Name: Live Workshop: How to Plug Competitive Intel Into Your Agentic Workflows
Uploaded: 2026-03-24T17:30:55.541Z
Duration: 59 min 53 s
Description: Live Workshop: How to Plug Competitive Intel Into Your Agentic Workflows

Transcript for "Live Workshop: How to Plug Competitive Intel Into Your Agentic Workflows": Hello. Hello. Hello, everyone. We are live. Welcome to our live session today. I'm Adam McQueen. I lead our content product marketing crew here at Clue, and I'm joined by my cohost cohost, co emcee. The person that's gonna be making sure we're running smoothly today is Jenna. Jenna, what's going on? Hello. Excited for this one. This will be a this will be a fun one. Yeah. My name is Jenna. I run enablement here at CLU. A big focus of my role is I also run our internal win loss program. So really excited for MCP and to hear from you all a little bit on how we're thinking about using MCP operating in this new world of AI. This should this should be fun. Everyone that's tuning in, I'm sure folks are starting to trickle in. Jenna and I don't wanna feel like we're just talking into the void here, so let us know in the chat. Where are you tuning in from? It's really rainy and gloomy in Vancouver. Winter is whoever's in the most tropical or sunny spot right now just to give us some FOMO. So let us know in the chat where we're tuning in from. Hi, Bert. Wow. Dubai. I don't know. We've we've had beautiful sun, but today, it decided we're gonna have we're gonna have a typical Vancouver day. So, yeah, let us know where you're coming in for Dylan. What's going on, Tara? Oh, we got some friendly faces. Excited. for this. Jenna, before I hand it over to you to kinda give a lay of the land on the agenda, what to all the fun things we can get into. We asked this question for folks when they're signing up, but I'm gonna throw it out there again. Home of two I'm Michael, Oh, I boy. March Madness. We got a we got a cornball on Friday night. Before we get into the session, I wanna ask the question again for folks. So drop this one in the chat, but there's so much there's so many workflows being built. Everyone's sharing all these different things that they can do now with AI. If you had a magic wand, you could automate one part of your CI workflow or product marketing workflow with AI. What would it be and why? I'd love to get a few in the chat. Maybe we'll address a few of them today. We'll see. But I'm just curious how people are starting to think about this as, as we get into this session. Jenna, you wanna let let the folks know what to expect? Yeah. Definitely. So we got a full lineup. We have quite a few guests who are gonna be joining us today. We're gonna kick things off by focusing on four we're gonna show you guys four different use cases for how our own internal teams are starting to use the Klue MCP server. So we've got a seller, Zach. He's gonna come in. He's gonna share a bit about how he's been using the Klue MCP server to write emails for him. So his prospecting emails, emails he has mid deal, how he's including in some of those communications that are going out. Then we have our own AI ops wizard, Isaac. He'll come on and show talk a bit about emails at scale. So how he's taken our MCP server and created a system for how our sellers can use MCP in their day to days. And then, Adam, you're gonna get on stage. Show us a little bit of ACI breakdown that you did of our web page, like a competitive comparison page, which is pretty cool that you built out. I'm excited for that one. I will show you a competitive overview that I did looking at some of our wins and losses using the MCP, and then we're gonna finish with two amazing customers gonna come on stage and we'll do an open Q and A. That's where we wanna hear from you and the audience what questions do you have. And we'll get the pros who are out there doing it in the real world talking about their experiences. For sure. And as always with these sessions, if you've been to these before, ask questions whenever, make comments whenever. I like to react to stuff that's happening in the chat. We'll bring them in as we go. We'll have some time at the end as well-to-do q and a with some customers, with some folks. But anything going on today, anything that pops up as we're showing stuff, drop it in there, we'll address it as as quickly as we can as well. Yeah. And, Love, it. Tim, this will be recorded. But that doesn't mean you're allowed to leave us early, though. way. Love some of the, yeah, the examples there in the chat, updating battle cards, keeping up on recent developments. Those are great. Okay. So we're gonna kick things off. I'm gonna hand it over to Adam. He's gonna give us a little bit of a peek behind the curtain, tee us up with our Klue engine. Obviously, we're talking all about the Klue MCP, but how does it actually work? Like, what's what's behind the scenes? So he's gonna take us through a little walk through of the Klue engine. Cool. Alright. Before we get into all of the fun stuff, I pulled all of these posts from my LinkedIn feed last week. I literally think it probably took me maybe, like, two minutes of scrolling. That's because everyone is building right now. Everyone's posting about building. Everyone's showing a new workflow. And, honestly, it's really, really exciting. I mean, it's why we're all here. I think it's super exciting for PMMs and CI leaders the way that we can think about automating more redundant tasks, even reimagining some of the workflows that we've traditionally had to do throughout our careers. And so it's really exciting. But before we get into the workflows, I just wanna take one beat to ask folks, is your data keeping up with what you're building? And what I mean by that is all of us have played around at this point. I'd say near unanimously, we've played with a GPT, a Claude, a Gemini, something out the box. Your reps, maybe yourself, maybe some folks in the org have asked a competitive question. And so it looks pretty good, pretty solid at first glance. And you kinda scratch the surface a little bit, and you see what's powering the answer. You've got broken links. You got four zero four errors. Maybe a Reddit thread from two years ago. Maybe a bunch of your competitors' own blogs are fueling the answer. And that's only with out the box LLMs. But the bigger picture here is whatever data you point your AI towards, the quality of that and how you process that is gonna determine everything downstream. And so not even just an out of box alarm. I think we kind of understand the the limitations that come with that. But let's say you dump a bunch of docs and decks into a notebook LM. Maybe you actually start to link up your live data, your internal g drive, your your call recordings. But here's what's actually happening on the hook with those systems as it relates to producing competitive answers as well. You dump in your data. Let's say you got your call transcripts, your g two reviews, your your decks, your docs. It's gonna get indexed by by the rack, and it's gonna be indexed into millions of chunks. So when someone asks a question, the LLM is then gonna go hunting through all of it to find the relevant pieces. This is what we'd like to call at Klue, the retrieval lottery. And we've seen this with a bunch of folks that have started to try and build their own kinda compete debate data engine. Sometimes it's gonna find the right thing, and oftentimes it doesn't. And that's because when you're running this across a really, really large dataset, you might get two reps ask really similar questions and get completely contradicting answers because the system is gonna pull potentially from