Philipp Baumanns

This transcript was generated automatically (YouTube auto-captions) and then edited for readability: names and obvious mis-hearings corrected, filler words removed, speaker labels added where the speaker is clear. Passages that could not be understood are marked [unclear]; short omitted passages are marked [...]. Timestamps are approximate. The recording is the authoritative version.

Philipp Baumanns 00:00:00

I think it's probably the best year to sell an AI product to enterprises. Right now, we're spending yeah essentially the same amount on engineering salaries as on coding tools. But I think let's say in like 2 or 3 years, I don't know, your dishwasher is broken and you want to find someone to actually fix this. Yes, you might still call someone or text someone, but you probably also have like a personal agent that you just tell, hey, can you like look someone up in the area?

Host

What do you think in the future are conversations that companies should never automate?

Hello, welcome to a new episode of Campus To Capital. Today we are joined by Philipp Baumanns who's the co-founder and COO of Telli which is a YC-backed AI voice tech pioneer based in Berlin. Philipp, great to have you on the show. Thank you so much for joining us today.

Philipp Baumanns 00:00:56

Great to be here. Thank you.

Host 00:00:58

And welcome from me as well, Philipp. Nice to meet you. And maybe before we get into Telli and what you're currently building, I want to look at your path because I think it somewhat presents a cookie cutter path of the traditional high achiever trajectory, right? I mean you studied at St. Gallen, then you had some PE experience, some VC experience, you were at MBB at McKinsey and then eventually you also spent some time at Enpal and yeah as I said on paper that's like the traditional, being a champion of this like hardworking and then high achieving linearity, but now you're a co-founder and you're out there challenging existing problems, providing solutions and engaging with investors and then trying to figure out your next steps. So when did you realize that you did not want to spend your career on this analyzing servicing side but rather go into the less structured and much more risky environment of being a startup founder?

Philipp Baumanns 00:01:55

Yeah, very good question. I mean I already knew that during university I would say, so I generally had this thought of okay at some point I want to do something on my own. At the same time I knew that I definitely wanted to have some experience and also learn some hard skills before doing that. But then when I was in these experiences, so different internships and experiences in like finance and so on, you always had this like very outside in view on companies and on entrepreneurs, but then when you kind of yeah met actually the people that worked at or that built up these companies that you, yeah, from your employer perspective invested in, I was quite impressed and I always thought okay, those are the way cooler guys than the investing people. I mean obviously they're also super sharp and super intelligent and so on but that's, yeah, I think that strengthened this feeling of hey I actually want to move to the other side at some point and actually build something on my own and work towards a product and not just analyzing it from the outside. So I think it's a combination of like early on motivation and then also yeah seeing a lot of people in these experiences and kind of understanding what these different paths look like in your day-to-day and then feeling like okay this is probably the better path to go towards.

Host 00:03:17

Yeah definitely. And I mean then when you went on, you were also at Enpal with one of your co-founders, Finn. You were employees number one and two in their heat pump unit and then reportedly grew the organization to more than I think 150 people within the first year. So you were surrounded by call center teams repeating the same qualification questions thousands of times, which then ultimately became the origin of Telli. So looking back, was that insight primarily technological, so that voice models yeah had finally become good enough, or was it more operational, so that because you worked at Enpal you really understood the workflow at core and that outsiders consistently underestimated?

Philipp Baumanns 00:04:05

Yeah. I think I mean in the end it was a combination of both. Initially it was very pain based. So we helped build up this company essentially within Enpal. And we automated almost everything that you could automate. So everything from like subsidy requests, emails, offer creations, appointment booking and all these things because we had so many touch points with the customer. But then the limit was still kind of like the customer interaction through phone and through voice. So that was kind of like, I guess, the blocker that we faced and where we thought okay there must be a better solution out there and there must be a way to also streamline these customer interactions. So I think that was the initial hook. But then as we got a little bit deeper and started talking to other companies that had very similar yeah customer journeys and very similar stacks, we found out that, or we basically came to this hypothesis of hey now the voice AI technology is like becoming better and better. When we started it was still really bad. So the first product was still really really bad. But we had this hypothesis of hey when voice AI becomes better and better then yeah why not use it in use cases that no one else thought of, because historically most people or most companies thought of it as hey okay I can use a voice agent for customer service. I can answer a phone call and so on. And I think we were the first ones especially in Europe that said hey you can actually use this in an outbound context, not

