Subscribe
Share
Share
Embed
AI in go-to-market is mostly bullshit.
In this episode, Toni talks with Koen Stam, a senior sales leader at Personio and creator of GTMOS. They break down where AI is actually helping in sales, marketing, and customer success, and where it’s falling apart.
They talk about why “AI SDRs” don’t live up to the pitch, why outbound emails generated by AI are useless without context, and why most teams don’t have the data or workflows to make AI perform. Koen shares how he uses AI at Personio for tasks like research, call prep, and churn prediction, and why the fundamentals still matter more than hype.
Visit Personio: https://www.personio.com/
Read Koen's Substack: https://koenstam.substack.com/
Learn more about Attive: attive.ai
Visit Personio: https://www.personio.com/
Read Koen's Substack: https://koenstam.substack.com/
Learn more about Attive: attive.ai
- (00:00) - Introduction
- (04:04) - The Importance of Context in AI
- (07:14) - Challenges in AI Implementation
- (10:47) - AI and SEO
- (14:56) - AI in Sales and SDRs
- (23:06) - Proactive AI Assistance
- (25:43) - Balancing Automation and Relevance
- (28:10) - AI in Sales and Call Coaching
- (31:35) - The Importance of Practical AI Applications
- (41:28) - Proactive Churn Prediction with AI
- (48:05) - Wrapping up
Creators and Guests
What is The Revenue Formula?
This podcast is about scaling tech startups.
Hosted by Toni Hohlbein & Raul Porojan, together they look at the full funnel.
With a combined 20 years of experience in B2B SaaS and 3 exits, they discuss growing pains, challenges and opportunities they’ve faced. Whether you're working in RevOps, sales, operations, finance or marketing - if you care about revenue, you'll care about this podcast.
If there’s one thing they hate, it’s talk. We know, it’s a bit of an oxymoron. But execution and focus is the key - that’s why each episode is designed to give 1-2 very concrete takeaways.
Why Most AI in your GTM is BS
===
[00:00:00]
Introduction
---
Toni: Today, Koen from Personio joins me to separate the BS from the real. When it comes to AI in go to market, we'll cover real use cases and explore why AI truly adds value in those situations. In other news, Attive AI is coming out of beta. Today. We are connecting all your go-to market data into one intelligent workspace to help lift everyone's performance on your team.
Take a look if you want to and now enjoy the show.
Koen: Everyone wants to have this shiny, uh, feature whatsoever, but we need to understand what is the real problem and what is the fundamental under that problem, and how can we solve the problem? How can potentially. AI help. Everything comes down by, is it trained well on the right dataset?
That's not only for ai, that's also for you. Me as an SDR, you need to continuously feed it with what good looks like, what good looks like, like those kind of coaching, AI [00:01:00] assistance, et cetera. They can help as long as they prompt proactively every day pop up like, Hey Tony, I listened to your call last night.
Like, um. Do you want to get feedback on this point or the, as long as if it gets back to you proactively. I'm a big fan,
Toni: so I have Koen back here on the show. And, uh, today we wanted to talk about, well, I feel like there's lots of anxiety out there about using AI and go to market and. I also feel more and more people are, mm, I wanna say either calling BS on it or don't trust that there is something or, or whatever they might think.
And um, today I want to have the chat with Koen on, you know, what him and I both see there, out there in the market that that works or doesn't work. Um, and I think, and this is the tee up here. I think what's quite interesting, uh, K you can kind may kind of say that yourself, but obviously you are a senior sales leader at person, so you are living some of that reality yourself every day.
[00:02:00] Um, but also you do promote, um, the G-T-M-O-S right? Kind of maybe, you know, talk about that in a second, yourself a little bit. There is a little bit of like, Hey, there's some cool AI shit going on, um, and let's promote that. Right? And same, same in my own situation. You know, I'm building a company that is around AI and go to market and, and people promote and sell things, but I'm also advising founders and kind of, you know, not directly, but to, to large, we need to live through my decisions there a little bit more.
Right? I think the, the idea of the session here today is a little bit like, um. What's the, you know, what's, what's the difference between, um, the dream that you and I sometimes are part of selling and what, and, and, and what we are actually living in reality. Because I think that that juxtaposition can help some people to sift through what is, what is the, the, the CO and Tony sales job, um, and what is maybe reality, right?
And I think that could be really fun for people to listen to. But Kun, [00:03:00] before I go further in kind of all of that stuff. Tell us maybe just a little bit more about G-T-M-O-S and what you're building there.
Koen: Yeah, maybe first before we jump into G-T-M-O-S, I think we fundamentally need to make AI not that complex and not just so over hype.
Right? And this is also, I mean, I'm do G-T-M-O-S, but like. It's just nothing else than like, what is the problem? Do we really deeply understand the problem and where are we trying to solve for and, and, and, and make it very, make it very simplistic. And I think that's like where, where, where everything goes wrong, when it comes to the hype.
Everyone wants to have this. Shiny, uh, feature whatsoever, but we need to understand what is the real problem and what is the fundamental under that problem, and how can we solve that problem? How can potentially AI help? And I think that's going where it goes wrong. And just as a bridge to our GTOS. Um, it's nothing fancy.
I mean, for three years I, um. As you know, I like to educate myself. I need to educate myself as a leader because I have all different kind of [00:04:00] cross-functional teams, whether it is SDRs, AEs, account managers, customer success at Personel.
The Importance of Context in AI
---
Koen: I mostly as well, like, uh, chap in, um, pavilion, and I found it very important to share among my learnings with my peers and vice versa.