different chunks each time. And so where there's a standard rack system, it's gonna store your raw data. It's gonna collect all that data. It's gonna generate insights or find answers from it on the fly when a question gets asked. It's gonna search the warehouse and try and find that needle in a haystack. But what Klue and what our Stacks engine, streaming agents of Klue as we lovingly call it here, how it does it differently is it uses streaming agents to go through your data proactively. We're gonna generate the insights first before questions are even asked, and then we're gonna store those insights. So by the time anyone asks a question, the answers are being built from scratch, from a pile of raw data, is being pulled from a pregenerated and organized intel. So the context window where your LLM is work working from is clean, it's relevant, and it's grounded in the same source of truth. So what does that actually look like in practice? I was talking to our VP of engineering, trying to wrap my head around. This is a lot of technical stuff. What what does this actually mean for us? And he shared with me, think about it like a mail sorting system. Before yourselves even ask a question, Klue agents are already going out there through all of your data, your call transcripts, your win loss interviews, your GT reviews, competitive page updates, all of that. And it's sorting through and organizing it. So pricing signals go in one box. And not just like, here's the pricing page, but pricing shifted and the market perception of the pricing model your competitors updated or objection patterns in another one. And not just these are the objections coming up, but these are the ones that came up in 11 deals, and here's the three responses that are being received most positively from buyers. Win reasons, breakdowns based on buyer requirements. All of these agents are preloading the context window for your LM with these insights. And so when a rep asked something like, what's the latest on a CompareX pricing? The answer isn't going through a mountain of raw data and hoping it finds the right thing. It's actually pulling from something presorted, pregenerated, and that was already there waiting. So I really wanted to kinda emphasize what this is what the source of truth that Klue system is providing rather than what we can build on top of this. Because this is what powers any of the workflows we're about to show today. If you don't have a trusted compete data engine, something you can actually rely on to provide accurate answers, then you all of the work we're doing on top of this presents risk. And so what we're doing here, we're gonna be producing a really rich set of data that's gonna be implemented into whatever tools, whatever LLM you're using, whatever workflows you wanna build, and that's the foundation of everything that we're gonna get through today. So there's a little bit of my spiel here, and maybe we'll flip it over to some of the fun workflows that we're building on top of this data now. Who we got, first, Adam. Jenna's Screenshare? First up, we've got Zach. So Jacques is gonna show us two seller workflows. Jacques is a superstar AE here at Klue. He's gonna come on and show us a couple of seller. reports. Oh, hello, Jacques. How are doing? What up, Jacques? Good. How are you guys doing? Good. What do you got for us today? Yeah. I've got a couple. I think they're pretty cool. They're probably gonna be pretty elementary compared to what Adam, Jenna, and Isaac are gonna show you guys. So excited to start the day off here. Yeah. I'm an account executive here at Klue. So these are some of the workflows that, you know, my myself, I I I go through, and I also work with hundreds of prospects prospects or customers a year in terms of developing them internally, as well. So so really excited to walk you guys through a couple workflows. Let me just start off by sharing my screen. There we go. So, yeah, two main workflows here today. One is gonna be geared more towards prospecting, and then I'll get into another one, when it comes to adding a little bit more deal context to a potential email. So I want to start off with prospecting because it's relevant for me as an AE if I'm trying to break into account or SDRs, BDRs. You know, we we we know maybe one of your or their accounts has a competitor. They're currently using them. So how can we keep position? How can we get that meeting set? How can we try and get them to switch over to our tool? So as you can see here, a pretty simple prompt. Can you help me draft an email to a prospect about why Klue typically wins against Crown? I'm using my demo instance, so a lot of this stuff is dummy data, by the way. But, yeah, this is really all you would have to put in. I added in a little bit of extra context because I find that Claude asked me a lot of follow-up questions, and I didn't wanna put you guys through that. So yeah. And then let's jump into what Claude starts thinking and some of the output. So right off the bat, you're gonna see a running skill right here, and that skill is actually my email writer. So Isaac, who you guys are gonna meet in a little bit, actually gave every go to market employee a writing profile here at Klue. So I put that into Claude, had it attached to a skill. So whenever I ask it to help me with an email, it immediately matches my tonality, everything like that. And then we can start to see Claude shifts. Okay. We have the email style. Let's shift into more of the competitive material. Right? Instead of immediately going out there, searching the public web, maybe pulling from untrusted sources, it's first gonna go in and say, let me check if there is data in clue that I can use. Okay. Great. It's gonna look through some tools. It's gonna go to the battle card, and then you can see down here, it actually extracts that battle card. You can also see over here it's connecting to our MCP, which is great. We love to see that. I didn't even ask it to do it. So it's knowing, hey. I brought up a competitor. Let's start looking through Klue. And then we can see it starts to pick out some really good information. Once we have some of that insight, it's gonna say, okay. Great. We have all this data. Now let's match it to Zach's style, tone, and then it starts to spit out a really good email. And it's not just, hey. Here's why Klue is typically better. It knows that the account I'm going after is currently using Crown, so let's talk about maybe some of the downfalls we typically hear from folks we're working with and then how Klue actually combats that. And it also attached to a case study, which is awesome. So, yeah, pretty easy way to immediately break into account, draft a prospecting email once again for AEs like myself or even SDRs, BDRs, whoever whoever is doing this sort of stuff. And when I'm feeling a little bit lazier, I can then ask Quad, hey. Actually, draft this in my inbox for me or or better yet, send this out. Right? This is great. Here's the email address. Send this, please. So, yeah, this is the first workflow that I want to show you guys. I wanna jump into another one that's probably gonna be better suited for account executives. Let's say we just had a demo. How can I take all the good insights that I just learned and then deploy that value in a follow-up email? So for this prompt, I ask, can you help me draft a follow-up email to a prospect sharing how we are better than Crown? Please refer to the deal tip I got from Klue to find the relevant talking points. Once again, this is probably all you need to do. I added in exactly where to find exactly where to find the deal tip, so it wouldn't show eight deal