How TextCortex is Changing the World of Google Chrome AI Agents

while you need to probably need to search five to 10 minutes to for a file in a file management system such a knowledge agent gives you your answer in 30 seconds plus you then make it also collaboratively what I quite often do is I ask what is Tex cortex user Grove or Revenue growth I get an answer and then collaborative iterative work on top of it I just say please write this into an email for our investers hi my name is Demitri panichi and I'm a content creator agency owner and AI enthus you're listening to the AI agents podcast brought to you by jotform and featuring our very own CEO and founder idin tank this is the show where artificial intelligence meets Innovation productivity and the tools shaping the future of work enjoy the show hello and welcome back to another episode of the AI

agents podcast here today we are very happy to have two of the most interesting people that I've met in the productivity and uh SAS and AI space from Tex cortex want to introduce yourselves sure uh my name is Dominic I'm one of the co-founders of Tex cortex um started the company three years ago together with my co-founder Jun um and what three and a half years later we have about two million website visitors every month on the platform uh multi-million registered users uh been in the space since it was still called natural language generation and yeah today have on the one side been named g2's fastest growing um one of the fastest growing products of 2024 by G2 and Gardner also recently put us on the short list for um yeah top generative AI platforms particular when it comes to customization knowledge agents are one

of our main use cases we are today deploying to bigger Enterprises and B2B cases and here I give over to Bish my my right hand hi hi I'm literally on your right um I'm burish I'm currently uh responsible for the growth operations at teex cortex pretty much everything related to growth it's been a little over two years now um it's been such a journey I'm I'm mostly worked in in in the startup landscape not so much of a corporate landscape and I think that has uh a lot of Ops for me throughout my career um it's like you I don't want to call myself like a growth hacker but yeah you you learn all all sorts of stuff uh to quickly leverage uh in whatever Market you're in so it made me I think quite flexible and yeah even with without any prior uh prior

knowledge in this AI industry or yeah in this natural language processing or generation uh yeah I learned quickly um read a lot of stuff stuff also on the technical level and also in the application Level like how the product building is happening in this scene and now I'm yeah responsible for the U growth operations at Tex cortex our Engineers also see barnow as a backend developer yeah I also build a lot of things on the side uh to to uh finance my marketing operations that's funny that's Frugal and you got to be frugal in the startup yeah so uh speaking of uh startup I know you guys have been around for a couple years I'm sure the journey to getting to uh millions of active users per month is not necessarily a um um a boring story so I I definitely want it's probably probably

going to be pretty exciting I'd love to hear a little bit more about um you know kind of the story of how TX tax cortex came to be and you know how you got here it it it originally started under the desk of my co-founder Jun is my technical counter counterpart and uh back in the days when I met him he was already fine-tuning gpt2 models under his desk today I call it I call it the heating because every time something happened uh he had already product description generator running in yeah beginning 21 2021 uh and every time somebody did something with this product description generator suddenly the heating went on H um but quickly afterwards we started working we first yeah built an AI writing tool this was kind of like the first wave of generative AI uh where hundreds of companies were built

thousands of products uh most of them also dead today uh it was the first wave of generative AI that was gpt3 which enabled a lot of people to build something with it many didn't do so very sustainably and and with a focus on reten and then you could see the waves I think the next wave was image gen then 2022 then came chat assistance yeah first of all not with the fancy apis as we know today when CH GPT was released one of our backend Engineers had built a similar context handling system already for coding product for his own work so we quickly could reverse engineer the reverse engineer C gbt that made our shift a little bit into the assistance Bas uh then early 23 were we were pretty dissatisfied that you know they all sound generic can't really hold you a pen uh

and we went deeper into yeah what what needs to happen that this AI sounds like you knows what you are from very early on also deep workflow integration one of our first products which which felt really well and was a nice experience was our browser extension so we usually saw this whole implementation in fre steps Behavior custom or behavioral customization knowledge customization and workfl customization and from there this this journey just grew further and further when one of these big AI writing tools uh raised the 100 million plus round I came from or I come from the VC industry beforehand um I basically told our team we cannot fight this Behemoth with ads yeah we need to find other Frugal compounding ways to grow and that's when Bish joined us roughly two years ago a fellow founder just yeah gave me his secret Source on

how they build a Content production engine which has scaled to millions of website visitors already and we basically took this learnings Bish mostly reverse engineered it and we applied it for us and today we have how many 10,000 plus landing pages under text cex.com for each intent you might have for generative AI you will most likely find an article of ours and this is how we Farm not only engagement interest inbound inbound leads today there were Fortune 500 customers coming through this content today uh German publicly traded companies many of these hidden Champions somewhere in the forest of Europe also now under our belt it's how this growth Journey happened I think that's really cool uh di let me take over next question um it's really cool that you actually um um use your own product to grow your own product uh and that's very