And therefore, I educate myself on LinkedIn because I'm active on LinkedIn. But I, I read and learn and engage a lot. I do the same thing on podcasts. I do the same thing on newsletters. Yes. I also do this stuff, but of course there are way better experts and more experienced or different looks at go to market, which I want to collect.
I want to make sure I collect those insights. And for years I was struggling. How can I make this? Relevant. How can I make this impactful and contextualized towards you? Tony and myself, Kun, were both different go to market leaders and therefore AI came into the mix and I just store all this kind of stuff in databases and actually let AI put it into micro learnings.
Um, and whether this is generative or eventually agent ai, you can build any kind of really fancy workflows to [00:05:00] leverage that information for your specific problem. And that's actually what I'm trying to do with go to Marketto. It's still quite complex, but if I want to translate this into, I want to help you as a go-to market leader, or I want to help you as an AE or an STR or an account manager, what is your best next step when it comes to go to market?
And that's what I wanna do with all those learnings and to AI in your context, but we will get to together in a bit. Um, but yeah, so, and I think the, the task here for today is
Toni: we're gonna go through. You know, I think disciples will say that the bow tie, but just the go-to market process from start to finish, um, and see where there are problems where I can help because I actually think the issue is still a little bit, it's very easy to find problems in the funnel.
I think that's super easy. I think what's really difficult though is, you know, learning and understanding where AI can help and how. Be that's, that's still so ambiguous and, and it's, in the [00:06:00] beginning it was well everywhere. AI is gonna replace everyone and everything and we are now learning more and more that that's actually not really the case.
Right? Let's maybe call out bs like things that I said a year ago, even six months ago. I will say like, you know what, actually we tried this out doesn't work. Um, and let's kind of go through the list. Um, I would propose goon. We start at the very top of the funnel on the, on the very left side of the funnel, which is kind of the, the marketing area.
I can go first. You can go first. How do you wanna, what would you wanna do? No,
Koen: go like, go ahead. Like elite, let's look at the entire customer journey or, we both are lovers of the bow tie, but like, because that's the thing. There are so many elements and every. Maybe things really narrow in on this AI SDR thing or the, the vibe coding.
Uh, with lovable, yes, but let's need break that down into the funnel 'cause there's so much more to explore.
Toni: And actually kind of a good point. Um, I think there are a couple of other areas in the business where IS found a really strong fit. Right. So you just [00:07:00] mentioned lovable and vibe coding and, and generally speaking, coding.
Right. Uh, I'm, I'm, I'm quoting now using Cursor, like, isn't that insane? It's completely insane.
Koen: But I, I still wonder here, like, and that's also a thing, and again, we need, uh, when we talk about bs we need to call this out.
Challenges in AI Implementation
---
Koen: If you want to get really good results with ai, it is a tremendous investment of time.
And the question should be, should you and me in our core job, who we, who we are, and what we need to accomplish? Do we need to do this? That's like a thing, like everyone wants to be this, this, this, this, this coder. Now all days, I mean, for two and a half years I pay an a, an AI engineer on the side to get that done so I can focus on, on the go to market matter.
I think that's also, we need to be honest to watch each other, like
Toni: do,
it's a hundred percent,
but also my position is a little bit different. Right. Kind of entrepreneur, founder, small team. I'm picking up little jobs here and there. It's not that I'm building the product, like God forbid, right? I'm kind of doing some website stuff and that's kind of easy actually.
But, uh, just [00:08:00] saying kind of, I think on the, the, the coding side, AI has like definitely found, let's just call it product market fit and more than that, right? Kind of things have been blowing up there. I think another example that seems really good, and we can go into this touch and go to market a little bit more is the support side of things. So, especially in B2C, right? Um, a lot of chat bots and so forth have become a lot better with ai. And then I think the third thing that I'm seeing works apparently extremely well. It's the whole legal space. So there are a couple of legal, um, AI companies now that basically are more or less, uh, you know, replacing a lot of the work tasks that come in the legal
Koen: process.
A hundred percent. And I think what is important here is asking ourselves, why is that the case? Before we go into market, but why is that the case? Why is a support use case or a legal use case, why do we think that is an impactful use case, which works?
Toni: I think the answer is there's a wealth of written and documented text information that the [00:09:00] I was able to learn from.
And in the case of coding. I was even trained on that stuff. Right. So when you, when you, when you think about it, um, you get an AI chat bot and then you teach it your chat history, like that's one thing. And now has that in, I dunno, like a rack, a context space. It has that in their memory, right? And can, can access that.
But with coding, the reason why it's kind of 10 x better with, with some of those things is because the AI was actually trained, you know, on that data set, the thought was locked up. Like now it's done, now it's released, and now it can memorize your stuff. And I think that's really important for people to understand that the training bit and the memory bit, they're kind of, they're different.
They're separated. Um, and while it can probably deal really well with your text-based stuff, it doesn't mean it's super optimized for that actually. Right. But that's, I think that is the reason, right. And, and when we say coding is, is text-based, well. There was a, there was a [00:10:00] website called Stack Overflow that basically completely imploded where developers were sharing their code, um mm-hmm.
On like specific snippets. And this is one of the sources that AI was basically able to train on and understand, like, you know, how does the coding thing actually, and on code is nothing else in a language, basically.
Koen: And I think that's like an important matter to take along. We start on the left side of the, of the customer journey.
It's about the rich. Context, content and context, and I think that's really, really important and there are more important elements, but this is like one fundamental where it goes also wrong in so many use cases because the context is. No quality data, and that's a really big one, but especially at, at the start of the funnel where sometimes not all the data is really well stored in the CRN to be used.
Yes.
Toni: Let's talk, um, marketing real quick.