tips and ask me which one I'm choosing from and everyone I'm working with. So yeah. But, for for those of you who don't know what a deal tip is, a deal tip is when Klue, is picking up, hey. We just had a demo. A competitor was brought into this call. Let's not stop there. Let's actually analyze the entire transcript to figure out what's actually relevant to our prospect or customer. Right? Let's figure out through the discovery process what exactly do they care about in this deal. Once we have, let's say, maybe five key requirements, then let's go into Klue and map back to how, we're better than Crown in this case for each of those five key requirements. So going really from, you know, your typical why we win follow-up. Right? Here are the top three reasons we typically see. All of sudden, just saying, hey. You told me you care about these things. This is exactly how we're better at each of these things. So really bringing in that value differentiation. And that's exactly what I asked Claude to do here. So if we go once again into the brain, it's saying, hey. Let's use the exact email writer skill. Fantastic. And then let's start looking through the Gmail message. One piece of the deal tip actually that's extremely important that I didn't explain is that we proactively send it to your reps. So that could be through email or Slack, which is why I'm calling it to look through email. So instead of necessarily going into Klue, I'm saying, hey. Klue already drafted me this incredible deal tip. It has all the all the deal context involved in it. Go look for that email specifically and bring that into a follow-up email. So you can kinda see here it's searching through the Gmail. It found everything that I wanted it to. It's bringing up those key requirements, which we can see in the in the deal tip right here, and then it's mapping it, to a follow-up email. And then for this case, I did say, hey. Can you draft me or or draft this in my inbox? So I'll I'll show you guys what that looks like. Let me just jump over. There we go. Okay. Hopefully, you can see my email here. And there we go. So it took that that that email pulling from once again my writing style, the deal tip I got from Klue, and then turned that into a really good follow-up email saying, you know, not just why we win, but based on what you actually shared with us, these are the top things that you care about. Here's exactly how Klue is better than Crown in this case, and it's polished, ready for me to ship. So, yeah, kinda two different workflows that I want to show you guys there. One for the prospecting that really anyone who, is prospecting can use and then a a deal context specific one. Right? How can I immediately get value to my prospects post call and not worry if we're pulling from hallucinated information? I'm I'm I'm trusting that it's pulling from Klue, which is what I care about. Do you typically see. Zach, do you typically see some what are when we talk about the trusted data, some of the things that Klue typically pulls from that kind of powers maybe the powers some of the, like, point like, the points that it's giving you on that deal tab? Yeah. Yeah. And I think I think the beauty so I use this point all the time where it's like, you know, if someone was to corner me in an elevator and was like, hey. How are how are you better than Crown, for example? You know, I'm I'm immediately gonna go back to the top why we win reasons that I know. But the beauty of the deal tip is that, you you know, it's taking not not the top win reasons. You know, chances are, what I'm getting in a deal tip isn't what I would say is those top win reasons, but it's actually going back to that value. And, depending on what, you know, my prospect or or customer told me they care about, it pulls from the unique things within the Saks engine that you shared earlier, Adam. So, of course, you know, the top tracks that win out there in the field that we're hearing pulls a lot from win stories, from recent competitive won deals. If there's something specific to product or pricing that it can pull from those stacks, it's yeah. It it's been super beneficial. Cool. That's awesome, Zach. I love that you're really, like, experimenting with co work within this. I know that's, like, a new hot tool that everyone's been playing around with. I'm curious. If you were to try and find out all that information yourself, like, what to say, how to position it, like, how how much time would that take you in a normal day? Yeah. So I can go back to what I would do before, you know, we got access to co work and everything like that. And and even that, I'm using a little bit of AI, but it's still fairly complex. Right? So, I'm typically going into Klue, thankfully. You know, we work here. I I can go into Klue. But if if if we didn't have Klue, probably going into a slide deck or something like that. I'm probably looking through those top win reasons, trying to think about what my prospect cared about. Right? Like, the unlimited competitor tracking, pretty pretty good example. If they only had a few competitors, probably not gonna bring that point. So trying to do a little bit of mental math there, in terms of what's important. Pill picking out those key points, writing it in an email, copying it into ChatGPT saying, hey. Can you clean this up? I'm in a rush not speaking very prop very clearly, and then probably copying that back into an email. So, obviously, not not only taking, you know, probably fifteen to thirty minutes to actually craft that, but also having to jump into so many different tools, which which is a bit of a pain. Co work and being able to connect to different tools has been a huge, time saver. Yeah. Cool. Love it. And Zach Russell in the chat kinda teed us up for our next, speaker too asking, can we could you attach, like, customer evidence, something that proves these as well to that email? And that's exactly what our guy, Isaac, is gonna share as well for his email writer. Oh, he's muted. He's done the classic mute. For my job being purely technology, I'm so bad at technology. As Adam said, my name's Isaac. I lead AI ops here for our go to market team. Yeah. Let's jump into this. It's so it's my job here, amongst other things, to look for places where we can build things that are better than what we can buy. And it's not because, like, there's any special background I have. It's pretty simple recipe. It's basically just where can we add context that that tool wouldn't necessarily have. So today, I'm gonna show you what I mean. I'm gonna use an example that Zach already touched on. Now Zach uses his co ed co work to do things that are specific to him, but there's some things that are gonna scale across our whole go to market org, and so it's a lot more efficient to do them centrally. Also, there's just some compliance, elements. And as Derek mentioned, we also wanna be able to tie in some of that cool content that Adam and his team makes. I'm gonna show you how we do that. So let me just start sharing my screen here. Oh, no. Not slide. Screen. Alright. Adam, do you see my screen alright here? I got you. Everyone. in the chat, you can see it. Do we need to zoom in, or are we good? I can see it fine. Alright. So I'm gonna show you how the entire organization here does follow-up emails. So this is an example. Like, every organization sends follow-up emails when they have a call. So here at, Klue, we're a combination Zapier and n a n's shop. So we build these building blocks in n eight n, and then we put the Legos together in Zapier. So this is what happens every time somebody at Klue gets off a call. So first step here, we use Gong. So Gong