00:05:35

in the sense of cold calling people but like reaching out to existing customers and then booking an appointment or doing a lead qualification. So really like the outcome and sales and operations focused use cases. So I think that was kind of like this hypothesis from a technological perspective that we had. And then the second thing that came later, I would say like half a year after actually starting Telli, was that every single company that we worked with, they have a very similar technology stack. So they have like a CRM system, they have a telephony system, they might have an engagement tool to send out WhatsApps and SMS and so on. And we more and more like working with these companies and automating parts of their customer journeys, we thought, okay, what does this stack actually look like in two or three or four years? And I think that's really the second big thing where we thought, okay, this is such a massive opportunity to attack this stack and to essentially replace it because it hasn't changed for the last 20 years. And that kind of like yeah solidified this initial hypothesis, I would say.

Host 00:06:39

When did you understand that you could sort of fix this issue through AI? Because I mean like based off our own experiences and like I think everybody like hates bad customer support, right? Like there's nothing that sucks more than like being in this hotline hearing shitty elevator music for like hours and hours and then you finally reach someone and like they don't understand you or they can't really help you and it's like basically pointless. And then I mean I remember like a couple years ago some companies started rolling out AI chatbots and I think at the start like they made the problem worse. Like they made you even more detached from like getting to, like being understood or like being provided the solution that you were looking for. So how did you come about with the idea that there may actually be this like hockey stick curve moment where I can actually improve this issue and kind of get rid of it as well?

Philipp Baumanns

So I think the main insight that we had was there are so many different interactions that happen between a consumer and a business, because we had, I think for the heat pump product we had like probably 12 to 18 different touch points on average with each customer selling the product. These touch points were from initial booking conversations, to book someone in to visit your house, to more technical conversations about the planning of the heat pump, all the way to like real sales conversations. And I think what we did well in the beginning was we didn't think, okay, we're just going to solve all of these or

00:08:09

we're just going to automate all of these conversations right from the start, because I think that's something where you would fail, because for example with this chatbot example, I think what a lot of companies did in the beginning is like, okay, any request that comes in just goes through this chatbot and I think that made the experience bad because you had to cover such a breadth of different use cases and complexities. But we actually started out saying, hey, what are the simple use cases that have a huge effect on reachability, on conversion, on customer satisfaction that we can address, that are relatively simple and that we can enter a company with? So that was kind of the insight that we had. So we started out with a very simple use case, which was typically this lead qualification case, which is a super basic two-minute conversation, and then we started expanding in use cases. And I think that way we didn't try to like tackle everything at once, but really focused in the beginning and that helped us to also show results to our customers very quickly.

Host 00:09:06

Yeah, I think that you really lived through those pain points yourself makes the FO [unclear] story very special and is probably also a good hook for potential investors, I can imagine. But now when we look at the yeah sort of baby shoe weeks of Telli, I mean you started working together roughly a year before announcing the funding, formally launched Telli months later and then entered YC around 2 months after your founding. Shout outs by the way to Eric, a good friend of mine who interned last year. He gave me a bit of insights.

Philipp Baumanns 00:09:42

Nice. Mhm.

Host 00:09:43

And yeah, when you pitched to YC, you pitched without customers and apparently without a conventional deck. You can elaborate on that if you want as well. So beyond your backgrounds, what evidence could you possibly present at that stage that made YC believe that this is actually a venture scale company and not some, yeah, automation feature for call centers?