interesting story because like you were able to get to one million users and that's an amazing number in just two years and you mentioned that you actually use your own product to to build this content uh engine to to to drive tra traffic to your website to drive like two million visitors every month to your website so this actually this actually proves the success of your product um so I'm I'm really curious about like more details about that that success like you know uh what made your product so special that you were able to like uh bring like many many people actually use AI but they weren't very successful they actually they didn't get uh they didn't get uh like they were like punished by Google because they generated AI content but there was something special about what you did that actually made a big

difference and that brought you this success can you talk a little bit about that so we do have processes and yeah there's all like there there's a lot of noise around Google or any other search Eng like analyzing it uh AI generated content but in in their own documentation it actually says like it has no relevance of it has not any relevance like it doesn't matter whether you use any automation or any kind of AI to generate those pages what matters is that does it look spammy or does it look like does it give you give the reader any value or not so this is this is the um the simplest thing of course to keep in mind but also the most important one so you don't want to spam your page like with so many uh of the same or similar kind of keywords

or phrases so we do have this kind of process like written out it's like a six seven page document that how we do content for example this is just particular to content yeah or any kind of landing page generation and we did have this content so when on our platform like it allows you to create a knowledge base and share it with your own team and um currently like we use notion to store such documentation and so on so you can basically uh connect this document and have ai basically access to that information and create something new out of it and our platform also for example allows you very granular customization options so that you can for example tailor the tone of the responses and um add some rules guard rails like you must always do this never do that so on so you can

basically have up to 20 different rules and we have a lot of users that mix out on these rules because they they want a lot of customizations and card D and I mean we had this uh incidents where for example I think it was Chevrolet know uh Chevrolet is AI chatbot selling or like promising to sell a car for $1 because of oh yeah like yeah I mean this is kind of the um issue issues you don't want to happen and and with the guard rails that we allow you to configure and like very simple steps uh prevent such such such incident so with the amount of customizations we offer to our users with the knowledge I don't know internet search and all this um T tone personalization you are able to create a very unique page and get a first draft I'm saying first

draft because it's this is what like this is the point and the extent the that that we use our product for Content creation and once we get the draft we still have humans in the loop that take it from there and just improve upon that I don't know and we for example use internet search again to find some some footnotes to back up our claims in that content give some more stats industry reports and to to make the content look more enriched so to say and those kind of stats impact the um credibility of your content as well which allows you to eventually rank higher on those um for example we do have a lot of uh market research I don't know generative AI in Switzerland Germany Austria and so on the the BL region so say and many of these Pages rank quite higher

so in and that that kind of content also allows us to be referenced on third party media out for example if it's like I don't know like like sites like uh statista or something like that for example look for um like oh generative AI how it's impacting the workforce in Germany for example and they come across our content and they reference to it and then when there is a backlink coming from them The credibility even boost UPS further and allows us to generate drive more clicks or leads essentially so this is basically we still have humans in the loop we still have editors content editors and um authors that uh use our product to a certain degree integrated like for example as Dominic also suggested we have a very integrated Chrome extension so if we are creating the content pieces on Google Docs you can

pretty much also use our Chrome extension on the side or write within the line you can rewrite stuff you can fix some of the grammar grammar mistakes and so on so it's truly feel into truly feels integrated so this is pretty much how we used our product so far and um created a little over 10,000 p was also a process ultimately yeah initially it was you who did the strategy and quite a bunch of the of the copywriting as well now also put some numbers behind it we have two and a half copywriters uh who are super charged with our own product basically uh also something like with this heavy with the heavy citations and finding those numbers this is something which we for example uh got inspired by a consultant many of our customers today May consultancies usually 50 to 100 plus Consultants then

within the company if you want to wow this people and that they site your content you need to give them numbers yeah there needs to be a footnote to whatever you do research can be easily done with an online search Bish nicely describe the customization in terms of you know maybe aligning to a brand script we have a very informal brand we are very direct very honest in general this just comes from from from the base company values and and Mission which we have set out based down on the Persona you want to connect the dots in between blog posts put it into a knowledge space ask for when did I write about something about how will generative AI impact productivity in the D regions or German Germany Austria and Switzerland and this combination in also making ourselves the the biggest and the most active

use of our own product yeah comes with high quality measures respectively also leads to a better product this was always the biggest goal and that's also what today I basically drill our sales team on I don't care about a new sale yeah I only care if the S if the same customer says after the initial period either they continue with us that's great or even they further expand with us we have a Fortune 500 customer who in the last year I think they started in March March April they tripled their account with us nice to a good six six figure account and and these are the stories which on the one side you need great product but then also great service to it and I think a little bit later we will also talk a little bit about how do people accept agents in the