sends a webhook to this workflow. We just use a simple little AI step to classify all the metadata that we need. We then load our AI configuration. So here at Klue, manage our AI centrally. So rather than have putting in, like, chat GBT 5.4, we put in we want our writing AI or our code AI. And then we update it centrally as new models are released so, like, we don't have to go through and update everything. Any changes may proliferate down through everything. So then we fetch the transcript. So this is the first of these building blocks that I was talking about. I'm not gonna walk through a ton of detail, but so we call this a little n eight n workflow that finds the transcript for that call and optimizes it for an AI. And I should mention here, for anybody that does want, like, the recipe, any of the templates, any of the information, just reach out to who whoever your contact here is at Klue. Happy to get on the phone, walk you through it, share anything that I've done, any of the prompts. So they'll set that up for you. They'll throw some time in our calendar. So cool. We've got the transcript. Next, we're gonna get the writing profile. And Zach mentioned this, in his part here. But every week, we map every person that sends out emails at Klue across about 300 different data points from how they sign off to how they challenge perspectives, how they present lists, even how they use grammar. Not necessarily, like, correctly, but just, like, how they use it. Zach Russell's I I always like kind of a funny one. The way Zach use emojis, very he seldom uses them, but when he does, he makes his smiley faces backwards. That, like, picks up little minute. things like that to make it feel like it's actually human. So then the next piece of context is the conversation dynamic. So there's kinda three core elements of context here that we're adding. The person, the conversation, and then clue, which is sort of the secret sauce on top. Again, another one of these little n a n workflows. Sorry. I got it there. And now what this does is it takes everybody that was on the call, and it looks at all emails that have been exchanged between them and Klue over the last three months so we can pick up on the dynamic of the conversation. You know, is it new and formal? Have they been a customer for five years? Is there inside jokes? Is there banter? That sort of thing. Then we quickly get our playbook, which is, like, our system prompt. We get any assets that are available to Derek's question. This is gonna include everything from case studies to, like, NDAs, basic documentation we have to send out, help docs, all that stuff. So we the AI can see what's available and decide if it wants to put something in. Now I'm gonna show without Klue first just so you get some contrast. So I've added this little extra arm in here so I could show without Klue. So, basically, it goes in, picks. Should I put in an asset? See what it says here. So cool. I wanna use the how Greenhouse built self updating battle cards, case study. Cool. Gonna use an AI to write it, another AI to format the HTML for an email, and we're gonna put it as a draft in my inbox. So let's take a look at it, see what that looks like. this is the one without clue or with clue? is this is this is without a clue. So. we just got a nice little email that sounds like Zach. I'm not gonna go through it a ton of detail, but you have this nice little email. Now when we go and do it with Klue let's go back over to this other one. Now what we do is we use an AI to take all that information that we now know, and we're writing a specific query for Klue MCP based on what this person is looking for. This is an actual conversation from Zach from last week. I've just anonymized the company the names, but it's this is coming from a real transcript. So cool. We take that, then we call clues MCP. I prefer not to do it directly. We, again, we use these building blocks so I can reuse them. I don't like, it's like, everything is kind of these LEGO blocks that gets assembled together. It's kinda how we feature proof as things evolve. We kinda sub in LEGO blocks or update the LEGO blocks rather than updating the full workflows. But just calls a simple little n eight workflow. 8x8 n workflow go into super deep detail, but, basically, it's just two paths. If I just wanna know what's available, I'll use this path. In today's example, I actually want the information. So it calls this bottom path, and it gets me a whole bunch of information on win loss stories, where we have strengths, where we have weaknesses, or sorry, where our competitors have strengths, where our competitors have weaknesses. Then we're gonna do the same thing again. We're gonna select an asset. This time, it's selecting a different one, based on the clue information. So it's looking at a Blackbaud case study. Write the email. We extract the HTML, and we put it in our draft. Now I'm gonna show you what the differences are. I've added a little extra formatting so you can see it. You don't have to read through all of the information here on the call today. So let's just let it load here for a second. Okay. So here is the new one. So I've added in some color so that you can actually see it. It corresponds to what it's doing down here at the bottom. So from the yellow, first thing is so we know from our loss stories, lot of time when we lose to crayon, it's because, it's too hard to unseat the status quo. So here, because we have that information from Klue, the AI is already trying to deposition the status quo, and you can see kind of the the explanation down here, by making the cost of good enough feel immediate. And so in this case, it's reclaiming time, and it picks up on the manual upkeep in the areas that they're looking at that wouldn't be manual within Klue. The other thing so if you look at the blue, now one of the other differences between us and Crayon, not good or bad, is just the pricing structure. So certain situations being more expensive in Klue, certain situations being more expensive in Crayon. So it picks up on what they have said in terms of their use case. I've used Jenna. I sub Jenna in for the actual person's name. So Jenna stated that they wanna add their tier two and their niche niche competitors. So they're gonna have a ton of competitors in there. Because Klue offers unlimited competitors, gonna be a lot less expensive for them to use Klue. So we're anchoring that in the follow-up email. Moving on to the pink. One of the other things that we have seen from a lot of our feedback from our win loss interviews is that in terms of just features to features, it kinda feels commoditized because a lot of them are very, very similar. So rather than trying to unseat somebody based on features that are perceived similar, we're, again, gonna shift away from that, and we're gonna look at adoption. Because if everything else is the same, the platform that gets better adoption is gonna be the one that drives the most impact. So we're anchoring towards adoption. And then this fire one, the green, now the green is actually not up there, but this is how this is why it picked that asset. Now this anchors better in a scenario that is gonna be very similar to what the person described on the call. So this is another one of our customers where we were able to unseat our competitor, and it calls out tangible metrics, that they saw. So, like, in a head to head comparison, this gives Jenna specific numbers to justify that switch internally, not just feature claims, which can kinda be easily dismissed by the the incumbent. So that is that's what we use pretty much every time there's a follow-up email at