AI and SEO
---
Toni: Um, I don't think you and I are the super experts on this, so I think what, you know, there were, there were some companies in the beginning that were really good at, you know, writing copy. Now Che BT does that also, and Claude does that extremely well too. Um, that was kind [00:11:00] of useful in the start.
I think all of us. Starting to get better and better and better at like, ooh, this is, this doesn't seem genuine content. This seems a little bit AI sloppy. Um, versus kind of what's the real stuff, right? We're kind of learning that it's really good at, I wanna say polishing and rephrasing and structuring thoughts, but it's not so great at coming up with genuinely good thoughts, actually,
Koen: right?
And again, here, because it doesn't know the right context. And it, and, and let's be also honest, and that's like the entire thing for all of us users or, or founders or revenue users is like, the question we are asking is super generic. And I think I always, I always trying to compare this with like, us being SDR naive or an account manager or leader is like the more specific question we ask in our context towards your context, that really matters.
So the more precise and specific and descriptive we are there. Of course your output will be way better. And then again, the [00:12:00] context database on the other end. And that's also still where it goes really wrong.
Toni: Let's just say what I think what works well with ai. And that's kind of almost not super correct, um, in terms of like that's, you know, that's a good AI application.
But for sure we are seeing that, um, sources like Chet PT and Claude and Gemini. Um, assigned to be big channels for websites. Um, um, we saw this, uh, previously in terms of e-commerce. Yep. We've seen this more and more in terms of B2B, uh, there's really strong advice out there now, how you can start ranking in this whole AI SEO space.
There's a, there's a craze of investments going into, I think, 10 or 15 different players that wanna. The new EMR and help you to figure out, you know, how you can rank higher and tracking all of that stuff. So there certainly seems to be something that's working and that people need to start probably investing in.
It's like, how can we make sure you know, when someone is asking a relevant question on chat, [00:13:00] GBT. That we are somewhere in this word cloud that the LLM finds and then kind of, you know, presents. Yeah. And, and again,
Koen: I find it also important for all of us because sometimes we need to stand still because AI goes so fast.
But if you, as a founder or revenue to have any kind of question in your private life, I can bet you, you most likely, I mean at least I do. Any of my questions, I go to chatty bit your cloth and not towards Google. If buyers are doing the same and they do the same, however they upload their entire list of requirements or what have you, again, the contact, which is so important.
So rather make sure that in your contact on your website is really picked up by, by those new browsers because they are the new browsers. And again, that's like a very underestimated element if you don't, with your content team and with your marketing team really, right. For eventually, uh, AI browsers to pick it up.
Toni: That's another thing by the way, and I think that will probably continue it. Um, people are putting text and information on the website, um, that isn't actually being [00:14:00] displayed. It's just then the code for the AI to read it. Uh, because they know, like, um, if you put anything on your website, it should be short, it should be bulleted, it should be like super understandable and everything, right.
Uh, because we humans, we just lack the attention. Well, for the ai, the same problem doesn't apply. The, the richer the information, the more you know the, the more stuff you can add into it. The bettered, we'll be able to tailor a specific request from someone on chat GBT to what they're actually asking for.
Because to your point, right, it's like, Hey, give me an HRIS that works in, you know, Holland is best for the, you know, the, the SMB market or middle, uh, middle stunt. You know, has these and these features can do this and this stuff, and then, you know, you want to have chat VD be able to serve as person, basically.
I'm not sure if I'm, you know, paraphrasing it correctly, but that's kind of the idea. Right. Other things in marketing, Kun, what do you see anything where that's where AI is about from?
AI in Sales and SDRs
---
Toni: I
Koen: think eventually, and let's be honest, if we are [00:15:00] looking. We have on our website or in our CRM, whether this is in HubSpot or Salesforce, I know we, we should not name, name brands, but like these are I think nine of 10 times used in, in by startups and scales, et cetera.
But like there's so much data or maybe let's, let's put it really ly like poor data into those systems and a lot of information which we are actually missing, which is out there online. There's so much information out there. 'cause it can scrape even LinkedIn profile, actually it can go over the entire web.
One of the key things where we also look into is, let's call it the research agent, right? Yes. All those intense signals and how I, I'm just sometimes imagin and that's also part of like one of those use cases which I'm just doing myself now. G-T-M-O-S, I'm just advising some, some founders and some sales teams.
There is like, if we have the context of the CRM. Ideally, um, we go out in a web and we pick up any of those signals. And again, those problems need to be very [00:16:00] specific and we can pick, uh, pick, uh, information, um, as semi, uh, clay or all those other tools who are yeah, because they're actually doing the same thing and we can enrich, um, our CRM or we can enrich the preparation for the next call, et cetera.
I think these use cases are highly underestimated and especially. Um, if you triple down eventually later on from like top funnel to middle funnel, when you restart as a human within a startup or scaler, talking to a prospect. But that can be, of course the agent can be ai, it can be the human. If we have those data insights, which are not that difficult to collect, but you want the AI to run this 24 7.
So when I'm Monday morning, I have those insights, which I can. Stake from my CRM, which is my still single source of database. Truth that that doesn't have any other other function anymore. But it preps me really well before I move into my next meeting. And I can do, and I always call this here [00:17:00] like, and that's also what we need to really think of.
Like it won't. Replace an AE or or persons, but it can really help you becoming this one, five or 10% better. And that compounds over time if you leverage those data, data inside. So, and that's like also how I look at ai. It's like, no, it's not like, oh, it has all those signals so it can run my outbound campaign.
And, and, and there we have the ai SDR. No, it helps.
Toni: Uh, I think you mentioned a couple of really, really strong points here. Um, I'm just gonna try and kind of pull them, you know, apart a little bit. You know, let's just linger a second on the AI SDR thing. I think this blew up a lot, uh, in the, in first positive and then a very negative way over the last 12 to 18 month.
