Adil Saleh 0:03
Hey, good morning, everybody. This is hyperring it podcast. And, you know, we've seen this all the technology wave from the year 2020 I would say early 2022 to now 2025 we've seen open AI growing up. They've just sending a few models. Initially, we had, like, some text generation tools like, you know, platforms that are helping with emails and all that was, you know, the first trench of AI revolution that we had. And now we are at a point that, you know, AI is now transforming the entire workflows and entire job functions. And, you know, and we had run two or three platforms in the recruitment space. So that was, we were so, so interested to bring some more, you know, I would say AI, genetic frameworks that are completely transforming and automating the way people, people hire, and people you know, have, you know, acquired the talent and all of this. So for that reason, we have the co founder and CEO of of nota, that's that's a European platform specifically specialized for the AI co pilots in the in the hiring and recruitment space, helping companies hiring managers, talent acquisition team, HR teams, to recruit better, recruit faster. So thank you very much, Alex, for taking the time. Thank
Alexandre Duffaut 1:19
you, Adil for inviting me and super happy to to be there and talk about AI and AI and hiring industry. I think it's one of the industry that will be most changed in the next few months, few years, it's going to
Adil Saleh 1:34
be crazy. Yeah, it is crazy. It is crazy. Thank you very much for for coming on, Alex. It's entirely My pleasure. So first I was talking about, you know, how you started a new time. I know that it's, it was a big problem. I know that as a founder, you just not just think about solving the problem. You also think about, like, how to stand out, marketing wise, go to market wise, from a commercial standpoint, how to, you know, have funds. So all those founder thoughts when you were building Noda just a few years ago, what was that? Just walk us through, like your initial thought process, your hypothesis, how you want to go about building a product and how I want to compete in the market that is, of course, ever growing. You know, especially with this agent, take AI, building agents, within co pilots, co agents. People call it with the different names, but at the end of the day, you it's all about, you know, doing it very, really smartly, efficiently, on top of the language models that that are present, you know, out in public, that's the game winner. So how did you think about all of this as a founder back then?
Alexandre Duffaut 2:37
Yeah, so if I went to start back like at the origin, initially what I had I looked into like problems that I met into my past experience, and the problems were the key. Instead of looking at an idea, I was looking to a problem. So the problem was really when you talk with someone like, all the information you are exchanging, just like now, our use case is podcast, but you're always having conversation into the professional world, and you're exchanging so many valuable information that you will never be able to reuse perfectly or to really capitalize on that moment. So how do you get that element and transform that into data? And these data now it's made through report. But how do you transform this data that is good to actions that really change the way you work? So by looking at all different industries that are using notes, we found out that hiring, the hiring process, they are doing notes that are more important than almost everywhere, because when you are hiring someone, you really need to take notes, and those notes will be shared with others. Will take, like very important decision on those notes, and from this decision, you will make a good or bad kind of good or bad decision. That decision can be very expensive. So I think the value of those notes were really, really important. So that was really the origin on how we went there. But then to go further, now, indeed, as you just said and mentioned, is the time where data is not it's not only data that does the stuff. It's actions. I love this period, because we have been in the data for, like, maybe 10 years. Our data was the key to everything. But now it's not anymore data. It's really the action. What you will do behind the conversation, what you will do with all those data, you will need to do actions. And so here the agents are coming. So of course, it's something that you think about, you know, deep in your mind when you're thinking about the product that way, it will go, and you imagine the perfect tool, the perfect solution that could help, like recruiting agencies, recruiters in big companies, what are the process and when you think on how you can help them and improve completely their daily task, you go. To actions, and that's where we're going. We're now having a very, very nice report data, and how do we transform that into amazing actions that will revolutionize the daily flow?
