Founders maze: A real look at a company in the making

No. This is not a tactical episode. But. You will get a real look inside of a company being created.

• (00:00) - Introduction
• (04:01) - Scar tissue
• (09:31) - Feedback and Market Research
• (13:28) - AI in Data Analysis and Future Prospects
• (16:06) - AI's Growing Influence in Business
• (18:06) - Exploring New Data Sources
• (19:05) - Challenges and Opportunities in AI Adoption
• (20:07) - Pivoting to Product Marketing Managers
• (21:50) - VC Inbounds and Funding Decisions
• (23:02) - Mickel's Departure and Reflections
• (25:05) - Breakthrough in Data Integration
• (27:26) - Introducing the GTM Graph
• (30:55) - Use Cases and Applications
• (36:31) - Call to Action and Conclusion

This episode is brought to you by Everstage - the highest rated Sales Commissions Platform on G2, Gartner Peer Insights, Trustradius with over 2,000+ customer reviews. Some of their customers include leading brands like Diligent, Wiley, Trimble, Postman, Chargebee etc.,

You can go to https://www.everstage.com/revenue-formula to check out Everstage and mention Revenue Formula to unlock a personalized Sales Compensation Strategy Session with Everstage’s RevOps experts—crafted for enterprise teams to maximize performance.

Never miss a new episode, join our newsletter on revenueformula.substack.com

Creators and Guests

Host

Mikkel Plaehn

Marketing leader & b2b saas nerd

Host

Toni Hohlbein

2x exited CRO | 1x Founder | Podcast Host

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.

[00:00:00] Toni: And I think what, what I learned is you have, you know, parts of the people that are too nice.
[00:00:06] Like, I'm doing lots of interviews and it's like, do you think that's valuable? It's super
[00:00:10] valuable.
[00:00:10] Mikkel: Totally. Gonna use it every day.
[00:00:13] Toni: it? No. You know, it's like, it's like, okay, cool.
[00:00:17] Mikkel: it's just generally unhelpful for someone to say. No, don't, you shouldn't do that.
[00:00:22] It's like, it's a, it's a terrible idea to go in this direction. It's, it's just not gonna work out for you guys. And then try and solve it by saying, if I were you, I would do this
[00:00:31] completely other idea I have,
[00:00:32] Toni: Today's episode is brought to you by ever stage, the top writer platform to automate sales commissions. You can create a single hub for your reps to track all their deals, earnings, and performance history thanks to EverStage's seamless integration with Salesforce, Microsoft Dynamics, Slack, and MS Teams.
[00:00:50] Reps can also know exactly how much they could earn with Crystal, EverStage's one of a kind commission forecasting module. Visit everstage. com to learn more and help your sales teams ace 2025.
[00:01:04] And now, enjoy the show,
[00:01:05] Mikkel: So, by the way, my kids, they're so dissatisfied with the decision making in this household. Basically, they are now changing their own names. Elena was like.
[00:01:18] Toni: I thought they were, they were deciding that someone else needs to run this
[00:01:21] house.
[00:01:21] Mikkel: No. Probably
[00:01:22] Toni: firing you, you
[00:01:23] know,
[00:01:23] Mikkel: I think they have some cloak and dagger meetings happening in the
[00:01:27] room because sometimes they just disappear and it's all quiet and I get
[00:01:30] really worried. So, you know, it feels like I have a target on my back. It makes, makes sense. Makes sense.
[00:01:35] Toni: I, I, I would, I would I. You know, not enjoy the next shareholder meeting.
[00:01:40] Mikkel: No.
[00:01:40] Exactly.
[00:01:41] Toni: these guys just might kick you out.
[00:01:43] It's
[00:01:43] Mikkel: Extraordinary.
[00:01:44] Toni: I'm sorry. Michel, we really appreciate all the hard work you've done and we've brought this enterprise, but I think we have to part
[00:01:52] ways.
[00:01:52] Mikkel: No, exactly. Speaking of padding ways, so you just came outta stealth.
[00:01:58] Toni: That is correct. I think we came outta steal last week for, you know, for the very dedicated folks listening to this on the Tuesday when it comes out. Like you should, like, that is the way it works, right? Yeah, we came out of stealth. The company's called Atif a TT v ai. Check it out. We'll talk a little bit about this and to a degree we actually.
[00:02:22] Wanted to talk a little bit about the journey, kind of, I think some people have heard about the founder Maze or the PMF maize or basically describing this extremely frustrating process, which certainly is that right? But also, you know, we know who's listening.
[00:02:40] Mikkel: And we want you to buy No.
[00:02:41] Toni: yeah, we, we won't, we won't be able to give everyone like the next GTM play
[00:02:46] or something like this, right?
[00:02:47] So, you know, if you're not interested in the story, you know what? Tune out, hit the subscribe button. But on your way out though, like, don't forget about that. But otherwise, I think it's either gonna be interesting for founders or just a fun story about the two of us which people probably will enjoy.