Um, there certainly is a category of software that's been created, like certainly there's something here. Some people have success with this. Um, and you know, if, if you think about what worked well, what didn't work well, I think what is working well? All the [00:18:00] tasks that are, um, you know, much easier for AI to do than for humans, which is basically reading a bunch of texts, a k, a research, whether that text is on a website.
On a news side or any CRM on a call doesn't fully actually matter. It's not like every SDR is sitting down and listening to, you know, the past opportunities that they had and kind of the calls and reading through the emails. It's just not gonna happen. Right. You don't, don't talk about, there's also no time for it.
Right. So, so that's, that's what I'm saying. AI is really good at that stuff. Like, you know, give me a bunch of text. I'm gonna summarize it in the best perfect way for you. Like really. So the research piece, I think that's, that's certainly something. Um, and then, you know, people were thinking about, Hey, you know, it can write emails really quickly and really well.
I think that is less so proven at this point that that was a great idea. I'm not sure what if, if you guys had persona kind of using something like this, but yeah, I mean, the writing part is kind of a iffy, isn't it? It comes
Koen: and, and, and, and I, I, I'm, I don't have [00:19:00] all the answers here, right? Like, it's only from the experience which you learn yourself and the teams, et cetera.
But like, I think again, it comes back towards this context. If you really have the right information and I have direct access towards my goal recordings and all the transcripts, I know all this information and, and I think if I had that information on itself, a follow up email. Can be done quite okay. A, a a, uh, initial thirst outbound email.
I think to, let's be honest, because that's also what we need to, we need to be aware and call out, right? Like everything comes down by, is it trained well on the right dataset? That's not only for ai. That's also for you and me as an SDR. Like, let's be honest, most of those onboardings are not perfect. And, and, and, and, and onboarding is done and we think we are done.
No, it's a com continuous iteration. I think Jason Lemkin is, uh, one of those examples I really like, like how proactively shares like what is building with, with disaster ai and [00:20:00] he is. Continuously training his outbound agent. And I think that's the thing. You need to continuously feed it with what good looks like, what good looks like.
And that's like, and, and there comes again, like the timing investment. We as humans, we think like, Hey ai, let's put all the information in and let's put it out and Oh, it's fine. No, I can get you, like I, I, I write on every Monday morning my newsletter. Every Sunday night, actually last night towards the 11 I wanted to go to bed, but like, no, I don't like this.
I need to iterate. I need to iterate, iterate, iterate. Need to iterate. No, it still doesn't sound like me. And still, because I also use AI there. And, and I think that's the thing. If you give it the right context and you really have the patience to train what good looks like, um, always still that think the sort of fortune ka Dorsey, but like that is really the what makes the difference.
And if you do that continuously, consistently, I don't see why, why ISDR can't work. Not only ai,
Toni: it's human in the loop. You know what's also really important to, you know, we talked about [00:21:00] why is AI good at certain tasks, right? I think what's actually really key to just keep in mind is none of us are actually training ai.
I just kind of be clear on that. No one is actually training the AI itself from our side. Right. Kind of the training has finished. When it's released. You can ask the AI kind of, when was your training done and they, we'll give you November 23rd, 2024 or something like this, right? Yeah. What we're doing is we're managing context.
And what we are kind of sometimes then doing in order to achieve that is we are adjusting the prompt a little bit and we are adding something. We're taking something away. Um, you know, I was, I was listening and I'm not by no means an expert on this, but I was listening to a podcast that apparently is very much a deep expert on this.
And they were talking about, it's a little bit like. Asking a student to come into the room and play the saxophone. Um, and then, oh, you didn't, you didn't quite hit it. Let me give you a new set of instruction. But instead of this being the same student, you hold in a new student, like a clone of that person, basically.
Yeah. Yeah. Give an updated, you know, set of [00:22:00] instructions and then they have them play the saxophone and they basically saying like, yeah. I think this is really difficult to achieve and this is, I think where humans, you know, because we have this ongoing training kind of still works out. I wanna move us ahead a little bit though, Kun.
So one other thing that I've seen that works extremely well and still talking SDRs here is, um, mock cold calls. So you are basically as a human during onboarding or training or. Ongoing training you are calling an ai. Um, and um, you have like an interaction with it. Um, and then you will be rated on this afterwards and, you know, sim similarly.
So, um, after every call, call every connect, everything that you could actually record. You get some feedback from an AI kind of into your Slack or wherever on like how well you did, right? I think some of those immediate like feedback things. I think extremely powerful. Honestly. Honestly, really powerful.[00:23:00]
Your manager couldn't listen to all of that. Forget about it. It's not gonna happen. Right. And like, I think that piece is really cool.
Koen: No, no, no.
Proactive AI Assistance
---
Koen: It, it, it is the only thing I want to call out and just to be open and honest, because like, this is, like how I look at it is like where I see the biggest challenge and struggle with AI is it's always the human proactively from themself.
Need to go to rt, AI to ask all those questions. And I think in order, because eventually AI is a hype, like with a lot of fancy tools, like you are, ah, this is fancy. I also want this. And then there is this hype at the start and then you see adoption fading away. And why that happens is it is this threshold of getting daily back into that matter, which makes it really difficult.
And I think that's like where any vendor on AI and we as leaders need to think of and we want eventually our people to really leverage on a daily basis. It sounds weird, and I think like to your point, like those kind of coaching, AI assistance, et cetera, they can help as long as they prompt proactively every day pop up.