Adil Saleh 5:11
Yeah, okay, so data that drives action, I get that point. It's not just about, you know, having raw data at different places or getting them all centralized, of course, back in the years, 10 years back, you'll see a lot of these recruitment teams, or I would, I would say talent acquisition teams, whether they are working as a recruitment firm or working within within a corporation, they are just getting data from different places, and all the things that, all that they were doing, that they were getting all that data in one place. But now, five, five years later, it was not just about getting the data and doing a centralized or unified that data, but also getting, you know, some manual actions like maybe analyzing that data. Now it's all about, you know, using AI to analyze all of those data that is centralized and give you the next spec sections, or, you know, you know, give some actionable insights out of those data points, that is. But what the AI is, so how do you think? Like, of course, it is a big up beginner from for in this hiring space. I know that a lot of these these companies in from the early, I would say from three years post, in the first three years, they're pre product product market fit. They're struggling. They just try to power fast. But when you get bigger, you need to have SOPs for hiring and everything. So you need to not just hire or fire. You just need to make sure that you hire the right people. You have the qualified people. So there's going to be now it's a big company that's even a bigger problem to get the right people for the roles that they're looking for to make the right, right amount of impact. Since the there is two challenges. There is one challenge that the you know, the technical people, and different roles, they're not easy to to be, to be to meet, like, to get, to find for certain locations, even if that is remote, it's the tech jobs are very much saturated right now, like everybody is trying to switch, because they'll always have like, better options that are paying them, better better culture, better opportunities, better growth. So that is also one competition, but at the same time, if you're not sitting on top of databases, you're not sitting on top of technologies to make the right decisions at the hiring when, when you find some people to have the right ask them, right questions, right assessment, all of this. So how, like, I know Nura is, is doing all of this now, you can make out calls and everything within the platform, but when it comes to the role matching. And when it comes to the neuroscience, sign of side of things like how you actually find the right personal traits for that role. A lot of roles, like customer facing roles, they require specific personal traits, not just about the degree, not just about the experience, not about just about the tech stack. It's also about, you know, the personal choice. So how is that coming along with Noda?
Alexandre Duffaut 8:01
Yeah. So I will start by breaking down a bit like the classic process of recruitment. So usually it starts always with a need. So as you said, like some people will have a need. So yeah, you need a software engineer with 10 years working on Python and two years working on AI, whatever. And so you say that, but then let's say you are a recruiter. You need to go deep down to that need. You need to understand what is the value, how that person needs to fit into the company, what will be deep down, the roles that he will have to do. And so from that customer brief. So when you're an agency, you're like talking to a customer. So when you get that need, you will get a lot of information. It's not only a job offer, but you start with this. We don't start with a job offer, we start with a brief and so there is so much more value into that. So we start with this, and then with all that data that we have in the pitch, in the original need, we'll use that to then do many things. We can generate the job offer. And then, of course, do the classic stuff, get the CVS. Will match the CVS with AI, and that's a bit different from what you had before. So because you you will match it with llms, which usually, you know, it was with NLP, stuff like and statistic data, which was really bad, to be honest, when you see it, what is happening on the big recruitment job boards, like today, it's it's not, it's crazy how bad it is. I'm surprised by this, but yeah, so after you, you get a few people to know, or maybe you, you go and you source them, and here it comes, like you have, like the moments when you need to contact them. So you always do the same thing. You will pre call them. So that's why we integrated, like, the phone system, like it's so important for the recruiters to be able to call, and today, like, except if you had, like, a dedicated VoIP solution, like you had nothing. So now with Nuta, we say, okay, like for the recruiters, they will be able to call from their phone, from the app, the Nuta app, or the directly inside the platform. You record everything. You have a super nice screening interview report that then is always up to date into your ATS. So I didn't mention that ATS is where you will store every data. So it's the database for the recruiter. Is the equivalent of a CRM system for sales, but for recruiter, it's called applicant tracking system. So that's really the key. Also, we were talking about that, and for forever, this, ATS, has been the center of everything happening into the recruitment, if you had the candidates, you could see what was his status on that ATS, can I call him where he is right now? Is he matching with something and everything inside is manual. And that, I think, will be the big change, because if you look at AI, it will be super cool to update automatically those kind of manual entry systems, like