workforce yeah how do people work with AI because this is still the biggest problem when many people still feel a little bit lost where to start everybody talks about ai ai ai let's think about some of the Google presentations last year where just the word AI was mentioned 50 times but how I I sat down with workers councils in Germany uh who were concerned about implementing those tools not helpful you need to you know this typical German angst it shouldn't be like this yeah give the people empower the people to test things generative AI don't always or they won't always have a set use case no but what we've seen from the interaction with a general purpose assistant so to say at one point people start building their own use cases yeah so out of our small assistant interface people started building invoice recognition tools

wow then they approach approached us to oh how can we make this uh workflow product how can we make this into an API product and that's slowly how you split up expand we come back to the topic of retention and expansion which is mattering the most to stay around in this market what were some of those specific concerns that you actually were here ing from the people at the workers uh I didn't I didn't hear if you said count conferences or councils one of the two uh workers councils I counil okay it's something very common in Germany okay um but people are I mean in particular German is a little special when it comes to data due it to history now um but employees are afraid that uh employers for example spy on their employees so this is in particular in in Europe it's always

a big a big question mark some of our us customers also work with us because we have a European infrastructure and they rather trust our data security and our transparency and our approach to the topic than for example some us competitors um so this is the part on the whole security privacy side where then people come in ask questions there are already for some of those big assistant tools they're not only selling the productivity gain but also the observability gains for management for example this is already something which you can hear in some discussions in bigger in bigger Enterprises not no um this much about the workers councils uh another anxiety you usually Al are not so much an anxiety it's more an an adoption blocker it is um what's the use case no what's why are we using this thing because the chat interface

and this is where often times it lacks in creativity for many people what am I throwing in there yeah am I is this just here to write emails for me or am I writing blog articles uh we have some of our power users they started do comparing documents with each other one of my most actively used agents on our platform is actually a legal consultant which I called Harvey Spectre like in the in the suit Series yeah um just to sometimes think about some legal perspective and you know that the the telephone of my of my lawyer is not you know he up because I constantly call him with all these Enterprise contracts at the moment but they need to be open to test those things out and you need to give particular corporate Enterprise employees the space to do so often times the buying

process looks like what's your use case okay then we go in now it it often times starts with okay we have no idea of use case everybody pressures us to do do to do something with a pocc or what we now do is service service let sales we do the most German thing and sell driver licenses for now it's kind of like a first first approach yeah let let us give you an idea what you what you can do and then we narrow down but always interaction uh have regular meetings gamify the whole process yeah put put some of the employees a little bit into the spotlight and say hey these five people here have outperformed everyone in terms of interaction with the tool also helps us in our support we've seen us plenty of times to do that for three to six months and

I'm very open about it because the more support and service we have given in the beginning the team itself at one point start onboarding I just had one of our oldest customers is now what two and a half years with us I just had an onboarding with this I still I still managed this account because it was one of the first ones these people were raving about us still after two and a half years H even with so much competition and they always find something new in the platform as well I mean essentially really the use I mean this is this is also why we um I mean I've been telling you to work on for example like together let's work on this use case videos and so on because we want to really make use cases more apparent and I think I mean the

value essentially lies under these use cases like yeah everybody says oh AI is cool and now they say AI agent is cool and but essentially I mean we will talk more about it but the the value lies under the in the use cases like what is it for what what what value you can derive out of it to make your life a little easier and on the product building level we also try to I mean the the technical ities or all these infrastructure could be awesome as it sounds or like could be really complex because our research team also like loves talking about a lot of oh let's build this let's let's build such a workflow and make it also visible to the users and it's it's not the best idea most of the time because it just makes things more complicated and that's why

we've been contemplating on these AI agent uh concept as well and how we should be um how it should fit into our current ecosystem in a way that still creates value uh with the respect like with the respective uscape but essentially that's really a challenge especially in large organizations like people especially in the management teams I think know what can be improved but then comes some of the concerns or the applications at scale and how it can be done and then during that process we usually just do handholding I'm curious about the um use cases like can you talk about like top five use cases for Tex cont cortex and like how did you discover them like uh what what did you use to discover those use cases and like um Talk little bit about like how people use Tex cortex yeah I mean discover

those those use cases were discovered through two enablers first of all we had a platform which was very easy to customize it was based basically build a way that we can customize agents quickly and that helped us when we approached customers what turned out for us as a great seller in Germany would say sell something like warm bread um was knowledge agents uh this described quite quickly yeah imagine many use a Google drive or a one drive for example example to file a lot of knowledge file all the information files presentations PDFs yada yada and then I constantly have this for example with my with my uh with my lead for for B2B in Partnerships the guy is producing slide Deck Over slide deck and sometimes I need something and I can't find it I ask because he's working in his own little knowledge structure