Philipp Baumanns 00:10:07

Yeah, very good point. I mean we had like, I think it was three pilot customers that we had using kind of like the very very rough version of our initial product, which was essentially us building agents for our customers on a platform that we put together. So it was very very basic. And I think I mean if you look at what YC yeah rates you at, or like what they look at essentially when they look at the company, it's a combination of the team and the ambition level that you have, and I think what we did well was on the one hand we had like a very very strong insight as shared before from our previous experience. We also had very deep technical knowledge. So our third co-founder Seb, he studied in Stanford, like played around with voice models quite early on and had like this technical insight and Phil [unclear] and I, we brought this operational insight. So I think that was something that helped convince Y Combinator and the second thing was the ambition level. So like we set out from the very start that we would not just be a voice AI product or just like an automation feature but rather that we would completely change the way how B2C companies would interact with their consumers, or how consumers also have a better experience interacting with B2C companies. And I think showing that ambition and showing the market size that we're going after in combination with this team insight, that was the

00:11:37

yeah most powerful thing. And then obviously it helps if you have like initial customers and initial traction and can really convincingly tell them, hey, this is just the start and we have this pipeline and we're going to grow this fast and so on. But I think the most important things are team and ambition.

Host 00:11:54

Yeah. I mean that's, when you look at who gets into YC, like I think that's always like, the resumes that the founders have are always extremely impressive. But then also I mean another thing that you need to bring to the table is to kind of show the ability and the determination to be able to massively scale up as well, right? I mean they're looking for these potential outliers that can reform and reshape industries. And earlier you mentioned that you didn't want to tackle all problems at the same time. But then maybe once Telli started scaling, what were some of the hardest technical operational challenges that you faced? Because I mean running a couple thousand calls every day for large companies, like that's a lot of stuff to handle with a lot of potential conflict points, right?

Philipp Baumanns 00:12:41

Yeah very good point. I mean there are definitely some experiences or anecdotes that we went through during that time, but I think from a technical perspective the biggest thing was essentially building up this infrastructure that could handle loads and loads of calls. So I think one thing that we ran into quite quickly was concurrency, so how many concurrent calls you can run through your platform at once. And obviously if you have like three or five customers, that's quite easy because you just have like I don't know 20 or 30 concurrent calls that you have because these customers probably don't do that much at the same time. And then once you're, yeah, that scales, that was probably the hardest challenge to build up, because you have it with every technology sub provider or stack that you have within your technology. So like the telephony provider that you use, the different models that you combine, the call workflows that we've built up and so on. For everything you need to build kind of like this capacity of hey this might scale to like multiple million calls that happen through this infrastructure every month. And that was the hardest challenge I think from a technical perspective and then that obviously meant like making sure that customers could run all of the calls they wanted to run. And there were a number of challenges that we ran into but then the second thing was also on the implementation side. So we are still

00:14:12

I think a very very lean team compared to the size of the customers that we serve and that is mostly due to the fact that we've built a product that is very, like our mission is essentially to make the users of our product completely AI native and to let them build their own agents, implement to their products, to implement into their CRM systems and so on, and to really like do it together with a customer, not just bring in I don't know like an implementation consultancy or something to implement the product. And there I think we made the right decisions in the beginning but then obviously as you scale up it's also kind of like implementation and support capacity that we have to have on the team, that we at some point found out also when scaling, that that was yeah a bottleneck at some point. So I think yeah these two technical and implementation limitations that we ran into are probably most notable.

Host 00:15:05

And then maybe one further challenge that I could imagine is keeping up with the competition, because AI voice startups, or like AI voice AIs, have been super popular in the last I would say probably 3 years and in our last episode we had the chance to talk to Christian Meermann, founding partner at Cherry, and they are apparently also investors with you.

Philipp Baumanns 00:15:35

Yeah.

Host

And they initially said that they normally avoid crowded markets and almost didn't invest because yeah there are many voice AI startups that appear to be building a very similar product. But they were ultimately convinced by your orchestration layer and that you can solve sort of the last mile, actually reaching the customers. So during the fundraising process, what did you show investors that could not be reproduced by another startup using the same foundation models or voice providers?