old CRMs and all. But I believe that in the future we don't need those, almost. We almost don't need those manual entry system. What we need is AI based systems, and I think that meeting note takers that are the based from the conversation the human interactions, will have the entire knowledge. And from that knowledge we'll be able to take all decision generate job offers, send message on LinkedIn like and do all the process that you were doing before on the ATS, but with an agent. So that's clearly where we, I think, the very near future he is, and that's where we are going also with Nuta. So integrating the all the possibilities to record everything makes the platform really more intelligent, in a way, like you can collect all the data. That's very the first step. The second step is, how do you treat that data? What do you do with it like so you need automation. And for recruiters, they know that sometimes it's not the same. When you're doing a screening call with someone, you need to know like, okay, is that person like, available tomorrow? Is that person geographically located here? Okay, so I don't need, like, a two pages report. I just need a very bullet point and short elements. Whereas when you are with some agencies, recruitment agencies, they need to have, like, dedicated report with nice formatting, with very dense data. They really need to go deeper to know, okay, that person said that he has been a lot involved into sports. So the AI understand that he's someone really that goes deeper into some some actions. And so you need to go so deeper, you know, pre, pre evaluate, in a way, absolutely,
Adil Saleh 12:40
and, yeah, also, as you mentioned, the CO piloting and agentic frameworks and how they are going to replace all of your workflows, and they are just going to be a smarter version of you initially, that, of course, you'll have to train them. So the what is the challenge that you're facing, Alex, while training making these agents specialize for hiring. And then, of course, you have different personas. It's not just about, like, you know, training one agent for everything. Of course, you have recruitment firms using it differently. You have hiring managers within companies HR using it differently. They have like, different goals that they actually slightly different hiring process and workflows. So how you're making your co pilots, really, really smart and specialized. And what kind of challenges have you faced so
Alexandre Duffaut 13:25
far? Yeah. So the fact, indeed, like, we are focusing on hiring, and so already we have, like, we need to split a lot and to have different challenges I just mentioned. So you will find also, like, co pilots and stuff that are doing for everything. And so usually, when you do it for everyone, you lower the quality of the outputs. So yeah, so even us, like, we have, like, maybe, like, we target specifically three to four different ICPs. So special specific people, like services companies, you know, digital services companies, they are hiring all the time. They are doing sales. So it's really niche, not niche, but specific needs that they have, hiring teams, also in agencies, or hiring teams into like B Corp. So basically, all of those has different needs. And the challenge is really understand deep down the use case, what they do in the daily job. I think that's also when you are, you know, doing a SAS. You, you sometimes you don't wander too much. And you think, you assume some stuff. And when you go deep down and you you ask them, you go with them, you understand, you're like, Okay, that's what you need, actually. And sometimes it's simple, sometimes it's hardcore, don't we will not lie about that. So when you know this, what we do is, like, we split. We try to personalize from the onboarding the the environment. So for instance, when the person is coming and we are really releasing a new onboarding like, I think next week, it will be even more personalized. The right, they know we know what they are doing, and so that help us to personalize all the templates they will have. So they don't have to do anything, they don't have to go and configure, they don't have to click any button. They arrive. We know what they need. We know they need an automation that works like this. We know that they will need a template that looks like that. We know that they need to be integrated with this kind of tools, and boom, in like 60 seconds, you have everything set up for your use case specifically. And that makes also, that allow us to have, like, a very high rate of usage, well, I think at 85% of daily usage, like in corporations that are maybe more than 500 licenses. So that's, I think, super good, like, when we see that, we're super happy, and the Okay, solization, I think, would be
Adil Saleh 15:40
okay. So you're mentioning that your agents are used now. They've been adopted by bigger corporations as well, some of your bigger customers as
Alexandre Duffaut 15:49
well. Yeah. Okay, so I wouldn't I think agents, people have used that word for so many things, and to be honest, I think we just thought to see real agents coming now. So you saw manners, you saw proxy from convergence, like, those are for me, like, what I think agents before, it was kind of automations through APIs. So maybe you can call it agent, but they were still limited into the number and the type of actions they were able to do. Now it's becoming wider and wider, and that is clearly agents, agents are, I think, still very shy in the corporations, because, you know, the level of quality of output is super important. So when you are able to measure that there is, like a 70% of success, it's good. But if the AI does a wrong move and have still accessed let's say to your mail and send the wrong email. How do you handle this kind of stuff?