World on Google Drive we internally use Google Drive um if you ask something there or search something there you might find it if you build the structure if you didn't build the structure very little chance to find anything if not you have very very set processes but they will always get broken no matter in which organization I worked there's always people who file information somewhere else a knowledge agent is there to you know use a tool such as access your Google Drive do searches find the right and relevant files and respectively very quickly uh this agent can give you an answer and what is then extremely important is proper referencing uh proper citation ideally and this is currently what we support with PDF files if you click on a reference it should move you to the sentence where the information comes from yeah and and

and this comparison maybe we can also later in the video we have one slide showing that H Google Drive versus us while you need to probably need to search 5 to 10 minutes to for file and a file management system such a knowledge agent gives you your answer in 30 seconds plus you then make it also I ask what is Tex cotex user growth or email for our investors beautiful my job is done I can fire myself as one of the one of the co-founders uh and can move on with my life so in in search operation it's only one step at every time and then a lot of manual work this has been one of the bestselling use cases and then you see other more workflow relevant use cases build I don't know invoice recognition is quite near metadata generation as well this is

a new product which we will launch soon such I mean imagine whatever files you upload um we had this quite often a first recommendation of the uh confidentiality level which typee of categories come are are in there um we worked on a project of a major German automotive uh vehicle Builder I don't know what the pro that's probably the right word they've hired an army of seniors and students to label a lot of files to see whether whatever they have created the knowledge a decade ago is still relevant today this is a task which you can also have have a large language model evaluate if you have the right ways of recognizing information but it was a it was a step like a lot of Step In also on the product building level like when we first introduced this um knowledge concept I mean we

we saw the need from the users oh can I um like the people at first like people weren't even aware of of such a need that like they it was way before this retrieval augmented generation concept and yeah first bring this onto our platform then people started asking questions oh it would be so cool if I can just uh uh integrate my Google Drive oh it would be so cool if you could use this on SharePoint and so on and then we started building those Integrations and basically tried to keep everything in sync and yeah with a with a Flawless infrastructure um but now uh it's not only about retrieving some information and referencing but now they want to build something new out of this stuff the current data for I don't know if if you have some sort of a situation for example for

project management and now you you are for example you onboard a new client but uh it's a very similar project that you did in the past and now you want to use that knowledge and apply it to a new new case so that's a completely new information but you can still use AI to look at the information in the past and just generate something new out of it and these will be eventually even more improved than will be more sophisticated with the concept of agent but then again uh I I yeah I don't I I can talk more about this later but I don't think you would need an agent or uh for for every single test so this is then it boils down to again use cases what kind of stuff you need an agent for and what kind of stuff you can just

get things done in zero shop we have internally a nice overview on how we see the this product AI product Market H we always talk about first step is generic AI this has been you know the GPT 3es Etc which were used today's models to just build a very plain one use case writing tool for example I've recently encountered yet again in Berlin a new LinkedIn post writing tool which everybody Hypes where I was thinking like wow that this train was already running two years ago finds a little bit here and there smaller Niche where distribution was key again so you have generic AI you have personal AI this is when you give people all the tools to customize this interaction and then later you really have this autonomous AI yeah where agentic you can you can play in there even though I mean I

see agents or agentic AI more as a spectrum then I would say this is this one specific thing because ultimately the the end user they don't care whether they call it a co-pilot an assistant an agent hairat yeah whatever we have had great experiences with our um with our Enterprise customers um C giving an identity to an AI and just calling it Marvin the whole time because it it binds the people a little bit closer it doesn't make it feel so strange if if you would say if you say in German it's an agent most of the people think it's a spy because Ag and and an AG in German is it is a spy word it's not what you would usually use it for in in or you know it's a individual agent to do something it makes total sense but for the common

user they also build some some anxiety you want to move around this now and every type of use case you want to have needs a different type of toolbox yeah whether that is data analysis image generation uh also rack yeah so retrieval of some information not all of them need access to all of them all at once it needs to have a strong large language model on the one set in English what we often times see as European players you know what's with French what's with German what's with Spanish we became extremely viral in Latin America as well because we had some great creators from from from down south and uh they brought masses of students to us I'm still amazed by how this happened um today this not is not so much of an issue anymore a cloud 3.5 son has great response time