Philipp Baumanns 00:16:10

Yeah, very very good question. I think, I mean it comes down to these two things that we always focus on in technical decisions but also in terms of promises to our customers. And the first thing is that we always want to reduce time to value. So time to value for us is, you start as a customer, but when do you actually do the first live in production consumer call, that like yeah happens in the real world? Like that's this time period that we want to reduce as much as possible. And the second thing is driving outcomes for our customers and that is mostly related to okay, how do we within a customer journey drive a higher reachability, higher conversion, a higher efficiency and so on. And these two things have been important for every product decision, for everything that we do, and it's also what I think yeah helps set us apart. I think back when we did the fundraise with Cherry and YC, I think the biggest thing was always kind of like the functionalities that enabled our use cases and that enabled this driving outcomes very quickly, because we initially focused on a lot of outbound use cases. And there you for example needed to build your own dialing infrastructure and dialing yeah product essentially within Telli. So customers would be called up multiple times if you didn't

00:17:40

reach them. Also long-running call workflows, so that if a customer calls back the agent would accept them and also recognize that customer. So a lot of things that are not necessarily involved with the voice AI technology itself but more with the workflows around it. And I think that was the biggest product defensibility that we had back then. And until today I don't think there are any tools out there that have similar yeah functionalities, especially around these outbound calls. And now it changed a little bit because with kind of like our second fundraise and also kind of like where the company is going, we focus a lot more, beside this driving outcome, also on this time to value. So what we've built internally is this agent Charlie, which is essentially kind of like your co-worker that reviews calls, builds agents for you, builds workflows for you and so on. And here the promise is really, hey, we are the product that gets you live the quickest. And we've proven that with a bunch of large enterprise customers that usually would expect like six to 12 month implementations. And now with Agent Charlie, we can essentially go live within a couple of days or sometimes weeks. But either way, way quicker than with other products. And I think that's the main kind of yeah product defensibility that we've built up now, that has helped us also pitching to investors.

Host 00:19:07

Nice. I mean, okay, Philipp, we've talked product, we've talked competitors and differentiation and now you briefly went to the next topic with product Charlie, but I want to look at your customers as well because when we look at your published customer cases, they showcased very different operational environments. I mean, the most recent one, Sky, they just went live in under three weeks. Then Zenjob automated 90% of their initial candidate calls. Enpal increased weekly service appointments by 30%. So when you evaluate a potential deployment today, I mean you just talked about how you can do that quicker and better and faster, but what characteristics are you looking for on the customer side that tell you that a company is actually ready for voice AI and is ready for Telli?

Philipp Baumanns 00:19:51

Yeah. Also something that we learned a lot about I would say, because I think in the beginning you build a product, you launch it, you have a couple of initial customers and then there are loads of customers coming to you just because they find your product cool or because they find your company cool. And I think what you have to learn is to like deprioritize or qualify very well because otherwise you have like very differently looking customers that then all have different requests and I think it comes down to really the ICP that you want to serve. And the first thing are like these hard facts that we found out over time that are today our qualification criteria, that our sales team finds out in a conversation with a customer, and that's related to the volume, so how many calls a customer makes, the tech stack, so what is their current, like do they have a cloud-based CRM? Do they have a cloud-based ERP or something like that that you can connect to? And what is the use case that they want to cover? Because those are the three hard facts where you can also quite well disqualify. If there's a customer that only has like 10 calls per month, it's probably not worth it to like go through the entire process of automating them. But if it starts at like a thousand calls per month and goes all the way up to like hundreds of thousands of calls per month, I think that's like the perfect fit. Same goes for tech stack and use case, where we really know now, okay, how do you disqualify certain customers where you just know it's not going to work out? And the second thing is

00:21:22

really this, like as a company is this a priority for you and are you really looking into implementing this, because what we also see is a lot of companies that think hey voice AI is cool, let's like look at it, but then 2, 3 weeks down the line you really realize okay it's not a priority, it's not really solving any problem for them, and I think that's the second thing which we found is super important, that there's internal priority, buy-in from the leadership and that they actually want to do something and solve a problem with it.

Host

And do you see a trend when you look at your customers that they, I don't know, in the last years or last year said that they go more into, or become more AI native, because I think that certainly also arriving at the yeah more senior C-level positions, that people realize okay they might have to build a personal operating system or like change the text [unclear] within the company. So what would you say, how's it evolving?