[00:03:00] Mikkel: I think honestly also, as you know, when you're an employee and checking on a job, you just take for granted all of the things that were built. And I think very few people tell the story of almost infancy and then what happened until, it's usually like a bi small part of the whole story that gets told.
[00:03:17] So I think
[00:03:18] that's why I was like, Hey, this could actually be pretty interesting for us to just talk through. And yeah, you're right. It's not like we're gonna share three different go to market analysis. You're not gonna go perform and how to slice your territory and commissions and all. That's not gonna be part of it.
[00:03:32] No,
[00:03:32] Toni: I had three people on LinkedIn, by the way, reach out to me after the self prospecting
[00:03:37] Mikkel: Like finally.
[00:03:38] Toni: and they were like, oh, you know, finally some advice
[00:03:41] again, you know, all of this 11 X Klan,
[00:03:44] Mikkel: Bullshit, bullshit,
[00:03:46] Toni: give me the real stuff. Sorry. Sorry, boys and girls. Not gonna be this today either.
[00:03:51] But maybe let's start the story.
[00:03:52] So where, where, where, where should we start? Michel.
[00:03:55] Mikkel: it's a really good question. I'm wondering whether it's the whole. We were kind of traveling.
[00:04:01] I were traveling, and then you were traveling and this whole company basically got formed between four, four wonderful people through some learnings we had at at Roblox together.
[00:04:10] Toni: I, I wouldn't say learnings. I, I would call it scar tissue, like all the stuff building roblox. So many things we love by the way. So many things we hated too. And we were like, oh, you know what? Those 3, 4, 5 things we hate, not that the opposite. Please, going forward
[00:04:27] Mikkel: So the CEO obviously hated it. Hated
[00:04:30] Toni: That's it. That's it. That's it.
[00:04:31] So I think this is, I think this was kind of the starting point, right? Us kind of thinking about what could be next. And, and we were very quickly, very clear on what, what is it we didn't want to do? Right? And I think one major thing in Roblox, which is like terrible was the whole integration help, I wanna say.
[00:04:52] Right. Basically, yes, we tapped into all of those tools, which is great. Great. Cool. Please do that. But everything had to get like super exactly right. And if it wasn't lining up, then the whole model was like, you know, it's, it's one thing that your MQL are off like one or one or two or something like this.
[00:05:10] But the thing is, if it's part of the model. Like that suddenly throws off your opportunities and your revenue and so forth. So everything had to be exactly perfectly right. And that was, that was difficult to achieve. So, I mean, in, in all honesty in all honesty kind of, we basically charged around 20,000 euros, like roughly plus minus and we always did like a consulting gig basically in the beginning for the implementation for the rollout. You know, map the funnel, you know, do the data model, do the growth modeling, all of that stuff. Each of these packages, you know, we actually now kind of added also, we have some consultancy stuff that we could actually announced yesterday.
[00:05:49] Each of these packages are now clocking in this area, right? It's kind of, if you think about it, the insane amount of consulting we've done for someone to use the product. I think that was also a thing for us. We're like this, this can't be,
[00:06:04] this can't be how you build a product company, you know?
[00:06:06] Mikkel: Well, to be fair, also, when we looked at other companies, Looker, they had that approach of basically implementing it. Usually company also will hire consultants to help implement your data lakes data warehouse and all that stuff. But yeah, it's intense. It's not like a. I guess the wet dream when you're in this, in the trenches that Roblox is the opposite, which is Oh, that PLG company where customers just self onboard. Please, can I have that?
[00:06:31] It's like, no, I totally get it. I totally get it.
[00:06:34] Toni: So I think that was, I think that was one thing we were like, not, not that again, please. Right. I think the other piece was that was also pretty difficult to manage for people, right? It was like pretty much I mean it was powerful, don't get me wrong. I think the model in itself was like really, really cool and I think people loved that part.
[00:06:50] But keeping the model running in the right way was not only a difficult. Thing to do engineering wise and development wise. We kind of spend a lot of resources on making sure the performance is high enough and kind of it's responsive and stuff. But also for everyone to understand all the intricacies and using it became like super difficult to scale.
[00:07:08] Right. And kind of, and, and I think it was also kind of a barrier for people to spend more time in it because it was like so far and so far away, so difficult to kind of use in that
[00:07:16] sense.
[00:07:17] Mikkel: Yeah. But I kind of got a I mean, I kind of liked that it was almost a standardized way to do reporting. That was because you'd done the consulting was totally aligned with also the team and stuff. Right. But even so, I guess what happened which, which is the last point here was the usage was not kind of there either.
[00:07:35] Right? It was, it was a bit infrequent, which is also like some of the things we discussed was, well, how. How much do you actively work with the numbers? Like, it's not like counting where no, you're gonna sit in your ERP system for a few hours today, my friend. It's like, no, like you and I, I, I think I at least have, have said a few times my morning ritual is looking at the numbers, but this is like a, we're talking minute, minutes exercise.
[00:07:59] Right. So the infrequent use is definitely like, I recall was one of the, the major like hurdles
[00:08:06] Toni: no, exactly. And, and then I think the, you know, to a degree, it was like, Hey, is that a big problem? Kind of, do you have to be in this thing all the time? Right. But still kind of looking at this, we were like, we just not comfortable. Like it's, I, I think you can run a really strong business with infrequent use if you, if you will, like a really strong product with infrequent use.
