follow me. I tend to fall behind pretty quick on DMs. But if you have a question, I'll I'll do my darnedest on keeping up with it. And if you have a sales team and you are struggling to up scale them to sell in today's market conditions and you wanna solve that problem, you can go over to PClub.io. That's that's what we tend to help with. Awesome. Sarah, same question to you. Thank you so much for joining. Where can folks follow you and learn more? Yeah. Well, I'm on LinkedIn, Sarah McKenna at Sequentum. I'd like to give a plug for the Sequentum Cloud. For the first peep 50 people that sign up from this webinar, we will write your first agent for free. That's for either WebScrape data or an intelligent agent. And you can, of course, connect with me on LinkedIn. I'd also sometimes fall behind. But, yeah, looking forward to working with you. You get a free trial when you sign up as well. Mhmm. Love it. Eric Quanstrom, what about you? I'm really easy to spot. There's very few quants in in the world, and so you can go find me on LinkedIn. And, I I'm generally pretty responsive. Awesome. Alright, everyone. Till next time. Again, next webinar is in a couple weeks. Hope to see you there. Until then, hope you have a great rest of your day, and thank you very much for joining us. Bye bye. Thanks so much. Thanks for having me.