Klue. My many of you probably have received one of these not knowing that it was coming from an AI because we try really hard to make sure that it sounds human. But that is how we drink our own champagne. Thank you, Isaac. This? is this one's really cool. And for other folks here as well that maybe I personally get a little lost in the sauce once Isaac pops a chunk on an n eight n workflow, hit him up. It is a little bit simpler as he walks through it behind the scenes. We obviously can't take all the time there, but maybe Isaac drop you a LinkedIn or something like that in the chat, and folks can hit you up to see how they can put together that workflow and see exactly how they can build that. Will do. Alright. Thank you, Isaac. Jenna, who's next? Is it me or you? Which who wants to go first? We rock, up, next, scissors? Adam. You're up next. up next. wants to see your competitor comparison page. Okay. Give people. what they want. So one of the things I think someone mentioned in the chat, like, AI, where where they could use AI is to help with stale documents, stale assets, and that's a huge one I'm thinking about. I'm a product marketing team of one, and my competitive comparison page is so out of date right now. Like, I don't wanna I don't want people to go find and see it. It's way out of date, and that's just kind of par for the course when you publish these landing pages and you're a lean team. And so we're thinking through how what are different ways we could use MCP? What are different ways we could actually, like, infuse Cluedent into some of the work we're doing? That was one of the first things that came to mind for me. And so I'll pull this up on Claude. I've already preset some prompts so we can be a little bit quicker here for this. But let's let's hop on over. Which window? Going into we're going into Claude here. Okay. So I'm gonna zoom in a wee bit here. Maybe that's a little much. So what I wanted it to do was to do an audit of our comparison page, one of our compare comparison pages, our competitor's own comparison page on us, and then go into Klue and just review, hey. Are these the most relevant differentiators we should be hammering? What's the social proof that I can use within this page, and what how should I be changing or tweaking my angle? And so what Klue did here is it started to look through the competitive landing page skills. So I built a skill in Claude on that. It's gonna start to sort of analyze both the pages, both ours and our competitors. And then it queries Klue MCP, and it's gonna start looking through our cards on that competitor. It's gonna look through win stories. It's gonna look through loss stories. It's gonna look through talk tracks. It's gonna look through what buyers are saying about the competition in sales calls. So And that's what it was doing during that analysis side. And here's, like, a high level obviously, I've anonymized this for marketing purposes. But what is really interesting here is before even getting into the differentiators, they mentioned that this is not a structural problem. This is around a '22 like, a 2022 era competitive narrative. And spoiler alert, it probably hasn't been updated recently than 2022. And so that core frame even identifies just like, oh, that stuff that you guys used to hammer and win on back in the day, but you need to update. You need to be back where you're actually winning today. And so it went. It looked at the central argument that the page should make, and then it really hammered our key differentiators. And I'm a product marketer, I kinda know, yes, this feels directionally right. I'm sure Zach, our AEs, can look at us and say, yeah. This is some of the angles that really work. And so it what it would look at is, hey. Deal specific competitive intel. What's the claim? But what I really like isn't it's not just making stuff up. It's not hallucinating. It's like, what's the evidence that is saying that why this should be your lead differentiator in your comparison page? And it's gonna cite multiple win stories because sales enablement was way better than the raw data dump they saw with a competitor. When we assess this through Gong, we saw, like, successful deals closed when we mentioned this sort of generation of auto insights or win loss stories for sellers. And then we even again, I anonymize this, but a prospect mentioned that the there was such, like, a low adoption for sellers and deal specific intel that their Slack channel had more. So I'm just getting a lot of evidence from Klue to validate this differentiator and how we should message. Another one, which is, like, a no brainer that's not on the page, and I'm just kinda outing myself of being a bad product marketer right now. So so it's okay. But unified CI won last. We know that's a big differentiator. Jenna knows that, and it wasn't in our page. And what it's done here is it's saying, yo. You need to bring this up. This needs to be a key point of how you sell. And we have win stories. We have we have cards. We have Gong calls. They're all validating. People are picking because of this. Now I'm sure high competitors are probably watching this as well. But, yeah, this is one of the one of the ways in which we are looking at using the kind of current verified data on what our differentials are, what's working in deals to then work into our marketing collateral. So quickly I'm just gonna quickly flip over because it's done this audit. It's done this suggestion, and I'm sure lots of folks have kind of worked on this before. But what it does really well is it then can bring this into, like, a really mock up quick template that I can get to brand. Now this isn't perfect. We won't publish this exactly. It's not our brand colors. I'd probably wanna do a couple of tweaks, but it's pulled all of those sort of key differentiators and laid it out for me with, like, building blocks I can tweak. It's even starting to get some of the quotes that are with Include from our Win stories and things like that that really gets me a massive starting point in terms of building one of these things. I don't have to, like, go back and forth with brand, building all the wireframe, looking for those differentiators. Like, it's already there for me, and now I'm in a way stronger starting spot to produce an asset that always, always, always falls out of of date if you're a product marketer. So that's my use case. I'm not sure if other folks have that resonate with the landing page, the comparison page, falling out date the moment you publish it. But yeah. Jenna, I flip to, your Adam. Yeah. I I think I just wanna call out what's really interesting there. Like, that would have been such a manual lift for you to have to analyze that information. And then even before, like, you would have then had to go back into Klue and sort of said, okay, Klue. Like, here's what I learned. Help me think of differentiations, and then back to another tool that's gonna do the end output. Doing it all in one place, accessing our MCP server directly in the tools that you wanna work with is so impactful, and it's really it was neat to see the output that it produced within seconds. Yeah. I was fired up on that one. I'm excited. Jenna, you've got a use case, and then we have some awesome customers that are waiting patiently for. us in the waiting room. So let's get to. your example. And then, folks, if you've got any questions, get them ready, and we'll have some customer sharing as well, what they're doing and what they're thinking. Perfect. Yes. Let's get these customers on stage here. So I will be quick. Everyone can see that okay? Yep. You see my cloud, I? got