[00:08:23] I don't think these things negate one another, but at least kind of, you know, walking away from it, we were like, you know, it would make us feel so much more comfortable if someone would have like a daily value from this thing. Right.
[00:08:35] Mikkel: I think it's also like one of the things we discussed, especially in building, additive here was how, what's the cheat code to PMF? And part of it is, well if you can get a lot of users who can give you a lot of feedback and when you have infrequent users, there's kind of a limit to also how much feedback you're gonna get at the end of the day.
[00:08:52] 'cause they're not using it that much. So I think it also that, that also makes things a little bit more tricky to a degree. Right. So,
[00:09:00] So definitely some scar tissue. I think there,
[00:09:02] Toni: So, and then I think from there, right, we moved on. It's like, okay, it should be simple. It should be, you know, often to use. And you know, with the team that we have, it's like, you know, probably something in the go-to market space, something in the data space, something with ai kind of like, like that was roughly kind of the shape it took right in the beginning.
[00:09:19] And then we started to talk to some people,
[00:09:21] Mikkel: Yeah, yeah, yeah, yeah. Yeah.
[00:09:23] Toni: and I still remember, I think.
[00:09:24] Mikkel: Was so annoying. It was, I was like, livid. I was
[00:09:27] livid from this process
[00:09:28] Toni: I, I think I talked to two people that actually were guests on the show.
[00:09:31] So people that like, I would say very, and let's just say highly revered, let's just say it like that. Know their shit like for sure. And we just told them about the direction and there's a, you know, let me stop you right there.
[00:09:44] Lemme stop you right there. Don't do anything in the space of data. Don't, don't do any data. Don't do it. It's that it's never gonna be better. It's done. No one wants to do data. It's difficult. It's like, you know, five year sales cycles. Don't do it. Don't do
[00:10:00] it.
[00:10:01] Mikkel: yeah. But it's so, it's so funny. Like I I remember that process of talking with. One was a co-founder in this, actually, two were co-founders in this space, and another was also working in a high growth company in this space. And it's like, it's just generally unhelpful for someone to say. No, don't, you shouldn't do that.
[00:10:20] It's like, it's a, it's a terrible idea to go in this direction. It's, it's just not gonna work out for you guys. And then try and solve it by saying, if I were you, I would do this
[00:10:29] completely other idea I have, which is like, well, we have these insights that folks struggle to work with data and we wanna make it easy, and you're telling me not to enter data. Because it's a tough market or like Yeah, it was. I, I think, I mean, you have to listen a bit to it. It's not all kind of madness, but I think there's also bias there. And I think for me, what pissed me off at least was there was just no, like, there were no help, no insight there that we could kind of work on the back off.
[00:10:59] It was just a, Hey, bad idea, don't do it. It's like,
[00:11:02] Toni: But that in itself is valuable. Let's just kind of be clear, right? I think I think you have to have this stubbornness, but you also have to be somewhat open. It's kind of a super weird balance you need to strike kind of in this journey, right? And I think what, what I learned is you have, you know, parts of the people that are too nice.
[00:11:19] Like, I'm doing lots of interviews and it's like, do you think that's valuable? It's super
[00:11:23] valuable.
[00:11:24] Mikkel: Totally. Gonna use it every day.
[00:11:26] Toni: it? No. You know, it's like, it's like, okay, cool. So, so, you
[00:11:31] know,
[00:11:32] Mikkel: It is like I will give you, I will give you access for free. Can I have your email and then crickets. It's
[00:11:35] like, oh no. Oh no.
[00:11:37] Toni: no, so you have people that are way too nice and there's some tricks to kind of fool them. So, so they aren't. And then you have people that I don't know. I think those are, those are on the, on the more smart genius end of the spectrum that are trying to find ways why this is a bad idea, you know?
[00:11:56] And the thing is, there's truth in all of that stuff, but it's like, you know, what do you fucking listen to? And I think that's, that's kind of the, the, the, the, the difficult piece. And you know, further on this one we also looked into the space in general and kind of where other people doing something similar and are they successful?
[00:12:14] And we're like. Eh,
[00:12:16] not successful.
[00:12:16] Mikkel: And I think that's a better clue. And by the way, in all this stuff, I think what would've been more helpful is for them, for folks to say, Hey, here's like three reasons I wouldn't do it based on what I've learned or experienced. Right.
[00:12:26] And I think one of them actually did that and said, Hey, you know, in this space you're gonna have a high a CB that lands you an enterprise.
[00:12:33] There's. Three default vendors that folks will pick. They don't want to get fired. You need all these certifications and it's like a lengthy process. And, and then it's like, okay, yeah, fair. If we start now, it's gonna take year and a half until we have the first
[00:12:46] deal signed and done like that. That makes sense, right.
[00:12:49] Then you can kind of navigate it.