Like, Hey Tony, [00:24:00] I listened to your call last night. Like, um, do you want to get feedback on this point or the mm-hmm. Because otherwise the danger I see is that this, A-E-S-D-R or whoever wants to be trained is not proactively going back towards that tool. So as long as if it gets back to you proactively, I'm a big fan.
Then the second thing is still where I see, because I, I asked it a lot when, when we also had community events and said like to leaders. It's like, Hey, who is talking to their ai? Are you talking to your AI on Tony, like texting or talking?
Toni: I'm texting. I'm a texter.
Koen: Why are you a texter? What is faster? How can give more context to ai?
Toni: I dunno. I'm a millennial also. I don't, I don't take phone calls. I text with people, you know, that's the same relationship I have with the ai. We, we have, but I, I know where you're going. Same age. We're at the same age. I know. Thing is, I, I, I tried it out. I dunno, maybe I selected the wrong voice or something like this, but it just sounded like this.
I didn't wanna talk to that person basically, but
Koen: like, but like, that's what I decided. No, it is. You can change voices. Uh, I would not recommend uploading my [00:25:00] voice, but like I can imagine like you can also, uh, eventually maybe pick the voice you like, which sounds like, Hey, I really want to get. Uh, coached by Tony and I just upload Tony's voice and continue to have Tony's voice.
But that's like the thing I wanna mention. Like, yes, if you as a person find it difficult to talk towards ai, and let's be honest, Siri, we never got like a very good relationship with Siri and Apple. You need to across that, that, that, that, that, that bar, because otherwise also these kind of tools are not going to help because you need to feel okay to have that interaction.
I feel fine to have this interaction with ai.
Toni: You had one point there, um, building in this di you know, in this, in this area. Um, and we're seeing real user feedback one way or the other, right? Kind of. And that, that informs me on like what I see is going well, what's not going well.
Balancing Automation and Relevance
---
Toni: Interesting learning on automations.
And this is kind of a little bit, you know, still maybe too early. Um, but we are actually seeing that you need to strike the balance really well in, Hey, this is just too much. And it becomes pages of stuff that [00:26:00] I'm skipping over and it's kind of slop like it's, it's good quality, it's trustworthy, you know, all of that stuff is correct, but it's just not relevant for you in that moment.
And I think especially for call coaching, I think that actually is. Um, that can work, right? You basically kind of summarize all the calls in a day, and then you get, um, then again kind of the update. Um, what we are also seeing a lot though is basically asking the same questions. And this is, you know, yes, the, the, the person needs to change their workflow.
And that's, that's always difficult to totally see that. But when people ask questions, uh, whatever comes back could be the exact same output as an automation. But it's way more relevant. It's way more specific to their pain. They kind of get to ask follow up questions, kind of go deeper into that. And we are actually seeing that.
Um, my, my whole point is not, you know, promotive. My whole point is like there is a balance between what you should automate. And what, what should be ad hoc and, and very kind of situational specific. Right. Kind of that's, [00:27:00] that's I think the insight I wanted to kind of get across.
Koen: I, I really think on this point, and then we can move on from like, let's say the still like the top and middle funnel, but like, again, it's this context like if, if you need to coach me.
You need to give the, the, the, the, the, the AI tool, whatever the context of, Hey, Kun, this quarter, and we've talked about this I think, what about one half year ago in one of, uh, one of your earlier shows like on, on coaching with ai. I mean, I built all those coaching framework, right? Yeah. And I want to make sure o Kun.
As a leader needs to be coached on those two, three things this, this quarter, and let the AI only focus on those two, three things because that's the thing it, and that's also with ai, everything, and is also not my strongest thing is it gets overwhelming, it's got too much and if it's too much and overwhelming.
No impact. Jesus. So, and that's the same with ai. Where do we start? We need to really break it down and really break it down towards one or two things. And that's it. Be specific. And if you're not specific, it's overloads overwhelming, won't have any impact. I think that's also very important learning.
Toni: But I guess the, the, [00:28:00] the coaching piece does carry over into the ai, uh, into throw the, uh, ae kind of Yep.
Realm. Yep. Right. So what I'm seeing so far.
AI in Sales and Call Coaching
---
Toni: No one is really seriously talking about replacing AEs. No. There's a little bit of talk in the s and b space. Yep. And, and I totally get that it's, you know, where, where someone is only a little bit better than a, than a chat bot previously. I think in those cases and, uh, kind of an AI AE can, can take over.
But then also what I'm seeing, and I think you and Nu and I were kind of chuckling about this, what it feels like in the AE space and similarly so for the AM space, um, is there are a bunch of mini use cases. There's no banger use case, you know what I mean? It's like on the, on the SDR side, you have the research piece and maybe the outreach automation that seems like banger use cases for the AEs.
There's a lot of small things. Um, I would love to kind of hear, you know, what you're seeing there before I jump in, but like, it seems to be lots of small things where people are a little bit lazy on [00:29:00] moving on that really
Koen: the question is also, and I I, I, I hear you saying this right. Why are we always looking for this big next thing is that needed to make impact? I, again, getting back to simplification and specific use cases, I mean, I'm trying to always break it down into phase of the journey. And if I, if I talk about, let's say the SDR, like, um, outbound or, or, or discovery calls whatsoever, you always have to deal with a certain preparation, execution, and a follow up.
And it's the same for nae. It's preparation, it's execution, it's follow up. Yeah. So first preparation, AE. I can help the e with exactly the same piece of information. And, and you know how I also look at those, uh, databases, which I've built, where I have the best tips and tricks on how to do a demo, how to, um, how to convince a CFO, et cetera.