you, Claude. Okay. Perfect. Okay. So here's a use case that I was working on. So, essentially, I was given a task. I'm sure many of the competitive pros sitting on this call can sort of relate to some of this. I'm new to doing compete at my organization, and I was just learning. And I needed to get up to speed on one of our competitors pretty quick. We noticed some trends in our deals. Leadership wanted a bit of a download on, like, what's actually happening in these deals? Like, what's going on with this competitor? How can we position ourselves? And, like, how can we start to think a bit more strategically about some of the tactics that we're using in these competitive deals? And I was kinda like, where do I start? There's so many different places, like, I could go to find this information. And, you know, I was like, okay. We're obviously working at Klue. I wanted to come to Klue first. And I said, I'm gonna I'm gonna try out the MCP server so I can play around with the end output a bit. So what you're looking at here, this is a prompt. I I'm a classic. I'll write the crappy prompt myself, tell AI to write a better prompt for me. So this is the prompt that Claude actually came up for me based on the project. I gave it a little bit of context. I told it to search our clue MCP. I was a bit more specific. I put you don't need to be quite this specific that it'll still pull the MCP. And then I asked it to figure out the findings, pull it into why we wins, why we lose. And then for the end output, I wanted it to be just a simple visual mock up that I could share with our leadership team. And then because I'm showing on the webinar, I also included a bit of details around, let's anonymize this data so what I'm showing isn't real deal data with you all. So I, yeah, I put that in, and I'm it takes a few minutes to load, so I'm gonna show you the end output. And so what Claude built for me is this live competitive brief, and it's all based on data that it pulled from Klue. I don't know how many of you have played around with our win and loss stories. I find those to be really, really powerful sources of information as well as I run win loss interviews as well. And so this is what it was able to distill very quickly for me. It looked at a number of deals. Again, not real data here, but it looked at a number of deals from Klue and started to pull out just specific win reasons. Something that's really hard when I've done these on my own in the past is getting quotes. Like, these can be I've searched through Gong calls. I've tried to find these clips. Klue already distilled these quotes for me in our win loss stories, and it pulled these directly into the actual findings, which I thought was great, and it linked it. So you'll see these will actually click. I won't open it, but it'll take you to the clue card that this information pulled from. So it's trusted information. It's backed up. You can take it to a call clip and watch the the prospects say that particular clip live if you're showing it to leadership. So this was, yeah, the report. It's, like, pretty specific into different patterns that we're seeing. It goes into where we're leading, where our competitor might have a little bit of an edge, how we can differentiate. And my favorite part is this action piece. I knew I wanted this to be around go to market strategy, and so it came up with five distinct actions that I could then use. And so I thought, like, a good one. Okay. Running a targeted campaign. Great. My next step would be, I would simply ask Claude, okay. Based on your recommendation, help me create a campaign strategy that we can go to market with. And so it's just really powerful way to pull in that clue data to just supercharge a lot of the work you're doing and a lot of the deep thinking research. So much valuable insights living clue and then bringing it to life through your LLM is a great use case that I've started to play around with. This one was so cool, Jenna. And I do wanna call out Jenna was we've been playing around with this, thinking about where the Klue data plugs in. And Jenna did this without using Klue data as, like, a test as well. And she started to dig and realized it was kinda making up a few things. They were made up, like, win stories and buyer quotes. The moment that she pulled that data out, the Klue MCP data was not being used, and it started to take a bit of a mind of its own. So I think it's always a good reminder, one, all of us here, like, be let's really review the things that AI is generating, but also when you have that kind of trusted data from Klue as well in comparison to some of the nice looking, but a little suspect outputs that that can come from it. Totally. The chest is huge. Alright. Should we get our customers or people that everyone's been waiting for on stage to chat a little bit here? Yeah. Let's do it. Let's bring them out. Drumroll, please. Welcome. Welcome. Welcome. We've got Jacques from 8x8 and Prathith from Sisense. What's going on, you two? How are we? Doing well. Doing well. This has been great. You know, you made a really good point just then, Adam, about the trusted data sources and the the AI hallucinations, and we've seen plenty of that when we're kinda adding in, you know, some questions and not going back to a good source of truth, which I think is a big part of competitive intel is having that right data. Let me hear do you have any stories? Like, what what is the moments you've gone through here in that in that regard? Oh, absolutely. Let's see. Where where do I start? No. I've I've done a lot where even just looking at competitive week to week and seeing, hey. What what is coming out? What are our competitors saying within unified communications or contact center? And, you know, as a practice, doing what I do, I actually like to read through the articles, let's say, that are coming out in the news. But even when you have that article, it can still hallucinate and say it'll pull out like you said, it will pull out a different quote. It'll make an incorrect conclusion about what that means. And I've seen that several times, actually. And it's it's worth noting where, let's say, someone just got acquired, completely put the wrong, dollar amount that was associated to it. And that's information that we need to kind of vet because it goes out to sales. In my case, lot of times, goes out to our executive team. So I have to make sure that's really buttoned up. So keeping that keeping the AI honest is a big big important part of it. Prathith, what does that look like in kind of your your your situation or maybe some of the ways that you've been experimenting, like the importance of the accuracy and up to date intel? No. Thank you, Adam. First of all, Jacques spoke about, you know, very important example, and you spoke about the challenges of, you know, updating information yourself. Right? You're you're swarmed with a lot of cards and everything, and you're trying to you know, the constants change. So what I think what I've been doing recently with the MCP server launch is kind of using Cloud Code and Core, extracting some of the outdated information into Cloud MD file files, kind of reviewing it, updating it with the power of AI, and also writing that information back into the platform pretty easily. So that's been kind of a game changer, to be honest. That one's really cool. That's a that's an interest one of, like, reupdating some of the assets that might be within Klue as well. Can you Preet, we talked about this, when we were chatting as a prep call here, but I'd love for you to share sort one of, like, sort of the first workflows or or things that you talked with