[00:12:50] Toni: no. And then, you know, we saw some folks you know, I think they built like around 22, so, so 2020 or 2021, like they kind of started around there and we were kind of doing some of the research in late 24, early 25. And don't get me wrong, they were, you know. They were basically, all of them were still 15 people.
[00:13:09] Like, you know, it's that, is that now a sign of a bad company? No, it's not anymore. It's like you kind of with, with all this bump, but it certainly isn't a sign of a great company. Right. Kind of that, that, that also kind of is true. Right. So we did see that there's obviously some struggle and there, and I think kind of throughout this period we.
[00:13:28] Realized more and more also that AI needs to be further and further in the back and, and in the foreground actually. Right. I still remember you and I starting out this journey with this basic kind of brushing off this AI thing. It's like it's never gonna work, you know? And it's like, hey, you know, look at this report from this Goldman Sachs guy that says like, AI still hasn't delivered jack shit.
[00:13:48] And look at this other thing that sounds like AI is gonna hallucinate until forever. You know, all of these, we were kind of very much like, no, you know, we want to not build an AI company actually.
[00:13:58] Mikkel: Yeah. Yeah. And I, and I, I actually forget what the moment was, but I think we just realized with this whole difficulty of analyzing data for a lot of folks AI could totally help solve that. Like just the, the sheer power of. Being able to prompt your way to an answer. And I think it's also like the more you start using ai, the more you realize, well, these other things should also be leveraging ai.
[00:14:24] And then it becomes a why aren't they? Why can't, and it's like, I've said it a few times before in the show. It's like, for me it's the iPhone moment. Like just before where I was like. Why can I not go on internet, on the internet on my phone? Like I'm spending so much time there, finding so many things, watching so many things, but I can't do it on my phone.
[00:14:41] And then
[00:14:41] the iPhone came as like, oh, that's what I needed. Right. So
[00:14:44] Toni: No, it's, it's a little bit, it's so, and we're kind of jumping a little bit ahead, so I'm just kind of teasing it a little bit, but to a degree, right, it's, we, you know, every single crazy bit of information on this planet is really just a prompt or a query way. Like really,
[00:15:00] Mikkel: yeah. A
[00:15:01] wife A new house,
[00:15:03] You know.
[00:15:04] Toni: it's so simple.
[00:15:06] And we expect, like it comes back immediately even though it, you know, you know, insane amount of data. Right. But what is so messed up is that. We're totally okay that it takes three to four hours or a week if it's inside our own company. Like if, if I wanna know something from inside of my own business, I'm okay.
[00:15:26] If this needs to run through a data analyst and the data scientist and takes two weeks and comes back and it's a dashboard, I'm like, that's not what I asked for. And I actually forgot what my question was to begin with. Right. Isn't it's so, it's so super weird that we are okay with those two completely different realities.
[00:15:43] Mikkel: yeah, yeah,
[00:15:44] yeah, yeah, yeah, yeah. It is. It is.
[00:15:47] Toni: No, and I think I think one of the reasons why we also kind of went into the, you know, AI being part of this, was we also started reading some reports from the industry. And, and this was like, okay enterprise spending for SARS in general. Down 12%
[00:16:04] Mikkel: Yeah, yeah, yeah, yeah,
[00:16:06] Toni: Enterprise spending for AI in brackets might also be part of us up 40%. And we're like, you know what? Maybe we shouldn't try and swim against the tide here. That's maybe stupid.
[00:16:18] Mikkel: And it was also like we, we checked all the reports for what are the challenges in marketing, in sales, and like what are their priorities going into 2025 and you know, top three on both challenges and priorities like ai, ai, ai. It was just like everywhere. It was just.
[00:16:33] Top of mind and then factor in even stuff like, not, not that this was like a pro I, I think driving our decision, but look at the stuff from Carta on AI versus non-AI valuations and investments happening. And I think when you just couple all that information together, like their companies growing super fast on the back of being AI first there's a growing trend that AI will rep replace people and services.
[00:16:56] There's a growing budget for ai. There's growing investment happening in ai. It's like. Do you really want to do a non AI company now? Seems, seems silly actually. This whole landscape is changing. In a couple of years. Maybe this SaaS UI is just not gonna be as it was.
[00:17:13] Toni: No, I mean, it's a little bit like, oh, I wanna be pocket hunters. I want to be the last SARS alive and no more, you know, behind me. You know? And it's and I think
[00:17:24] Mikkel: great investment rate. We are gonna do railroad in Europe.
[00:17:29] Toni: I, I
[00:17:29] Mikkel: It's what?
[00:17:30] Toni: I think it's also silly to go into it just because it's a hype, right? Kind. It's like, oh, Web3 and crypto call.
[00:17:36] So that for us it was extremely, you know, the, the writing was clear, but also there had to be a good reason, you know, to actually kind of do it. There had had to be a non phony reason to use ai, which, which we kind of clearly, clearly, easily found. Actually kind of once we started to kind of look into this a little bit
[00:17:53] more.
[00:17:53] Mikkel: But it's like you look at the business case, like with anything, if you wanted to enter a new market, you look at the business case, what is on, on the mind of those folks? I think for us there, there were a clear trend pointing that it would have to be ai. We saw a lot of opportunity in the data space.