Adil Saleh 16:45
So I mean, keeping, keeping human in the loop is super important. I know that. But like having, having AI, I would say vertical agents to work for you and do all the research. Let's say just for this use case, a recruiter needs to know a lot of things before getting on the call. So he needs to have a quick blueprint of having all the, you know, analyzing all the data points and all the external, qualitative, quantitative data. Both of these sources in front of him have next web sections for that call to be successful. So it's not just that he's sending an email or making a call, it's having human in the loop and making human, making human to review to that before making an action. So you can absolutely do that right with, with, with this, these, these co pilots that that are pretty much specialized for some use cases. Yeah,
Alexandre Duffaut 17:35
completely. I totally agree. But I'm just dreaming about the full automation that one day we will have,
Adil Saleh 17:40
yeah, that will take a bit of time. Take a bit of time. Yeah, we'll see. We'll
Adil Saleh 17:47
see, of course, a lot of a lot of industries, especially healthcare, you say oil and gas, you would say accounting, they need, like, they don't even need 95% accuracy. They need that 100% accuracy. Yeah, you know, so the efficiency is, is the key
Alexandre Duffaut 18:02
kind of safety really high safety issues there, clearly,
Adil Saleh 18:07
absolutely, and especially in these big corporations, when they have, like, of course, data security and regulations, they are, they might not be directly impacted by by the responses or actions the CO pilots might take. But of course, they're, adjacent integration partners. Or there's like, big corporations, big problems, you know, so you need to make sure you always keep human in the loop. So I think for the next, I would say, minimum, five years, it will keep it the same way with human in the loop. And it gets smarter and smarter. Maybe customer support tests, they get automated, completely replaced the tier one and tier two support. But in me to be tech, it's mostly, you know, human in the loop, especially tech and support, it's going to be always human in the loop. But I see that you've got some cool integrations and cool features. I know these recruiters, they use lock so and platforms like these. So it's good to, you know, have them everything you know, pretty much integrated within their system. So they, they, of course, they, they have the insights. They provide the insights. And your co parents or your platform analyze all of those insights and give them the next section for them to review. Or if they, I mean, if they approve to you know, approve to you know, for know how to take all the actions they can, they can make all the actions as well. Yes,
Alexandre Duffaut 19:23
exactly. And and today, like, you have already the possibility to do a lot of things, but with the agents coming in, like, you can imagine things that getting automated a lot more. And one solution, also, if you really want to automate and you don't want the human in the loop every time you're doing a single task, is to make sure once this task is working well, and then you can probably automate it like and make sure you have a very higher level of accuracy, getting close to like the 90 95% and I think that's the path we choose, like, like, others, others. Agents, systems, and for this, it's, it's super nice to be able first to test rapidly what will be happening and how it will work. And the people will also the recruiter will be able to to use it dynamically every time you finish an interview, every time you finish customer brief every time, every, every, every week, maybe. So, yeah, that that's, that's
Adil Saleh 20:24
so I'm not that technical, Alex, but just a question. I was curious. Like, how do you train these? I'm only going to talk about COVID Just like vertical AI agents. Like, how do you train it? There's one way to, you know, get the data from the customer that works like, let's say that's get the best practices, best workflows, from from your prospects or customers upfront, and train your co pilots on those for some use cases, maybe you can have multiple practices or playbooks you can get from users. There's another way. Is there any other way? I was curious to, you know, have it internally before, like bothering the customer. Maybe you can get their websites or get their, let's say, their their product, or how they hire on, or maybe what kind of roles they have open like job boards and all, and you do it internally, doing some experimentation as you're of course, prompt engineering is the real thing now. So you got to make sure you have, once you deploy your co pilot, it's already smart enough get know small customer, and how do you make
Adil Saleh 21:26
sure that without interacting with the customer? Yeah,
Alexandre Duffaut 21:28