Philipp Baumanns 00:22:25

Absolutely. I mean, I would say this year it's like crazy how high the demand is and even like traditional companies just coming to you and telling you what they've done with other products, what they've built internally, what, yeah, what they maybe also changed within the organization already. And I think it's probably the best year to sell an AI product to enterprises right now, because everyone is just thinking about how to adopt this and how to bring this internally. So I definitely see this priority shift. But even a year ago, I was surprised how many companies would like seriously go into deploying voice AI, especially because it's like something very sensitive. It's to end consumers. So I expected a lot of companies to be very careful with that. But yeah, last year it was already a priority. This year it's even more of a priority for companies. And it's a very exciting market to be in.

Host 00:23:24

Happy for you guys that it's going that well. And maybe talking about tech, this is a question I tend to ask in the most recent episodes we recorded. Like what's your personal tech stack? Do you have a second brain with Obsidian? Like how do you really operate on your own?

Philipp Baumanns 00:23:43

You mean internally? Yeah, I mean everyone is obviously, like every company, very Claude Code hooked I would say. I think the biggest effect we probably see within engineering, where we would have had like double the size at least for the same output that we have currently if it weren't for Claude Code. But even in all of the yeah more business related departments, I think it's incredible what we can do there. Here I mean we've built like this internal agent that has access to every single data source internally, so it has access to our product database, our billing tool, our HubSpot, our CRM system, our customer support tool and so on and it's kind of like the single source of truth that we use to evaluate any decision, to understand kind of like developments of product, of customers better, and I think it's the biggest game changer for us internally. And for me personally it's also something where I can just see kind of like the productivity increase, where I used to do a lot of like admin stuff completely by hand and now you have whatever, your co-pilot in Chrome and your personal agent and so on, which makes any admin work so much faster and so much easier. So AI native.

Host 00:25:11

Beautiful. But while we're on that topic of, first of all, asking questions that we've asked more frequently recently and also on the topic of AI, we recently spoke with the new chief of staff from Peec AI and kind of discussed this whole topic of having a strong technical moat in this AI wrapper and AI leveraging tech industry. And I mean Telli describes itself as model agnostic and I mean that gives you flexibility but it can also mean that your most visible quality depends on suppliers such as OpenAI, ElevenLabs or other infrastructure companies, maybe even Claude, what you just mentioned. So in your view, which capabilities should remain interchangeable commodities for you guys and which are some that Telli should actually maybe own themselves in order to establish this strong technical moat that you can hold on to for the future?

Philipp Baumanns 00:26:03

Yeah. Generally we think about it in a way where like everything where we have the strong hypothesis that it will be commoditized and that it is not worth investing into, that's something that we're happy to yeah essentially buy from someone else. And I think that's very true for a lot of the underlying models. I mean we still build a lot of stuff around it. So I don't know, we have like our own voicemail detection feature. We then combine it with like background noise reduction models and so on. So like we build scaffolding around it but I think where you know it's not going to make sense to invest into, when you know the kind of like voice AI or text to speech stack is commoditizing, that's something that we would probably deprioritize in building ourselves. And then in terms of focus on what we actually want to own ourselves and where we want to build up this moat, I think it's very yeah in line with what I shared before, it's always what drives outcomes for our customers. So for example, yeah, building a more advanced dialing strategy or having functionalities in the product like WhatsApp and SMS to send out a booking link or a checkout link during the conversation and stuff like that, that you just don't get out of the box from anybody else. That's like stuff that we definitely want to build and that we definitely want to own. And

00:27:33

the second thing is again like what reduces time to value for our customers and what makes our customers adopt AI faster. That's also something that we want to focus on. So especially Charlie, this internal agent, any feedback loops that we can generate from all the conversations that we're doing across our platform to then improve conversations again. Any integrations that we want to build to make our customers adopt it faster. Those are all the things where we are focusing on and where we think we can also build up a technical moat.

Host 00:28:10

And then in terms of your product offering and also Charlie, do you then maybe in the future see that your long-term product is not necessarily just a voice agent itself but maybe providing an automated management and quality assurance system for your customers to basically provide an unlimited digital workforce for them, or how do you see that?