[00:18:06] I think like after this whole feedback, we got. We, we did some some soul searching, I guess. And we realized that there were another data source
[00:18:17] we had not considered, which was previously, it was like, let's say structured, highly structured and quantitative data. So think about the data you have in your CRM, number of opportunities, leads, et cetera. And we just realized, well, with AI you could. Tap into all the qualitative, all the unstructured data out there, which was kind of like, okay, let's, let's discover this. And kind of a, I guess a new journey ensued from there.
[00:18:40] Toni: Yeah, so I think we showed some of the first data products with some folks and they were like, oh, this is really cool. And then, you know, they told us about their problems and the problems was like. Hey, you know, my customer service reps are saying stuff on the phone and I don't know what it is. And, you know, we have all of those gong recordings, but we kind of don't use any of that stuff.
[00:18:57] And like, there were, and, and there should be so much value in here, but, but no one is using it actually. And we are kind of more realizing, okay, wait a minute.
[00:19:05] People in fact don't get super excited about a ai, you know, text to SQL thing, which we by the way, build at the
[00:19:13] Mikkel: No, no. I remember like, oh, that was underwhelming. And then you also see 10 under vendors having done
[00:19:19] pretty much the same thing. It's like, well, okay,
[00:19:21] Toni: Not the same, obviously, you know, snowflake, databricks, all of that and have that can basically on top of it. Yeah. Anyway and then we were like okay. Oh, okay. Qualitative data. Conversational data, unstructured data. That's kind of interesting. Let's kind of dig deeper into that. And I think there, we fell a little bit into rabbit hole and we're gonna, you know, like a little bit.
[00:19:40] I think we first of all, we kind of realized this is super interesting, and then we asked ourselves a question. I. Well, for whom is
[00:19:47] it
[00:19:47] Mikkel: yeah, yeah, yeah. Yeah.
[00:19:48] Toni: Right? For whom is it the most interesting and kind of, Hey, the textbook startup 1 0 1 is like, Hey, you need to, you know, one problem, one persona, one kind of ICP.
[00:19:59] Let's, let's go kind of, and who is that? And I don't want to tell the whole story but for some wonderful reasons.
[00:20:07] We ended up saying that product marketing managers. That's the thing, that's what we wanna do.
[00:20:13] Mikkel: yeah, yeah, yeah, yeah.
[00:20:15] Toni: And, and you know, what was the reasoning? I mean, it was it was it was kind of compelling on a couple of areas, which was like, Hey, those product marketing managers, they need to come with messaging and, you know, the story and landing pages and product stuff.
[00:20:29] And there are no, so sales calls, they don't, they don't read. You know, those tickets and if they do want to get insights, they need to spend like hours manually listening to sales calls, et cetera, et cetera, et cetera. So there was like a compelling case. We talked to a bunch of pmms who also kind of confirmed that compelling case.
[00:20:46] But I don't know what happened with that idea.
[00:20:48] Mikkel: Well, someone on the team shut it down. No, I think I think number one we saw there were a few vendors doing that stuff. I think that was one. We also figured out that some folks were already building kind of their own solutions. And then we also realized these guys, do they have any budget?
[00:21:05] Toni: Yeah.
[00:21:05] Mikkel: Do they and no, they don't.
[00:21:07] And then you're also in the, if you've ever seen Scott Brinker's MarTech is it called MarTech scape or something? He has
[00:21:16] basically a map where he takes every software logo that sells to marketing and plots it on there every year. I wanna say there's 10 x the amount of local. So it's just exponential growth.
[00:21:27] It is insane. So we were like, is this, is this really the space? To like, to build in. Is this, is this really it? Is that the wedge, is that the starting point for this thing? And I think just on, on some of the budget pieces and it became difficult to say, yeah, that's, that's where also like, do they have access to connect the tools and a million other things gets brought into question
[00:21:49] there.
[00:21:50] Toni: And you know, kind of chronologically around this time we started to get some VC inbounds actually people tap on the shoulder, Hey, what are you guys doing? Let's kind of have a conversation, yada, yada. And, and really around this time I think we had a couple of conversations. You know, investing and kind of funding is very much like dating.
[00:22:09] I learned this from a French guy, by the way. So he must be right about this. But you know, we had a couple of conversations. We were like, oh wow, this could be pretty, pretty great. But, you know, it's, it's a little bit like the, the hot girl that that clearly is out, out of your league, right? Kind of.
[00:22:25] Clearly that's a no. And then we had a couple of good conversations with like. Not the, the, the prettiest and the, and the dowry also wasn't that significant. So we were like, you know, maybe, maybe we don't need to do it. Right. Kind of, if we don't find the right fit, maybe we just don't actually do it. Right.
[00:22:43] Which is actually also a plug for some of the services. So if you want to you know. Customer fund the business. We have a bunch of services that might be actually super interesting for you guys, so kind of check it out at the AI slash services by the way. So really, really cool stuff there.
[00:22:59] But it also led to something else Mickel.