great multilingual generation is overall quite stable and now in social interactions you can already see comparisons shall I use 01 or I use son most people prefer prefer son for simple task because you get a response in three to five seconds while or one thinks for a minute I remember the first versions of our platform I think they needed on average 90 seconds to respond yeah we had we had these gpt2 fine-tuned models or early GPT Neo Etc models which were fine tune on blog post Etc they were extremely slow they couldn't they couldn't stream tokens stream words basically such that you could immediately get the feeling as a user of value is already being created for me um that some of the users we had back in the days were actually paying for this I'm still wondering but this this were the early days

very slow needed time I think one of the biggest challenges also for agentic AI is solving context as well an agent can only be as good as it knows things about you and we have some extreme users who have uh conversations they are hundreds and hundreds of thousands and millions of tokens long yeah so now I think Gemini is is one model which apparently has two million tokens I'm still looking for proper research where I would say I fully trust that it finds the needle in the Hast stack h i mean I was also very skeptical back in the days of CLA when they released the first model with 100,000 tokens uh today 200,000 tokens no issue now we slowly get into the millions yeah and I think what we see currently with agentic AI and and and core generation yeah look at cursor lovable

and how these to tools are all called they currently have this the similar Hye curve such as AI writing tools now because they became stable enough and iterative enough such that you can use them on a daily uh on in in your daily work but you can also clearly see if you if you check out Hacker News if you check out on Twitter what people say if context gets too large it becomes also difficult to work with them well so context by far one of the most critical issues in this in this whole space before me we move to another topic I I I thought that your vision was uh really good like uh you guys started first with like really doing search in knowledge base like you have this big database of um uh like knowledge base people are uploading their data and then

they are able to search on it and they are even able to like find a sentence in like that uh big PDF file whatever they are looking for I think that's big especially if you work for a big company if you're a junior employee just starting out trying to figure something out like just search the knowledge base find exactly what you need yeah it's amazing I love that idea and I really love the next step like the vision which is um you guys are going to you guys are doing this already uh basically um like learning from this knowledge learning from the work right so if you think about it work is about like inputs and outputs right a customer comes to you they need something you have the skills you have the knowledge and then you there's you give them an output outcome right

but if you think about it agents can actually take care of all of them if they know if they if they have the skills and they have the knowledge and they have the previous uh experiences like if they can see the previous jobs done they can just like take the input uh do the work using their skills tools uh whatever you want to call it use the knowledge and then give them the outcome so the the agent can actually use that knowledge and the key thing here is that have that knowledge have have all those like conversations have all those like PDF files documents that have the docu that have all the information needed and then the agent can take care of everything and I love that like Vision um if you want to talk a bit about that mission yeah that would be great

well thank you very much um I mean the vision is ultimately also requires to be infused with a little bit of reality this this is something which a lot of people who are also newly entering new Founders who are newly entering the space uh where now it's it's it's amazing because many people talk about it U many companies do something uh it's a little bit there is there is a VC here from Berlin yans from9 Capital he made a great comparison in a blog article uh comparing the adoption of these AI tools generative AI to the adoption of crms and this is what happens already yeah there were in the first versions of crms companies tried to build their own stack which if you're a machine producer somewhere out of the forest of Germany you shouldn't build software now Volkswagen is also not doing the

greatest job with their software um if uh and and or or you buy for example a software which is a little bit premature is not developing fast enough uh uh we we often times have conversations of how will the market develop some of them do buying decisions now uh some of them also do building decisions at the moment yeah one of our largest Ducks it's a German the German Stock Exchange company huge multi or International Company everybody knows the brand um they had built their first the first agent interface themselves yeah a good six bigger amount of money they invested in this uh it was built on GPT 3.5 I I believe which was immediately outdated when they had done it uh the problem is they hired some Agency for it nobody could update this API call to a different model this is in in

in in in in the in the viscosity of of an Enterprise uh difficult to handle so we came in now as vendor as Tech Service as Tech provider and service provider as well um respectively you need to I mean there will be plenty of demand to fulfill such a vision of building a big and Broad um agentic AI Hub and platform there will be in in my opinion for any Market above let's say 50 to 100 million inhabitants think about Germany think about France think about UK each of them can host two three multi-billion dollar companies in in the space the US can hold several of those some of them will at one point go together and become the next Google if they make it easy enough um and this is also what we play on at the moment sometime it sounds a little it

sounds a little bit too too too German but look at sap this one te tech company which software tech company which Germany has uh is Unthinkable to take away of erps why because they have the best customizability this product is so complex that they need to build the services around it AI agents are also complex yeah our goal is still to have it you know easily usable that's that's usually what I say why does B2B software need to be like a brick I use or I I'm we are required to do some ESG reportings Etc and there are platforms which you simply cannot use cost a good good amount of money yeah we don't pay for this at least yeah but that's where we always this is this is my example why we said a bigger Enterprise also just consists out of individuals give them