so indeed, it's always to keep like the thing as simple as possible, and to prevent like multiple steps. And also, we have another element to put into that reflection. Is like the security for us, customers are super afraid of training AI with their own data, so that's something that makes them really afraid. So we are kind of prohibited from doing this. And this is in our CGV. So when you put anything on Nuta that's not used, your data is not used. It will not train anything. It's secured and etc. So what do we do is like, as you just mentioned, we will collect data from, let's say, the URL of the customer, from what he tells us, from the documents he will share on us, and dynamically, we will apply those elements through a chain of thoughts into like prompt engineering, and those will dynamically be applied every single time, like they need to be. So it's not, it's on the the not the last layer, but let's say the second layer that will apply those, those elements. It can be like the the user can give us, like the URL of of their website they can give us sometimes, you know, they will upload like the value guidelines. You know, when you recruit, you have like value guidelines that you need to make sure this person is respecting those they can have also like job offers with the description of the company, this kind of thing. So all of this will will be possible. And to give more detailed example is now you can go on new time, and you know, you can create templates. But some recruiters, they what they want is like to have a report and a score card dedicated for each and every single job. So once again, you cannot come on new town on anything and just create that manually, like it will take ages, and that's not worth it. So now you come with your files, you prompt, you upload, and dynamically, Nuta will create everything inside the platform. So you will have the scorecard dedicated to that specific job offer. You will have, like, the the right questions to ask you were talking about that just before, like, you know, the question the interviewer needs to ask. You'll have the right template structure, and boom, this is already you can just move into your next meeting, interview and record it like this, and that will be applied dynamically so
Adil Saleh 23:49
interesting. And that means in simple words, if I put it simple words, just like company or recruitment from hires a new recruitment partner, or on boards and recruitment partner, they need to work with Nuta initially for the first two weeks to, you know, get them all the information, feed them with all the information, get the all in all the knowledge and education nota needs to, you know, work with you smartly, and then do it as, I mean, very, very awesome.
Alexandre Duffaut 24:14
That's something we do for enterprise companies. Indeed, like, I think it's important as to keep the quality as high as possible. So for enterprise, we have like a CSM dedicated and account managers that we will come and audit the company, like ask questions, talk with some people in different use case, collect this, and then we generate everything before we put and deploy on X number of people, and that's how we do the processes, perfect.
Adil Saleh 24:45
And what do you call it smarter when it comes to co pilots, whether it's the CO pilots, where you have your CSMs dedicated and doing all the asking, all the question, time to time, making and of course, training, I would not say the word training. Would say, feeding the right information to make those co pilots relevant and get the efficient outcomes, versus these self serve models that people you just go, you ask some questions during the onboarding, and you know they want to use your co agents for different workflows at different tasks that they have. What is more? Smarter?
Alexandre Duffaut 25:18
Smarter? It depends, because on the platform, you can do both, like, we can do it for you, kind of, you know, you come, you're a corporation, and, like, you don't have time, and you don't have the knowledge specifically, so we do it for you, so we know that works, that will work well, etc. Then there are, like some others, like, most of the time, smaller companies, you know, there are five to 10 people, and they, want to do that quick, they have to have hands on. And what's cool, so Nuta is like, you can edit everything. We can completely change everything and adapt it to your own usage. And we saw some use case sometimes really funny. So you have both. I would not say one is smarter than the other. It will mostly depend on your your AI, if you are an AI adopter, or if you are like you prefer to have us to do the thing. So you customize
Adil Saleh 26:10
it, how you customize information, and, you know, set up use cases and templates within Nuta matters, you know. And it impacts amounts. Amount of, yeah, it
Alexandre Duffaut 26:21
impacts the results in the end. Like, clearly, yes, yes, and, and now we do that all the time, like, and we can measure really good ROI, and that's first the goal. Like, you know, you you create a tool. And I heard so many stuff about AI before, like, people using AI when it was not yet, the llms, you know, like, in 2018 or this kind of stuff, yeah. And sometimes they were, like, it was useless application, like the result was not good at all. Like it was, like, average, but they were using it. I just want to make sure that it's useful and that they save time. They have better quality output, they have more homogeneous data. They are more focused during the interviews. And that's our other KPIs. You know, we're tracking, like, after 123, months, we go with them. We have a questionary, oh, okay, are you, what do you think about this? What do you think about that? Etc, etc. And that's how we measure and we improve also,
Adil Saleh 27:12
yeah, okay, perfect. You talk about the measurement. You talk about the having a dedicated customer success with, like, slightly mid sized companies to enterprise companies. So how do you measure the success of an account? How you're internally having data metrics, triggers, success metrics, how you're basically measuring the interactions inside the platform, and then based on those interactions and different data points, whether it's meeting notes, whether it's sales calls, whether it's follow ups, cadences, QBRs, all that qualitative data, mayors, and of course, you can train an agent for that too, you know, giving all the data internally. So how you're managing the success and risk for the existing customers. So it's