[00:23:01] Mikkel: Yeah, it led to me leaving you guys
[00:23:04] in the dirt, basically. No, I think it's like so we leave on good terms obviously, otherwise we would do this episode and still talk. I think what became clear to me was I can't live off of , ramen noodles. There's just no way. Like, I have three ki kids, they would like to do something during their summer holiday and you know, I also have a mortgage to pay.
[00:23:23] And I think, you know. Reflecting over this. Actually, funny enough, the other day, it's like you can, you can laugh at people or tell people off for pursuing vc. For some, it's just that's the only path actually they need. They need an income. You can't live on zero at some point
[00:23:41] because you've leveled up your lifestyle.
[00:23:42] Not that it's lavish lifestyle I have, but I got bills to pay. And I think that became very clear and not, you know, throwing VC completely off the table for me was like a yeah. Then. Then I actually do need to hop off this journey and kind of let you continue it without me. I, I was even reflecting, should I just bet on two horses?
[00:24:00] And I was like, after a second of reflecting, like, yeah, that's gonna be. Really terrible. I'm not gonna do well on additive and job hing. So now I have like some companies swag on here in my new role that I landed in in ver little time over at Get y who is funny enough, an AI company that works with qualitative data which is kind of cool.
[00:24:19] So not the, not I will say not purely the analysis part, but basically. What it does is it recruits folks for, for serving and interviewing. So if you're like carlsberg and you wanna test bottle design, it goes out. Recruits folks, does the survey. Interviews them, delivers the analysis. And because it's all AI ran, run, it's, you know, two days and you have the findings versus, you know, using one of those consultancies you pay a million bucks for and it takes them one to two months to deliver.
[00:24:48] It's pretty, pretty massive gain. Anyway, that was like a, a, a sidetrack to the whole thing.
[00:24:53] Toni: no, that, that was the sidetrack. Right. But the, you know, coming out of the pro marketing saga, basically unstructured saga kind of, you, you kind of, get the, the lead in actually extremely well. Right now we kind of.
[00:25:05] Realized something pretty interesting, right? Kind of. We were sitting in the product and we were asking things like Hey, how much pipeline did I lose against competitor X?
[00:25:15] Right? And we were like, Hey, you know what? That should, that should be totally a thing it can do, right? Now what happened is that the, the AI basically it crept out. It just didn't, it actually didn't really work, right? And here is now why it didn't work. The, all the data on pipeline and, you know, accounts and deals and all of that stuff that was in a, you know, very much traditional, I wanna say tab.
[00:25:42] Think about a tabular, tabular data, you know, base as Snowflake kind of, you have
[00:25:47] columns and
[00:25:47] Mikkel: a spreadsheet basically. Yeah,
[00:25:50] Toni: And the, all the unstructured data, so you know, for example, a competitor or objections or pain points or kind of what was said by whom. Like all of that stuff. It doesn't quite fit into, you know, rows and col that well it's more like a network.
[00:26:05] So in fancy words, this is called a graph, it's a graph database, right? So what the AI had like extreme trouble with is like, you know, it could, I could ask it to say like, Hey, you know, which competitor did we come up with? And then told me, and here's like quotes and everything. And I could say like, well, how much pipeline do we have?
[00:26:21] And it would tell me, and, you know, give me all of that data. It basically had extreme trouble to traverse back and forth between those two worlds. So this is then I think where we had a little bit of our our breakthrough, at least technologically speaking. We realized, you know what those two things need to live in the same world.
[00:26:41] And all of this table, you know, snowflake kind of structure data, actually doesn't lend it, it lends itself extremely well to this whole thing. So really kind of those two things basically need to live in the graph. Right? And I think that was really you know, the guys hacked it together. We asked those questions and it was just like magic.
[00:27:02] It just could, you know, answer the whole thing back and forth and, and what have you. And suddenly we kind of had this ire moment where we felt like, you know, the, the, the shape of the database. So this is like a little bit technical now, but the shape of the database gave the AI an understanding of the data itself, which is kind of like a, sounds like a circular era in a, in a spreadsheet.
[00:27:24] But that's basically what happened there, right?
[00:27:26] And that's, you know, that we now call GTM graph that we know. We'll talk a little bit more into kind of what this actually means and kind of who else is doing something like this. But I think our foray into unstructured data. Helped us unlock the technological side of, oh, this is actually, this works pretty nicely actually.
[00:27:45] Mikkel: Yeah.
[00:27:47] Toni: And then, you know, we did some digging around that like, you know who, you know, what is, how do you mark this thing? Who cares about this? And so forth. And what we actually pretty quickly found was the whole Klan story.
[00:28:01] You know, we, we covered this last week on the show. Surprised why. But the thing is right what they also realized, you can take all of your data, you can put it into a data lake, house or warehouse or whatever.
[00:28:13] And if you deploy AI on top of that thing. You still have a super intelligent ai like that doesn't change. But basically the stuff that it has in front of is just a, just a, just a bunch of, you know, messy, comprehensive data to it. So what you will basically get in return is you will get a very confused and kind of untrustworthy ai if, if you just take, you know, super crazy AI, put it on all of your data, it will actually not figure out what these things mean, how they're connected, you know, what means what in relation.