a pleasure to use and they will return and they will continue using it this is also how we beat co-pilot many people compare us to co-pilot because the Microsoft 365 subscription is not far away H let's be honest um but many times people started using our product and they started simply loving it because it's built for usage it's not built to get sold as fast as possible and I don't know with some arbitrary USP some but essentially also I think it would be a good Segway into the product Vision too um so we built the product I I mean before that let's proface this um I think in a couple years I think nobody will question which model they're using and it will still it will still be only about the value they get out of it they essentially want their response or the thing

they're working on to be as accurate as possible and is just AI agents also are just just a lever for that I mean AI agents essentially do what they yeah think with the reasoning models they reflect on the previous answer they find some revisions to be made and they improve that response and yeah give you a better one say eventually after you wait for maybe two two minutes um but eventually the value you get out of that is higher than zero shot um use case but similarly how we build the product also to give them a much Superior ux that's one thing but also the I think this this reference should be emphasized more a little bit more because like when especially in organizations like Enterprises like you work with large large amount of data and you have all that complex access levels so you

already have such a customized and sophisticated structure in your organ organization and that also needs to be mapped out to our environment where you can reliably work um and eventually get reliable results and accurate results but we still give uh we allow for for example like error in in the AI responses that's why we give you the most transparent UI on the side that you can immediately reference and just navigate to that file or that source of information and just compare the response uh to that Source whether it's accurate or not so this kind of like it seems very small but it's such an important ux that we want to a huge deal just over this small thing it's it's not even related to AI itself uh it's just some small infrastructure Edition that we built ourselves just point to a right reference it seems

small but it's a crucial um crucial feature that might prevent um huge failures I I was once talking with my co-founder when we were were uh in our summer camp somewhere where it's warmer than current Continental Europe it's currently minus 3 degrees here uh and we were talking about what would be the perfect customization of an AI and I do you remember those Furbies those hor not horrific horrific describe something beautiful like nightmares of these things but usually you know if you want to have be beautiful customization to a human being everybody from us will have you know from from the get-go there's this small thing which we because how we develop which will also help human beings more or less develop educate themselves and this this knowledge this memory this context of whatever this this this thing this toy has collected will just change

device device is from Bish I will briefly take your laptop you know this is the current Western way to work if you look more to self or or or or to more developing economies you have phones you have tablets where people work on and also I mean what is the hardware device of the future Jun would love that at one point we have an engineering department which does Hardware because he comes from the hardware space originally we're currently laying out the the tracks that this company can have enough cash cost to also do R&D maybe at one point into this into this space if you think a little bit like what happens in 10 to 20 years down the road I bet glasses and chips here who knows maybe it's a contact glance yeah we don't really know um I actually had a question regarding

something you both had like alluded to a two-parter but kind of in the same realm uh you essentially talked about how you feel like some of your key differentiators as you have gotten some large um customers were usability you know when you're referring to uh co-pilot and also the customer service that you provide obviously people don't necessarily on a public level actually probably associate customer service with SAS outside of when they're complaining that the customer service is bad um for companies they like so how how have you found at Tex cortex that you've managed to uh make both of these things um you know like at the Forefront and and what have you done to make sure that it is so effective I think this this comes from two spaces I mean first of all I grew up in those German festival Marquees tents yeah

these what you usually know from October Fest so I come from a Hospitality service You're simply nice to people yeah you try to create honest value for people so this is kind of like a source route where this comes from sure but then also just the expectation of uh delivering people something which creates value this requires and holding uh and then I mean it's a little bit of it's also the American way a little bit that's what what I have socked up in terms of American education on how to build great businesses technology is one part customer support is the other thing yeah you need to to some degree also care about the people who are using your platform and you can do this better on a B2B in Enterprise scale then for example with the millions of users we have this is where we

said okay what is the next best level to educate in masses this is blog content this is also video content you're well a well aware about what we are doing there um and showing faces there are so many there were we've seen so many AI products which sold some shady lifetime deal and you knew that if people buy this for these 40 50 bucks or it ran down at 1 Point up to N9 a lifetime deal for AI tools where you know this has a lifetime yeah with the API calls it does to the big service um at one point it will be done and dusted uh I can also just recommend anyone who is in front of a buying this ision thinking about is there a faith to this company do they have mention that they care about that you will also build something

with it that you will gain value out of it there's enough also we had faceless YouTube channels who will tell you oh this is how I made 5,000 with CET GPT this is also not the right thing to do yeah it's it is a journey it is a service now and this is one big I think this this is one of the big USPS you can have in your company setup that's plain simple that's why we also decided to build such a way that's why even us too are still joining uh bigger Enterprises in their slack and teams Channel and we support employees right on the ground where the problem happens yeah no lock behind some emails or whatever people pay us well it is the least we can do to aim for you to be retentive or to expand your account that we give