Philipp Baumanns

Yeah, I think, I mean voice will still be quite core of our product, but if you look at what we're trying to build, like our kind of like statement is that we want to build AI that essentially runs customer-facing operations and that goes, I mean that goes beyond voice just within the vision I would say. And I think if you look ahead like two to three years and look at how customers will interact with businesses, it will be very different to today. Obviously, we started with voice because a very very large share of this market is still based on phone calls. So like especially within our customers, a lot of these businesses interact with their customers via phone calls. Now we added WhatsApp and SMS because that's obviously a channel that customers also prefer. But I think let's say in like two or 3 years, I don't know, your dishwasher is broken and you want to find someone to actually fix this. Yes, you might still call someone or text someone, but you probably also have like a personal agent that you just tell, hey, can you like look someone up in the area, get me a good offer, and this agent just goes and

00:29:40

calls businesses or, I don't know, maybe businesses have like their own MCPs or something, and basically tries to get information from that business on what a repair would cost. And I think in that world we want to be this interface for businesses to interact with the consumers no matter the channel. So yes voice will play a massive role in that but it could also be that we are just an interface to help businesses interact with personal agents of consumers because that will obviously be kind of like reality in very little time as well. So I think it's a very interesting yeah space that we're moving into and it will not only be voice, it will be a lot about other channels and it will also be a lot about how you build these interactions, how you evaluate interactions and that's why Charlie is such a focus for us.

Host 00:30:28

Well and then who knows, if the agent is really really good maybe the agent will recognize that something is about to break before the customer even notices and then has proposed solutions ready.

00:30:38

Exactly. I think that's a super interesting angle because that would expand, like that shows how much untapped potential there is for you, right?

Philipp Baumanns 00:30:48

Yeah, definitely.

00:30:50

That was pretty cool.

00:30:51

I agree.

Host 00:30:52

But what you then essentially are doing is that you sell to companies a way to scale operations without adding the same amount of headcount. While you guys at Telli are hiring across the board, I just saw your postings on LinkedIn. You're hiring across sales, engineering, customer operations, recruiting, by the way, to anyone that's listening that's also young and ambitious. They can probably reach out to you or your colleagues, right?

Philipp Baumanns 00:31:21

Yeah, exactly. You can like check out careers.telly.com [unclear] or just reach out to me. Exactly.

Host 00:31:28

Yes. And in regards to that, I think one interesting fact is that in one of the most recent 20VC episodes, the Mccora [unclear] founder mentioned that they are spending more money on AI agents than on actual headcount. And he also predicted that in 3 years, every S&P 500 company will do the same. So what's your take on that? And I mean we already touched upon Telli's tech stack and so forth. But where have you learned that you still need a great human on the team as well?

Philipp Baumanns 00:32:06

Yeah I mean to your first point, definitely agree. I just got chased on Slack again by my co-founder Seb who basically said, "Hey, please build into our kind of like financial plan and budget that we're spending yeah essentially the same amount on engineering salaries as on coding tools." So that's something that we want to budget in as well because it has become massively important for us internally to yeah increase efficiency and essentially have the same amount of money spent on coding tools as on engineering salaries. But I think if you think about the yeah key roles, I mean for us, because we're a company right now very much focused on mid-market to large to enterprise customers, there is a sales process that's still very relationship based. And I think, I mean we've proven that we can grow quite aggressively with essentially two full-time people in the sales team. Up to today, we had like the first yeah new sales people joining last week and that is only possible because they are so efficient, because they use AI tools internally and they can essentially leverage that to be more efficient. But I think that in order to become very very big and to grow very very fast and to close the leading companies in Europe and beyond, you need kind of like this relationship based approach in sales and

00:33:36

you need very strong salespeople that have this experience, that are AI native, that understand the market, that can essentially consult also our potential customers how to use voice AI and Telli within their processes and I think that's a role that will not be replaced so easily. So that's also something for everyone, I'm telling that to whoever's going into like yeah more financial jobs and so on. I'm like hey you should probably consider tech sales or B2B sales because you can earn a lot of money and it's still a job that will stay human for a very long time I think. And then the second thing I would say is also everything related to creativity. So I mean our engineers are very close to the product. We don't have a product manager internally. They're essentially product managers because they think about what does the feature look like. They research what does the customer need and so on. So they need to be very creative about that and work then with coding tools to implement that. And same goes for functions like marketing and so on where I think it's still very important to have this human focus as well.