If I have eventually the right context for the upcoming demo demo, which I've, um, which I've scheduled, and I could provide you, Tony, as an AE with all that information so you [00:30:00] feel more. Confident and you know what kind of questions you can ask the CEO and the CO on that next goal. Um, those, those, those, those tinier details maybe doesn't seem big, but I can tell you the impact you have on that one demo.
And I may hope you have three demos a day, 15 a week, multiply by five people in the team is 45. Should we calculate the impact of that use case by preparing you better and so you are executing better. So there I see a potential. And then eventually, and, and what I like, I mean, I'm a big fan of, uh, digital sales rooms, for instance.
I'm only working with my teams for digital, digital sales for three years, and also there you've seen this tooling because we shouldn't always look at AI or JT PT, clot and Gemini. You should also really embrace your current tech stack. Consolidation, yes. But your current tech stack and see what are their use cases, uh, which they are adding, uh, with AI and for instance, um, eh, our current vendor with, um, with digital sales rooms, um, they also really, they call it the CRO in the pocket.[00:31:00]
So continuously after you have done a, uh, conversation with your customer, it also helps you what kind of information you need to send, watch which person. Uh, and, and it suggests those follow ups and I again, then I think from a. Preparation perspective so you can execute in your demo or prep a negotiation call or eventually, uh, afterwards, which you'll follow up.
Whether this is, uh, an email or video or whatever. I think this is actually very, very impactful. However, we are looking towards the fanciness, but we should look at the practical implication and impact. If you ask me,
The Importance of Practical AI Applications
---
Toni: I think the reason why fanciness and big impact, uh, usually helps is kind of a turning the tide kind of moment, right?
Kind of there's this saying, Hey, you need to be 10 x better than the previous solution for someone to move on and adopt you. You can't just be. Double as good. You need to be wildly better. Right. And um, and I think kind of what gets people moving is, you know, you and I talked a lot about optimization and incremental, like improvements and all of that jazz.
Um, [00:32:00] but there, you know, very few people get super excited about this. I think that's, that's just simply also true. And, and I think while this might be wrong in general, I think that's just how humans behave, right? And I totally do agree. Um, in terms of call prep, in terms of, you know, next steps in terms of CRM updates, in terms of like pipeline analysis, you know, what is slipping, what is not slipping, what's looking, well, where should I focus my time?
Like all of these things, you know what, AI can actually help you with that extremely well because there's a lot of data that needs to be crunched and consumed. Right? But again, like we are looking at, um, basically managing your context, managing your memory. The AI itself has, you know, it doesn't, is is not trained on your CM data, for example.
Right? So it's really, really trying to figure out, well, how can you, um, give the ai the context to be really smart for you? Right?
Koen: But like, and that's the thing. And, and, and, and, and, and we need to understand as again, leaders and, and that's like I, I've done quite some workshops now as well on AI powered go to markets.
Um, and if I [00:33:00] talked to founders and leaders. We need to understand that we need to provide this to watch our people. So the thing is, if we let the people do it, like same thing, preparation, et cetera, follow up. Let's be honest. It's not the quality we stand for as a company, as a brand, and therefore it impacts even so it can even harm.
But if we collectively understand where in the flow can we, I always say give 30% back towards our SDR, ae, account manager, marketer, whatever. And we really need to then make sure, and there again to watch your. Vibe coding skills, which I don't argue with Tony, but like, I think eventually you and me need to work together with, with, with, with the engineering team to really make sure that we, to your point, connect the CRM database, connect the call recording, connect the calendar, and really have that context.
And have the context. Yes. If Tony asks a question to watch a specific ui, ai use case or Kun, we both in the same company had a different role. So also that element, and it sounds very weird, but that's like where [00:34:00] I already started myself, I think more than one and a half year ago with my AI engineer. So every time when I have a prompt, it's first ask, who are you?
And for now, I just type my email address because then it knows all the context for myself. Then in this case, we've connected calendar, we've connected, um, CRM, we've connected call recordings, so it has all this information. In this case also, and this is actually a nice use case, which we call like the daily Prep agent.
It actually goes towards the web latest information on the context. We are joining that conversation on the company, on the industry, uh, maybe even on the, um, competitors because if I add some value to you in, in the upcoming conversation, you trust me more. And, um, then in this case, also it goes into this risk database of the latest tactics for.
Providing a demo towards the CFO or whatsoever. And, but having all this data connected for your context, for your upcoming meeting, and I know I'm talking to you, I get you not, but this makes a lot of impact, which all our people, whether you are the most experienced [00:35:00] ae. Or AM or CSM, we don't have time for it.
We don't like it, and please let the AI do this. So there, the AI is really a team member of your organization. And again, that sounds super boring as a use case, um, but super impactful because that compounds in the funnel.
Toni: I totally agree. And I think the, the other piece there is that it simply also needs to be super easily accessible.
Right. I think what, what I've, what I've. Also seen, especially in this context, is, um, you know, yes, we are saying kind of, Hey, it sounds super boring, but first of all, ask yourself, well, what's the prep my a are currently doing? Like, like ask yourself, you know, my, my experience is, my experience is they don't even read the SDR notes.
Like, you know, that's, that's the level, right? So kind of what can you really do to make this easy and better, more impactful for them? And I think, you know, those AI solutions certainly kind of help with that. But then on top of this, there's. They're like 20 different use cases for [00:36:00] ease. Um, I think with ai, and it's, I think what's, what's sometimes overwhelming and difficult is like, while.
Should I now buy 20 different solutions in order to kind of have 20 different use cases be done? I, I think, I think I'm just saying kind of, I think this is what people are struggling with sometimes. Right. Kind of, how do I get all of that stuff into kind of one, one connected piece? Yeah. Um, but let's, let's not kind of, let's not kind of get stuck here, right?