you a great that's where this whole thing come from fair enough yeah and what about the the other part U the usability um you compared your own uh to to Microsoft 360 or sorry it's called co-pilot it's funny that I stuttered on remembering the name I think that speaks to the the failure of that hitting the mainstream yeah I mean it's I mean there's co-pilot there's co-pilot Studio when we started integrating one drive for some of our customers there's also storage products in teams in one drive personal business it's difficult for people to understand um now but simple things we have this huge Automotive supplier 880,000 people globally uh they named their AI assistant Marvin and I've once visited them in this uh yeah it's not a small City stutgart somewhere somewhere in southern Germany uh their headquarters is huge uh I was part of

an internal conference about what they are doing in Innovation know representing our product and there was this one lady who came up to me and said because there was some she was excluded of early te of early testings for some reasons I can't remember at the moment but she came up to me and said Dominic I miss Marvin they gave them the identity one part it was it felt personal it was useful yeah it was also alongside the values of the company build and then simply uh one of our first product Visions still when we were in plain writing to more or less as a Chrome extension we always said we want to be in every text box there H in German there's the saying um you know the the bone is not coming to the dog the dog needs to come bone and if

we look currently at how many people need to work today is that they bring their work to Chet GPT instead of you know the tool coming to wherever they are working and that's where where much of this this usability and and product Vision has built on ultimately because to be honest uh you you might know this example if you go to the kitchen yeah before you went before you go through the door frame um you have a very straight goal what you want to get out of your fridge as soon as you walk through this door frame you have forgotten what you want yeah if it was a drink a small snack or whatever it's the same thing if you switch those steps and I'm one of these horrible hoers who have five Chrome tabs open with 40 tabs each yeah and I can't find

where I last time worked yeah so that's why I say please I need this uh right where I work yeah that makes sense I think um having used the product and worked with you guys I think it is a very usable product and I I really like the uh maybe you can speak to this as well as a Founder adakin but I really like the customer service um just to you had you you said it in such a a way that was just felt genuine regarding the appreciation you know they pay us well it's the least we can do you know like they're that's I think a lost on um a lot of people especially like some some of my generation I think even gets into the workforce and they they assume you know their value is inherent so that means that the correlative um

economic value is you know just required to be to be given to them but especially when it comes to a business that's something people don't understand like you have to provide value like constantly um else you know the money will stop coming in it's not the same as having a job so I think just as you know just as a business owner having that you know approach is um is invaluable because it it sets the ethos um for the entirety of your company so I think that was great yeah I think uh the customer service is going to be change a lot uh with with the with the AI especially AI agents that they're going to be able to take care of many of the things um and uh and and people are okay with that uh people are okay because they're getting their answers

they're getting their resolutions and uh as the service providers we can focus on like their success uh more like making sure you know building a rapport with them uh making sure that we are really understanding them improving our product while the AI agents actually provide the customer service and so we are actually also working on this AI product I don't know if you have checked out the previous episodes but uh it's going to be released next month in February it's called jot from AI agents and it's like uh basically it has like many different channels I saw that you also had this uh like chat bot I think I think you also have a chatbot but it will have the chatbot mode like phone agent mode like you you can call uh WhatsApp we just built WhatsApp as well like you can just uh have

a conversation with the brand with with the company over WhatsApp so we are also building these channels and uh we started the journey because we were actually like um uh we were actually thinking about like hey why don't we make forms uh like why don't we make make it so that AI agents can actually help people fill out the forms like just have a conversation and fill out the form right we started that way but we saw that beta users were actually using it for mostly for like customer service so we are kind of like also improved there uh but it's it's really exciting I love working on AI uh and so I've been working on this for over a year now I feel like working on an a product is kind of like a taming a beast it's like you have this Dragon but

you want to make it do things product productive things like you you want to make sure that it does things correctly and you can't do that uh it's just like you just have to spend the time like make sure that it's working well uh test things out like uh give it to users see what happens and just uh just spend uh some time on it and like build a good product so it's it's usability it's also about the usability like it's um basically working on the user experience um right uh just putting those uh guard rails putting those like structure around this uh this Beast so you have this Dragon right uh if you if if you're not careful it can burn you as well but you want to be able to ride it so you know you you can do something productive like you

know maybe burn your enemies or something right but uh I think a is amazing it's just it has so much power and every day it's actually getting more and more power you started much much earlier than us like uh just like uh you've been doing this like two and a half years now and you already released that product two two and a half years ago so you learned so much so I also learned a lot from you so thank you for sharing your experiences as well yeah that's well put well put I mean I would say just one thing I mean all should I include myself yeah I mean all of us are founders here I also founded a couple side things uh but I mean essentially you guys deal with stuff that nobody could handle and it just uh went up to you and