as you were working with kind of the MCP server and some of the APIs. Oh, yeah. Sure. So as you all know, as most of the people here as product marketers, you know, we do the balancing act. Not only do we support the GTM teams, but we're you know, the onus is on us to kind of work with product teams to, you know, shape up the road map. I think what's been really cool with the win and loss stories is kind of using that information and to find out, you know, hey. Where are we winning and where are we losing? And what can we do to kind of, you know, shape up that road map so that we can, you know, win more deals? And like Jacques Bégin said, this information is completely grounded. Like, the we're not, you know, doing guesswork on which, you know, functionalities would will will yield actual revenue. We're using solid information, backing that up with examples so that the product team can kind of, you know, prioritize their road map because, you know, guess what? Everybody's building a lot of team things like you said. And, you know, prioritizing what customers want you know, require is definitely a key. What's interesting as well with when you kinda share what you've done there is not only, like, showing why you've won and lost, but I I if I'm correct, you were mentioning sort of, like, the frequency of how it was mentioned, like, the total ACV or revenue attached to maybe a a a a feature or an area of focus, right, in terms of what those, like, win and loss stories look like as well? Oh, yeah. I I wish I could show the data. It's just proprietary. But not only that. Right? Like, you you won't be even able to see the opportunities, the ACV. You can you know, with the power of AI, you can segment. Like, let's say, Klue has or any company has four or five product verticals. It also segments that into the different categories and helps helps people, you know, really analyze, hey. This is where we're losing, and this is this is where we need to, you know, come up as a team. And, you know, this is where, you know, we're actually winning. So how can we double down on where we win, handle objections, and how can we, you know, better our road map. And not only in terms of product. Right? Well, as a, you know, product marketer when we're, you know, developing battle cards for information, this data is kind of invaluable. And I think, like, what Zach used and and how Jenna was mentioning about the codes, then the actual info rooted information. I think it's a it's a it's a really cool day. Zach, on your end here, I I you mentioned sort of that sort of perspective of, like, hey. As a CI analyst, as a pro marketing leader, like, I I can even pick apart some of these news articles. These are really important. I'm sharing them to the right stakeholders. And you could pick apart maybe the acquisition was the wrong value and things like that. Have you observed or maybe this maybe I'm gonna give you a little bit trigger you a little bit here, but have you observed even other folks across the org that might be, like, looking at the un like, the unfiltered data or going to AI without any verified truth and looking for answers, whether it's your reps or other folks in the org? Yeah. I mean, I I think the default for, all of us right now I mean, we've been playing around with this for, you know, pretty solidly for about a year. But I think, you know, as anyone who uses it, you you you kind of feel that temptation to let the AI just do its thing because it does it so quick, and it sounds like it's correct. And I think what we've seen sometimes is even from a sales perspective, it's kind of magnified because there is a more, faster need for it. I need this information right now because I'm on a deal. So trying to get that, and if you're just going to a a ChatGPT or, you know, whichever one else you use, it it might give you some information that's either old, outdated, wrong. Sometimes it's stuff that hasn't been vetted yet, and it's just kind of out there, or it's just fully fully unmade up. So I have seen that. We're trying to rein a lot of that in. I think one of the good things is being able to pull that information centrally. The CRM data, whether it's what we have in Klue or whether it's something that we have, let's say, messaging from product marketing, and being able to marry all those together in a lot of different kind of steps and AI workflows so that it when someone goes and tries to do that self-service model, it's taking that part or at least mitigating how much of that kind of wrong information is gonna come out. I love that, Jacques. I think you touched on something really critical is how fast sellers move in those moments. Like, they're like, I need this answer now, and so I'm gonna just quickly search where to find it. And if they don't get that quick answer, like, they'll have to make it up. They'll have to divert. So knowing that in those moments, they can route back to trusted answers in the tools they wanna live in, that's, like, really powerful. For sure. I we've got a question here from Chloe, and it might actually relate back to Begin. What are the use cases you were walking through of this sort of, like, product, road map suggestions, areas of improvement opportunities to double down on. She asked, how are you using the CI Insights to then discover these net new customer problems and then use AI to potentially ship features faster? So maybe it's kind of this PMPMM relationship of, like, are there opportunities, or are you thinking about how to build prototypes, like, based on this quickly to get something in front of the team to build on? No. This is a phenomenal question. Right? Like, like, most of you all know, it's not only about new business. It's about I think she's she's talking about, you know, customer retention and figuring out, you know, how we can kind of mitigate some of the issues. It's it's it's a really cool question. I think we're we're we're moving into the age of knowledge. There's something cool that I've been thinking about using not only Klue, but also Gong and Salesforce APIs, you know, to build an application to figure out, you know, hey. What's going on with the the implementation? How are they onboarding? Are they are they are they are they kind of getting the feedback that they need? Is there is there a drop off? Everyone you know, none of the CSMs need to get a surprise, right, at the time of renewal. Hey. Something's up. So there's some like, you know, Isaac showed, there's some really the the the landscape of competitive intelligence is definitely changing. So I think the more and more PMNs who are thinking about, you know, building these prototypes and apps using Cloud Code, integrating these different APIs or MCP servers depending upon their availability and tech stack, and kind of presenting this information not only to leadership, but to customer success and sales as well. For sure. I think that's a cool thing as well. We would, like, a lot of folks I've spoken to throughout years of Klue customers and folks in this profession. Like, a lot of deep research needs to get done, and then it's like, how do you actually move that into action? How do you actually, like, create change? How do you start to get stuff on the drawing board? And, like, this is a great example of this deep kind of win and loss research product road map gap. And then you can mock up something quickly to sort of, like, move forward on, hey. This is where we could build next, or this is like, you kind of iterate and build off of that. It's actually something our own PM has been doing with with Klue, like, sort of testing these, understanding how customers are receiving