No, no. The But ultimately I think for the AE thing, the, the struggle that people are having and you know, that we are seeing is like on the one hand side. There's not the breakthrough massive, uh, use case. There are a couple of really good incremental ones, but you need to figure out how to, yes, make it context aware.
Connect it to your call recorder and your CRM and your product and your ticketing and what have you. Um, but also, you know, have this surface and all the different kind of places where, where it should. Right. Um, and I think that's still to be overcome with the, with the AE side. Let's, let's kind of move on to the, uh. [00:37:00] CSM am piece. And I think some of the same pieces, um, I feel are also relevant here. Thinking about preparation, you know, handover. Yeah. I think especially CSMs and ams, you know, when they do onboarding calls or prep calls or whatever they do, um, there's a lot of context that's locked away in the, in the sales funnel, um, that.
The sales guy, when you set them down to have the endo was simply forgotten. Right? Again, AI can help you summarize some of these pieces, pull them up, make them available to you, et et cetera, right? There are a couple of those items that are popping up, I think, for that rule that are extremely similar to what we just discussed with the, with the AEs.
Koen: And I think, uh, that comes maybe back to towards the discussion we had before is on, on the boringness of those kind of AI use cases, but the handover. That's on itself, of course, a use case, which, let's be honest, DAE doesn't like to have all this information, again, [00:38:00] stored in somewhere in your CRM or in a certain Doc Jo, jump on a call internally to discuss the handover.
Nobody likes it. The CS M actually also doesn't like it, or the implementation manager. Why not? It's not presented in their context because, I mean, it's a salesperson. They wanna move on like, Hey, uh. No, but if we there have all, again, the CRM data, the call recordings, and all the information stored, and this handover is done by AI agents.
But here, I think, again, the context for the implementation manager or the CSM to understand. In their context for the customer. And I think it, it, it, and that's like also where I could do, we, you and me, we are both different. My information, you know, now, after a few times talking with myself, the way I communicate, I can't help myself communicate very different.
But I also know that maybe there are some information lost in translation because the way I communicate to you and you [00:39:00] receive it a different way. And I think that's also where ai, uh, can really, it really, really help because maybe. The way you perceive that information is really significantly different than the way I pronounce this information.
So getting back towards the handover, super impactful use case and we can element as well of, uh, also know like to who and I'm talking Yes. Eventually.
Toni: So I think you, you phrased it in a very positive and very good way to say like, well, you know, the, um, the way you communicate might be different from the way I communicate.
And there will be something lost in translation, right. Also what is also true is, um, because you're the ae you might emphasize some things that make you look better, uh, and maybe de-emphasize some things that make you look worse. Um, but those things might still be in there, right? There might be things you dimension that's important to the customer, but maybe you know that the product is super strong.
We sold the
Koen: integration where we know actually it's still, you know, but it's very important for them [00:40:00] that we have this in, in that integration. So they sold on I, you
Toni: know, it. You know it, but my point here is. Screw all of this communication style thing, which I think is, is euphemism to talk about. Well, wouldn't it be good to have some objectivity objectivity here, actually, right.
Wouldn't it be good? And this goes by the way, for pipeline reviews forecasting too. Like, let's have, let's have a, you know, second opinion on what's actually going on here. And let's not say that's the leading opinion, but it's good to have another opinion. Right? And the same thing also in the handover. Uh, there can be stuff that's maybe a little bit more objective, maybe.
That's good. And this is not solely an AE thing. I think that's also a CSM thing. I've had, you know, multiple, extremely talented CSMs that lost deals, um, or, you know, had customers churn and, you know, they didn't flag it because. You know, either they didn't, they didn't necessarily wanna say it, or they didn't hear the same things, or they forgot about this a little bit.
And I'm not saying people are doing this on purpose, [00:41:00] but there certainly is a bias in all of us. Um, I think it's called self preservation, uh, to kind of bring up the more positive things that are connected to us than, than the other pieces. And having some ai, some objectivity on top. I don't think it's the end all be all, but it certainly is a use case.
Koen: No, and again, like, um, I think especially on the AM CSM side, and I know, and I sometimes refill for them because like they have a big book which they need to manage of customers.
Proactive Churn Prediction with AI
---
Koen: Yeah, I think, uh, for instance also within Personio where we've been working on quite a bit like the last, I mean, quite some time actually, again, cross-functional is what we call this proactive churn prediction modeling, right?
And, and I think, yes, we are all busy, but if AI can help you. Provide insights for the coming week. We are now like you start of the week and hey, I think these are the accounts where you really need to focus on, because D is according all the data. And that's the interesting thing when you talk about so much on the marketing SDR side.
But the [00:42:00] further we come down in the funnel, the more qualitative data we have on the customer. So actually we can be way better. So also ourselves, the persona. For instance, we are spending way more time on. Uh, and uh, right side bow tie. So existing business, uh, data and see like where AI can help there because we have all this information, which we know is quality.
And I think then I can help the AC CSM and AM to see like, hey, these, according all the data is really where we need to put, uh, put our emphasis. And uh, and actually we also do it on the, um, 'cause everyone talks on, especially for that way, everyone goes now in planning for next year, and, oh, we need to do more.
We need to grow. But I think there's so much hidden information on. Preventing churn to happen, or expansion opportunities are possible by me that also AI analyze good adoption and good impact, and la la la and bring that towards the equation.
Toni: So I think there was a, there was a really, you know, strong point and out of many strong points, Kun.
But one thing I just wanted to hook into [00:43:00] was really the, let's, let's kind of take a second here. You know, this could, this applies for AM CSMs and AEs. Um, yes. We as leaders are sometimes a little bit like, oh, you know, they're lazy. They could work more and stuff. But also if you flip this around. These days have a pretty big, uh, book of business.