those kind of stuff you deal with and I think similar thing also applies in customer service use case so we have different layers like I think we will still have I mean for the next three years I would say we still have humans in the loop um just overseeing and AI customer service Bots I would say will deal with uh the first layer operations and then maybe they move on to second layer more complex stuff I don't know I want to cancel my subscription for example I think they those kind of AI agents will be able to access your database change the status of a subscription and just update it maybe that's possible that's possible you just need to give right access so that's that's where it gets a little bit tricky so the yeah front injection stuff gets tricky have the human in the

loop right just tell it like you're going your refund is going to be issued but then make sure that some human is actually checking making sure that not something is wrong but uh yeah it's it's amazing uh and but what we are seeing is that like all the examples I've seen is like talking about how to issue refunds even like agent force like changing clo but what we are seeing with our users is uh we have around like 400 beta users that are actively using it uh a chat form most of the time it's it's not even like it's it's even like they're just putting this chat bot on their website and people are asking these questions or they're just like giving this phone number to people and then people are calling and asking about like maybe they're running a festival and they're asking about

like when is this conert what's the price where can I park Mar cart uh all kinds of questions it's AI is perfect like it's just like it can answer the questions correctly you know this you have been doing this for two and a half years with uh with with knowledge based search as well but uh like just even those pre-sales or like just uh like public questions are something most of the time actually the companies don't do it because they it's just too much cost like they cannot put customer service people 247 or even like you know uh in regular hours when you call like you have to wait like 45 minutes but with AI it's just instant and the answers are great and it's just uh it's already uh there like it's already providing value to people great the international problem of waiting too

long in a line of a tele telecommunication provider 25 and we couldn't solve this hopefully hopefully there what I'm curious about is also this all these AI sales staff SDR companies how far they will go I think both of your inboxes as business owners are also full with AIS I mean I I I had what's recently the first the first AIS SDR which called me in German somebody was cheaping out who wasn't using 11 laps for The Voice they were using uh the azour standard voice no it wasn't a great experience that was an instant bounce I would suggest oh my goodness yeah we were talking about um those type of sales um well some of the sales products that were funny um on an episode a couple weeks ago and we were just saying like all of these AI products that are trying to

get you to utilize their product for sales you end up having to book a call with a person in order to purchase the product and we were just ding like that doesn't I don't I'm not convinced like all the bat I was so surprised that they are selling as sales product but their ai ai sales product but their AI sales person is not available like you have to talk book a call a real human being and agent force like uh they are hiring actually like one to 2,000 employee [Applause] like yes yes that is uh how would you say you're hypocritic yeah yeah I mean you can also see it in public discourses already that I mean it was an amazing pitch if you saw 11x artisan and their rise it you know if if this AI could do the job of five of your

sdrs uh at 50k USD ual cost maybe it was an easy buy for people to say like oh I want to test this let's test this out this is also why these companies just shot up in Revenue very quickly respectively we see is just gathering around like it's Mecca H trying to invest in those company but now in the public discourse you can see how many people are turning away again uh this is where we then come back to this principle of bring value be a great customer success supporter uh build retention expansion this is yeah sounds like an easy easy function but many people chase the hype I also love your vision of uh seeing customer support uh is like the Holy Grail of of TX cour like if you're giving customer support like you reach the very senior level I I mean that's

that's some tough that to Craig I mean I still I still love to do custom support sometimes I just randomly jump into some emails which when I find a ticket flying around somewhere I shouldn't do it but I still love to do yeah well um I think I think we're kind of wrapping up it we actually hit an hour it flew by for me um I really appreciate your guys' time um I uh I just want to make sure that you obviously plug whatever you you want to plug you know um we're going to make sure that we uh um get you guys uh as much much exposure as possible with your product amazing thanks a lot uh banish by the way let's send them for for the plaque let's this one slide with this comparison for Google Drive send this over such we can

PL this in somewhere but I think otherwise it was a beautiful conversation thanks for having us and it was an honor meeting you I taken as well thank you for uh being a guest in our show and I really enjoyed the conversation if we can help just let us know and demitry also to you we haven't seen for a long time but thank you for all your work recently I haven't joined any call recently B just is managing you so well that I'm just trying to put out fire somewhere else but thank you a lot for your I I have no I have no uh understanding how Founders could ever be busy so you know no I'm just kidding all right well with that being said thank you guys for uh listening to this episode uh make sure to leave us a review on Apple

podcast and watch the video version And subscribe on YouTube as well we will see you in the next one thanks bye-bye