Host

And when we talk about these kind of costs going out, both for engineers and for coding costs, like how does it work for you guys with money coming in and margins essentially, because I mean voice AI is not a traditional zero marginal cost SaaS tool, right? So every conversation has transcription costs, inference

00:35:06

costs, voice generation costs, infrastructure expenses, and on the other hand you have customers that probably look to improve their margins and have lower cost instead of doing everything manually. So as usage is scaling for you guys, where does Telli's gross margin leverage actually come from? And do you believe that the best voice AI companies can eventually achieve SaaS-like margins as well?

Philipp Baumanns

Yeah, I think it comes back to driving outcomes for customers again, because if you achieve to drive outcomes for customers, so you book a qualified meeting, you convert a customer, you help a customer resolve an issue, you increase the NPS for a customer, I think that's the moment where yeah it has value for the company and they will be yeah ready to pay you for that value that you generated as a yeah tool provider and I think that's where long-term we will go as well. Like right now our pricing is still very much volume based because of the breadth of use cases that we cover but I think that we will move to an outcome based world where the customer will pay us for the outcome and I think that's where our margin leverage long-term will be, because if our product is better, if we drive better outcomes, then we'll also earn more money on it and that will also beat probably current margins that we have. And the second side effect that you have is that I'm very convinced that any model

00:36:36

costs will come down quite strongly over the next couple of years, again increasing the margins.

Host 00:36:44

Yeah. I mean we now intensively talked about your product, how you built Telli and everything so on. But now looking maybe a bit ahead in the future, I mean we talked about that you raised your pre-seed relatively early and since then Telli has grown as we talked about. So could you maybe tease what the next phase of the company might look like? So is the priority now to go deeper with enterprise customers or expand internationally or grow the team faster? I mean, we're probably going to bring out the episode at the end of August. So maybe some things might have happened until then.

Philipp Baumanns 00:37:27

Yeah, definitely. I mean by the moment this is released, we've probably announced our second big round. So we just closed that round and for us the main priority is yeah investing that money into scaling the team and growing faster and with that it's multiple new segments that we address. It's new markets that we go to internationally and it's essentially a new phase for the company where our core focus will still be mid-market to enterprise customers. But we are also experimenting with ways of how to give access to like smaller theme [unclear] customers to our product, because generally with kind of like this vision in mind, hey we want to be the platform that any company of any size runs their customer-facing operations with, this can be yeah from a small one-person business all the way to a large enterprise. So I think we will, yeah, also by the moment this is released our product offering and our kind of like customer group will be much broader by then as well.

Host 00:38:31

Do you think that Telli eventually in the future might also have some use cases for B2C? I mean one person businesses are very close to B2B but

00:38:41

what do you think?

Philipp Baumanns 00:38:43

I think, I mean in general we're focused on yeah anyone providing products or services to end consumers. So we'll always have this mission of hey we're enabling you to become AI native and run your customer-facing operations more efficiently. There might be a world where, I don't know, this personal agent vision that I showed, where businesses have to react or be able to react to personal agents of consumers. There might be a world where we might yeah also be involved on the consumer side because maybe it makes sense. But as of right now we're fully focused on yeah businesses of any size. Yeah

Host 00:39:24

and maybe back to what we said earlier about team growth and you're hiring people. I mean you're hiring across the company and when we look at what you're offering, I mean there are some key startup buzzwords in there like low ego, high intensity, extreme ownership and I mean all of these sound nice of course but when you look across like other hot startups that are raising and are doing well and are growing fast, like I think you could probably read these in almost every job description. So maybe what are some operating mechanisms or measurements to establish like a strong team culture that actually in your eyes allow someone to be like okay but this is why I want to be at Telli and not at startup A, B and C, and also this is why I'm like proud to be at Telli and working for you guys and for your vision?