[00:28:45] And then it will become a very underwhelming experience. Right. And the, the only use case that works is okay basically this text to SQL, you know, find me all the deals where this is true and this is untrue. And then it kind of gives you the result back. But it doesn't give you the intelligence of, oh, I actually understand the whole thing in front of me.
[00:29:05] Right. And. That is actually what Klarna, I think what is pretty, pretty interesting about this, this is actually one of the, the things that they mentioned as the major unlock, right? They understood that fragmented data, you know, data silos, however you want to kind of call it, was a detriment for their team to be collaborative and pro productive basically.
[00:29:26] But it also disabled the AI to kind of help in any of that stuff. So kind of the first thing that they did. And this was for us, like, oh, we, we figured it out was basically they build a a knowledge graph. They build a knowledge graph on the same, you know, data infrastructure that we are also using.
[00:29:42] And I also want to say that that kind of knowledge graph idea also actually has been used by a company called gle gle.com. Check them out. They're kind of pretty awesome. What they're doing is they're basically taking all of the unstructured data that you have in Slack and notion and. Google Drive and your SharePoint and like PDFs and stuff and make it available for you.
[00:30:04] So kind of questions, questions people usually ask, like, you know, where does the team go to dinner usually, or, you know, who's in charge of this project? Or you know, what's the maternity policy? Kind of that kind of information, right? The category is called enterprise search. By the way, what we are doing, which is kind of different from that is.
[00:30:23] We are using as data sources. Your Salesforce, your Slack sorry, your Zendesk, your gong. You know, all of these. Structured and unstructured sources that are specifically for the go-to market, right? And basically with this GTM graph we are able to deploy a bunch of structure and understanding of all of your data that you already have.
[00:30:45] We are then able to plug on top AI that helps you to. Traverse the whole thing and even can give you a recommendation and insights and all of that, you know, wonderful stuff.
[00:30:55] And really the the use case for now, like we are super early on, so kind of, you know, but the use case for now is, you know, number one people that need to work with multiple tools at the same time.
[00:31:07] And you know who, you know, who is that? Well, it could be our wonderful crowd, the rev ops guys. That maybe need to assemble a, a report or get an insight, you know, more than just Salesforce, for example, or more than just HubSpot, right? Kind of. They need to pull in some product data, some finance data, some whatever data.
[00:31:24] And you know, those two things kind of, it's really difficult for them to stitch these thing together, right? Kind of. That's one, I think a more standardized one is actually CSMs. Like we had Marty from Pilot on the show you know, he was also talking about how difficult it's for CSMs to, well they need to get information from the C they need to get information from the ticketing system.
[00:31:43] They internet information from. Call transcripts and to get information from product data to get one picture of, of a customer. Right. And you know, the last one I wanna say, and there are a couple of other pieces here actually as well, but the last one I wanna say is like CEO. It's, I had this myself so many times as, as, you know, reporting to his CEO, but also being CEO myself, it's like.
[00:32:05] You have so many questions about so many areas of the business and sometimes almost revealing if you go and ask someone for a login for something or kind of have them create a dashboard for you you know, this takes this friction away. It can just go and ask a question and it kind of takes you in all the different areas of specific, the go-to market that you might be interested in.
[00:32:23] Mikkel: Yeah. I can also just say now, like starting a new job. You wanna see, I literally asked for the first report almost now. Right?
[00:32:30] And because I don't know where to look, but it would be so much easier. It was like, oh yeah, we're using this tool. Just go and ask it a question. Right. It was funny. You also mentioned the CSM.
[00:32:37] Like when you think about sales, I. AEs, they spend what only 30% of their time selling, I think is
[00:32:44] the stat out of Salesforce. Something like this,
[00:32:46] we could call it BS on it, do all kinds of things, but that's like ballpark the number 'cause there's so much admin research and other things happening as well. I think it's the same for cs, by the way. I think there is a lot of admin and prep work there. And I also think to a degree they wanna know if you own a portfolio of 1 million Euro, I would be interested in knowing what is happening on my accounts this week. Like, or
[00:33:08] last week, what, what changed? Are there some who didn't log in?
[00:33:11] Who logged in? How many tickets were there, like a spike on an account? And
[00:33:15] I, and I think this is, this is where it gets interesting that you can start combining. Some of the data sources and, and make it available to to these teams,
[00:33:23] Toni: No, think also for CSMs is like you as a manager of CSMs is like, cSMs are wonderful people and they're always telling the truth but they're also people and they're subjective. Right. Kind of. It would be helpful to just understand what was the sentiment on this call, really kind of, do we have an issue or not?
[00:33:38] Right? You know, this kind of information would actually be great to have this as a, as a, as a separate piece that factors into your churn risk. Right. And since you mentioned sales. What, what we found is, is less an issue for salespeople. Ideally they live in their Salesforce and we know they don't.