Uh, they, they have a lot of things that they're supposed to be totally on top of, like researching those accounts, you know, building account plans, you know, having a point of view when they're kind of being fully prepped, doing this whole medic thing, doing the follow up, doing, you know, all, you know, and this times, what, 100 or something like this, right?
And then, you know what I just said? On top of that for the A-M-C-S-M product usage goals that they wanted to achieve but aren't or are right. Kind of roller plan, onboarding plan, all our, that stuff kind of happening on top. So if you really think about it, um, those, those people are, you know, for better or worse, juggling a lot of different relevant pieces of data and trying to make sense of that.
And I think that's, [00:44:00] that's asking a lot. I think it's asking a lot. Right. And I think there are two AI solutions to this. One is around the, um, summarization that we've talked about about already. Yeah. Kind of summarizing calls, summarizing, um, you know, usage data and so forth, and kind of trying to combine, you know, summarizing all of those things together.
Um, but I think then the other piece, and maybe that's a derivative of this. Is prioritizing, like figuring out who needs focus right now, who needs to be taken care of, right? Um, and again, I think the, the trick here is genuine thought, genuine ideas, you know, you know, sifting out what's b and what's not. I think those are unique capabilities of humans.
Like, and, and you know, I don't see that necessarily changing for, for a little bit. Um, but then the ability to sift through a bunch of data really quickly, um, may be guided by the person by the way. Um, I think that is trying to combine those two things in a really [00:45:00] nice way. And I think, and I think this.
Um, when you, when you, when you look at, uh, AI and coding, it's, you know, taking your thought and writing some text, right? Kind of. I think that's, that's pretty great. I think for the, for the go-to market side, it's way more to, you know, thinking about. Um, searching, summarizing, prioritizing, analyzing. Those are the kind of types of verbs I would use for jobs that AI should do in the go-to market side versus the, the coding, you know, writing text side.
Yep.
Koen: Yeah, and I think you are actually nailed to the point of prioritization, and I think that, that, that is, if you also almost like. There is this overload of ai. Yes, we have all this information and I can send you so much information to prep for your next call in the entire funnel. We need to make sure, and definitely for AMS to CSMs is where to then prioritize what specifically for that specific customer [00:46:00] and at a specific moment knowing that the renewal is coming up in three or six months.
And to your point, which actually we liked is like our bias towards. Being overly positive and like hiding, kind of like, well, we actually know we should put our focus on. But like, Hey, it's gonna be different conversation again. Um, I just put it aside, it'll solve for itself. And if we want to become like this 1% more sustainable, 2% growing next year, the answer compounding with our teams, and again.
We can't blame any of our ics. It comes from the leaders. We need to facilitate them with those tools so they can maximize their time with the customer on the right thing to create the right impact. And I think if you really will break it down, that's where I think the value lies. And the problem is everyone also, all those vendors, they're coming up with all the shiny tools.
What is the true impact you have been able to make for your customers? So that's more from the vendor side, right? [00:47:00] They haven't much, that's why you're reading in the news, like 90% of the AI use case trending. Why we as leaders, we also didn't quantify. And I think that comes down to myself, um, as we are both for, uh, lovers of the bow tie, but we need to understand the entire customer journey.
Where does it really, eh, where where does hit it's the fan? Where do we need to kneel down? Focus there when, when it comes to AI use cases, fix there, the fundamental fixed area data. And then to our point, I think make sure that we have collecting all this data and know that we know the context, but please help us whether you are a leader, even a founder or a sales leader, revenue leader, or whatever leader, or whether you SDR and AM or CSM on ae.
Help me prioritize. I think that's give you me a lot of focus. And of course then we need to still consistently execute. But I think there lies a big help in this madness of ai, which goes so fast to really focus and give some time [00:48:00] back to also rest and prepare for the next call because we are keep rushing in our jobs.
Wrapping up
---
Toni: But, but I also think, um, I think maybe kind of let's wrap it here, um, because I think what you also said there is something that. More and more people are waking up to, in general, this whole AI is gonna replace us. Likely not gonna happen for a while. Um, but AI really great at augmenting us. Like all the things we suck at, which is like reading a lot of text, sifting through a lot of data, doing all our, you know, trying to connect the dots, trying to see patterns in it.
We are really terrible at that. Um. AI can, can help us with that, especially on the go to market side. Right. And I really see it, and I think this whole copilot thing is a little bit overused, but basically, um, having the ability to get those, um, those pieces of information kind of served up to [00:49:00] you, um, you know, that that symbiosis, I think can be extremely, um, extremely powerful.
A hundred percent. Thank you so much. Um. For everyone who wants to check out what, uh, the two of us are doing. Also kind of G-T-M-O-S. How, how can we find that Kon? Um, yeah, that's
Koen: a good one. Or, or just through LinkedIn or my substack kto substack.com. I mean, I'm a fan of your substack, but I'm trying to still educate all my learnings and give it back towards the audience.
I'm still an amateur substack, but I'm learning a lot. And this is also, again, like how I learn on a weekly based on ai. Like I'm not perfect, but we need to keep learning and event with the goal. To put this back into our teams, like they need to benefit from the right prioritization proactively, not reactively because we are not good at these kind of things.
It needs to be really not a co-pilot, but it needs to be proactively delivered in our email, slack teams, wherever we are really, uh, working on a daily basis.
Toni: That's it. And, uh, obviously kind of in my case at ai. Uh, cool. Thanks a bunch [00:50:00] and, um, I don't know, maybe see you next time. Cool. Let's see. Who knows.
Bye-bye everyone. Cheers.