Philipp Baumanns 00:40:14

Yeah. The first one I would say is a very intense hiring process. So I agree that these words are quite common across the board with different companies. But what we hear on a regular basis is that even for candidates that go through our hiring process and are like, I don't know, maybe struggle a little bit or maybe also like excelled at this hiring process, the majority of them say hey this is a hiring process that I've not seen at any other company. It's very intense. It's reference based in a way because we do a lot of references. We do very very deep dive interviews and so on. And I think that helps us find out about this person. Okay, do they fit into this Telli culture? And I think it's a great way of like already putting the filter in the beginning and seeing okay who do you actually let on the team, and that I think has been quite successful to keep this very high intensity but still low ego culture. And then the second thing, more towards your question of how do you yeah make sure this culture actually remains and people want to work in that environment, I think it's a combination of different things. So, we have a very honest and very direct communication style. We have these regular retros across teams where every, I think it's four to five weeks, we do these retros within teams where

00:41:44

you can essentially scream at anyone and tell them what did not go well and why. And I think we're very big fans of never letting kind of like passive aggressiveness build up, but like honestly just saying what you think. Then I think the second thing is just being able to shape the company direction. Like we have the yeah ability to, you can post like any product learnings, any product feedback, any customer learnings quite openly, you can share that with the team and also push for implementation. So we're not in the sense of okay the founders decide everything but it's really about the person that has the strongest opinion. And then the last thing are typical I would say like team things that we do quite well, like quarterly offsites that we do. Like last year in October we went to San Francisco with the entire team. In July we're going close to Barcelona with the entire team. So like really building that very strong team connection that I know a lot of other companies are doing as well, but I would say that ours are way better.

Host 00:42:47

No, I mean that's great. That sounds exciting. And I mean obviously they might be better just because then the right people are at the right offsite, right? But maybe before we round up our conversation today, in terms of your hiring process, what does your ideal candidate look like?

Philipp Baumanns 00:43:04

For me it's, I mean depending on the role obviously, like it's not so much about the experience they bring. It's very trait based. So they need to bring the traits and skills to be able to do that job. And I can also see that person developing to do that job. Like that's the important thing. It's not that much about experiences. But I think two things that they always need to bring is number one like very very high agency, which is essentially yeah you're never blocked by anything. If someone tells you, yeah, this usually takes three months or I don't know, our legal process and data privacy process takes five months and has these stakeholders involved, like not accepting that, but like really thinking first principles, how do we work around that? How do we accelerate it? How do we find maybe another way to start working together or things like that? So like really being very pushy but then also smart about finding other solutions than just accepting yeah norms. The second thing is very high ownership. So I think a lot of people yeah come in and they have the feeling in the beginning hey, maybe for a month I'm being micromanaged. But I think this is super important because micromanagement in the beginning for like a month, it sets the foundation for giving you complete ownership over something. So if I build the confidence within 4 weeks, hey you can own this topic completely by yourself because I was very close to it and very close to you working on it, then I feel very comfortable giving you like complete ownership

00:44:34

and complete responsibility for it. And I think that's the second skill, like this ambition and this will to also take over responsibility and own things end to end.

Host 00:44:45

Yeah, I really like that. So honest feedback by me, I really enjoyed the episode. I think we've learned certainly a lot, also our listeners. And closing off the episode I would like to ask you maybe a bit of a contrarian question. What do you think in the future are conversations that companies should never automate?

Philipp Baumanns 00:45:09

Very good question. I think, like a year ago I used to think okay very complex conversations where you would have to look up multiple things across different tools and so on, very long conversations like 30, 40 minute conversations and so on. But by now I think that you can automate these conversations. I think in the end it comes down to like emotional conversations that you might not want to automate. So if it's like a super sensitive topic where you generally value the advice from someone that has the experience and that yeah can give you advice, that's probably something that you don't want to yeah automate with a voice agent or through any other channel, because that's where really the human touch is still very much needed.

Host 00:46:08

Can never lose that. I agree. It's a good point and also really enjoyed that conversation, Philipp. It's been super interesting. It's been a pleasure. Thank you so much for your time.

00:46:18

Thank you. It was great. Thank you.

Philipp Baumanns 00:46:19

Yeah, thank you so much for having me

Host 00:46:22

and for everyone tuning in until now. Yeah, we are now back from our summer break and you will hear in the next two weeks from us. So, thank you for listening and hear you soon.

00:46:32

Yeah, finally back from the corporate grind and ready to get entrepreneurial again.