this, or maybe there's some problems or gaps that we need to address. And then just VIBCO mock something up and then work with the product team of, like, is this the path we wanna take? Can we get this in front of customers quickly? So it allows us to learn very, very fast. I have a question for both of you about what's the next workflow you wanna build? What what's the thing you're excited about building next or, like, the outcome of business outcome you're trying to accomplish with it? So, Jacques, maybe I'll flip it to you first. Yeah. Sure. I think Prathith really hit on a on a good point there as well is being able to look at things kind of before they come up. I think competitive intelligence historically has always been very reactive. We're finally getting to a point where we might be able to take some of what we're typically doing and, let's say, outsource that, you know, via AI and then move in forward to looking at, hey. What is coming up? How do we get ahead of a customer renewal, for example? How do we know what's going wrong in a deal before that deal is even getting closer to that point? From a perspective of looking at what workflows, let's say, in the next twelve months that I would be looking to to use AI to automate, it's gonna be a few things. One of which would be getting that sort of executive level competitive intel and market intelligence, not just a snapshot of what it is, but how does that compare to what we saw a year ago or three months ago, and then trying to get ahead of it. What are we going to anticipate in the next three to six months, let's say? So getting that right now, that's a pretty time intent intensive process to get that all kind of wrapped up. I've automated, probably about 80% of it, so I kinda wanna move that even farther off and then start diving in. Like, Prathith was saying, like, can we jump into these deals, you know, sort of more proactively and get ahead of that? I think that is where we'll see the big change. To Jenna's point, it's always moving very fast, and I think it's just gonna continue to move faster. So as we can kind of lock down that basics of competitive intel and then leverage the AI to let us kind of go out further down, you know, whether it's down the pipeline or get ahead of the, the market intelligence as it is. What about yourself, Prathith? What what are you excited to build next? And, you know, that's always an interesting question. Right? So I think I I I used one of my examples previously, the Gong, Salesforce, and kind of the Klue combining into, you know, one sort of monster intelligence ecosystem. But one thing that, you know, that I actually, you know, got a light bulb and I've probably not used is it you know, it's actually very simple is using the deep research mode include to analyze data from the call alerts and everything and kind of building battle cards for the CS folks. So I think, you know, that's something I could be doing a lot more and possibly using that information. But, you know, to all the PMMs out there, I think that's a, you know, cool way to kind of use case to kind of prevent share and then think of it that way. But, you know, there's so much you can do, Adam. I mean, it's not only the information. Right? Like, imagine a seller like Zach is on a call, and you, you know, on a live transcript, we're able to, you know, surface information to him so that he's able to kind of answer that questions on the go. So I think we're heading into a new age, and I'm cool to see what interesting people like Isaac and, you know, other people and are willing to build with any day and Zapier or any of these workflow systems. So it's gonna be fun. Love that. K. I have a last question for you both before we wrap things up. Obviously, so much is changing in the way we're working and operating. I am curious to hear from both of you. What do you think today makes a great CI practitioner? That is a tough one. I think there is if I put it into two very simple buckets, it's you have to have the foundations of what good CI is, and, I'm actually gonna piggyback on what Chloe's saying in the chat right now with research being passive. I think it always has been, and it was very time intensive. And I think that is that's changing a lot. The other, you know, sort of skill set that I think someone in this role needs is you need to be adaptive. Things are always changing fast, particularly in the in the tech industry. But with what a is AI is able to do is take some of that sort of passive research, which was very time consuming, and speed that up to a point where you can start jumping into how do we actually affect a deal, for example. How do we, let's say, our company is going to acquire someone else. How do we make the best decision with that with the market intelligence? So it is actually kind of branching off. I think the fundamentals still need to be sound. You still need to have good data. You need to have a good research focused mind. You always need to understand what sales is doing. That's that's very foundational. But then being able to leapfrog forward and take out that kind of, let's say, drudgery or that that heavy intense, time and start focusing on that just in time, just in time for sales, just in times for renewal, and really go beyond just that immediate competitive intelligence into the broader organization, essentially, which is I think was always the goal of competitive intel was to have it across the entire organization, whether it's product or the GTM teams. Totally. Scaling that reach, you're still the expert in the loop. Yep. Jacques said on some of the, you know, points pretty well. So I think there's a reason why we, as humans, still exist, and it's not AI doing this webinar for us. So I think what's critical for for us is, you know, CI professionals or product marketers is business strategy. Like, ultimately, like, I've been in my org for five years, and I understand the analytics place pretty well. So how much do you understand about, you know, the business that you're in right from pricing or packaging and, you know, what buyers actually need is gonna be pretty crucial. So even as a seller, thinking of, you know, putting themselves, you know, in in the shoe of who's gonna buy. So that's definitely critical. So business strategy and kind of decision making, critical thinking, all those things that we read in a book and we don't think it matters, it's gonna matter a lot now. So yeah. Love it. Well, thank you, Prathith. Jacques, we've just come overtime. Folks, if you stayed a little over, appreciate that. If if folks wanna reach out to you, Prathith and Jacques, where can they where can they reach you? Where can they contact you if they got questions coming out of this? Oh, yeah. For me, probably my LinkedIn is probably the easiest way to find me. I'm, usually pretty responsive on that, so that's an easy way for it. Me too. You can hit me up on LinkedIn. I have a different name, so it's easier to find me, or I can drop my, you know, link into the chat. So, yeah, happy to chat. good. Well, I appreciate both. of you kind of sharing your expertise, where you see both. this going. You guys have been awesome. that. And thank you everyone for being in attendance, being active in the chat. I hope there's some fun stuff that you can take yourself. We're gonna be building more and more. I think Preet said it. Kinda the limit does not exist on what we can build on top of this accurate data, and we'll be coming out with plenty more. So recording will be sent out, probably some more content, and stay tuned. Share with us what you're building too. Thanks. so much, share. and enjoy all next. time. K.