[00:33:53] They live in wherever they want to live, frankly, in a spreadsheet, on a piece of paper and scratch pad and slack, wherever. What we found was really telling in the, in the sales use case though, is sales management, and especially specifically for performance management. Lives in, you know, needs to pull data from all kinds of different tools.
[00:34:09] It's like, okay, we need to understand what actually happened in Salesforce, but also in your sales engagement, also on your call recordings. Also, you know, compare this to quota and what you should have achieved. It's like you have a bunch of different things you need to pull from in order to get like one holistic picture of an employee.
[00:34:25] Whether specifically sales rep, whether or not, you know, he or she kind of performed really well. Right? So that's the, that's another use case where this is pretty prevalent. And the last thing I wanna say on this plug here, by the way is really what I think is starting to be freaky. So a couple of few people might have noticed MCP.
[00:34:44] There was like an MCP thing
[00:34:46] released.
[00:34:46] Mikkel: SPI also releasing the air and
[00:34:48] usage is just like quant, tippling or something is crazy is
[00:34:52] going on there. Yeah.
[00:34:53] Toni: So what is that? You know, I'm just, it's not that, but I'm just gonna say it's an API for ai. Let's just say it like this, right? Kinda. It's not that, and you know, all the people that know, it's like, Tony, this is not correct. And you're right, it's not, but let's just say it's an API for ai.
[00:35:07] So what happened is we basically took the GTM graph kind of for on, on synthetic data internally to, for testing reasons. We set it up on an MCP server, and then we had Cursor connect to it, like connect that. Basically Cursor is kind of this developer tool, AI developer tool and connect into the graph.
[00:35:24] And then, you know, you know, co-founder of of ours. He basically just said like. Build a forecasting dashboard, enter, and then there's a, and basically, and then it was done,
[00:35:36] like
[00:35:36] Mikkel: you had Clari. Then you had
[00:35:38] Toni: pulled it it, pulled it from the GTN graph, it was all there. We used the MCP and plugged into cloud sonnet, kind of the, the frontier model that they have.
[00:35:46] And you could ask questions on top of it and basically get access to your whole go-to market data and would give you really smart stuff back. Right. And this is. You know, if you, if you try to do that by plugging into Salesforce and whatever, like it's, it's just not gonna work. I can, I can tell you that.
[00:36:01] Right. And then there's some really cool things here coming out where this maybe doesn't only need to be something that people interact with because Right. It's difficult to kind of have people change their, their ways. But this could also be the one plugin for all your go-to market data directly to your preferred AI vendor, for example.
[00:36:18] Right. And kind of use that and leverage
[00:36:19] Mikkel: Like the connective tissue between Yeah,
[00:36:22] I like it. Nice. Where do they go to if, if someone wants to try it now and go, like, you had me at ai. What, what, what Now? What Now?
[00:36:31] Toni: Go to Active AI, A-T-T-I-V-E ai. And we're currently opening up for, or last Thursday we've opened up for early access or, you know, private beta or whatever you want to call it. Basically lots of the stuff is built and now we just, you know, need to work on actually fine tuning this stuff.
[00:36:50] And what we are looking for is you know, not, not crazy amounts of people. It's like a handful or two of of companies to work with. That ideally are a bit more mature and they go to market like five plus million, 10 plus million, something like this. And have this problem, you know, to a degree that I kind of just outlined.
[00:37:08] Right. Kind of probably have massive fragmented data. Maybe try and use AI on top of this, keep failing. And basically kind of need a better way to solve that. And then, you know, the use case enabling that for, you know, our rev ops friends and report building or for the CRO to understand what's going on for performance management.
[00:37:24] You know, customer 360. All of these pieces you know, I think I think can be really powerful actually. So if you, if you think you have that problem if you wanna just check us out, ping me on LinkedIn or kind of, I think there's an inbound thing on the website, kind of you can use that. Yeah.
[00:37:39] Mikkel: Yeah. It looks like, it looked like you were looking at me for an answer to the inbound thing, but it's obviously not my remit anymore anyway.
[00:37:49] Toni: Yeah, I think, I think I'll get used to
[00:37:50] that.
[00:37:51] Mikkel: No, exactly.
[00:37:52] Toni: It's like, I think, I think what will be difficult is you and your terrible jokes. You can't, you know, you won't have an outlet for this anymore.
[00:37:59] We
[00:37:59] Mikkel: I'll just go and request,
[00:38:01] I'll just go and request demos and in the comment field I'll add information there.
[00:38:05] Let's say I, we'll, I'll find a way. I'll
[00:38:07] Toni: find a way. You'll find a way. No, exactly. But yeah, otherwise that was kind of the roblox to active and active may story. And, you know, let's keep you updated on, on, on where we go from here and where, I mean, where Michel is going from here, kind of, you mentioned this already and otherwise, thanks everyone for listening to the two of us babble about this journey. And you know, promise next time we'll have some go to market insights for you guys
[00:38:32] again
[00:38:32] that
[00:38:32] Mikkel: Something with C payback or something.
[00:38:36] Toni: have a good one. Bye-bye
[00:38:38] Mikkel: Bye.
[00:38:38] Toni: bye.