Alexandre Duffaut 28:00
for like customers, like the success of the information station at one customer of for Nuta, like for internal sales, like, what, what the I
Adil Saleh 28:08
mean for all of your customers, existing customers, how you're measuring success, customer success, that account is going to retain and expand, how you're okay, you know, indicates close opportunity, yeah, risk, opportunity, risk indicators, all of those. So
Alexandre Duffaut 28:23
it's close to what I mentioned. Like, clearly the key data is the usage. That's the simplest elements, and that's their most real one the concrete. Like, let's say you have 100 people in at one customer. Let's say, after two months, you have 10 customers, 10 users that have been using it. That doesn't smell good, in a way, I guess, when you have like results of like after one month, 80, 90% of the users of people that have been using it daily, it's it's great success. If you are 100% of the users that have at least used it once, that's also super good. So that's the kind of KPIs worth tracking. But to give real precise stuff, it's a daily usage, weekly usage, usage, total account, total count, usage. That's for the usage. Then after, we have a calculator of ROI, so we ask quantity qualitative questions like such as, how long did you took before to make a report? And now, how long does it take you? So that's really qualitative, and we make estimations. And so we were able to calculate dynamically, then the ROI, and we can say, okay, so this month you save 35 hours. You save 100, 200 hours of work using Utah, then you will have also the conversion of this time into like money, depending on the hours paid rate. Yes. And so that's one thing another, another one is also the improvement of the quality of the output. So here it's, it can be qualitative also. So you can ask people what they think about their deliveries. So you can imagine, among like, let's say, 300 people, you don't have the same quality of reduction of reports that you will generate manually. So of course, you will have, like, the new, new ones that will take, like, maybe one I want to have to generate, like, very detailed reports, and it might not be really, quite, really good yet. And you will have some others, like, in 20 minutes. Boom, that's done. Okay. So how do you homogeneous? You homogeneous, like the entire quality of the best ones for the everyone, and that's what you are trying to search. So we have other data that we can keep track, you know, more quantitative, like the evolution of the sentiment analysis per people, or among the organization, the evolution of the question asked also during each meetings for the organization we have like, I think, a dozen of like KPIs like this, when you can keep track and see the evolution. And so when you have a trend that see like, clearly the things that evolving, you also mathematically can know that all of the meetings of the company are getting better and and that's also a way to look at things, because you like to see numbers, not only qualitative response. So I would say, would say that,
Adil Saleh 31:33
yeah, I mean, and all of this is pretty much data driven, like you have systems in place. You don't have to look through these. I mean, you're, you're using product analytics platform like the tech stack. Are you using any kind of platform that gives you triggers, especially for the bigger accounts to your CSM, hey, this, this. This guy has not answered some of the questions we asked for the value ROI calculator last month. He'll just follow up. Maybe dig in. Some reason, maybe he's not using the platform enough. Here's the user of class one month. Maybe this user came into a meeting and shared some budget constraints, and ever since then, meeting the users drop, maybe some, some triggers, some next back sections, as you always say.
Alexandre Duffaut 32:13
So you're totally right, and that is in the checklist. That's the kind of improvement we should do. I saw a few, a few tools that helps a lot doing this. And for now, it's really everything is homemade. We did it internally, like with some some graphs and data, and we do a lot of manual work on this. That's why
Adil Saleh 32:34
I'm building that platform. It's still in the making. I've done more than, you know, 100 user interviews, and we're building a co pilot for success agents, especially for mid market and enterprise, and we're trying to work closely with them, making sure collect data.
Alexandre Duffaut 32:48
And you, you collect the data, right? Okay? And, yes, I'm
Adil Saleh 32:52
not a technical guy, but, but all that I've heard from my CTO, there's a bit of it I'll share. And you might be thinking, hey, this guy knows how the back end. No, I don't know a lot, but yeah, we are building doing the right engineering. We've been doing this for about seven months now. A lot of these our Customer Success space is kind of you say it's not as populated as like, more sales or marketing tools that are building agents every day if they wake up in the morning. So we are trying to make sure that we do it right as as, as you also mentioned, that efficiencies are good, so that is why we are collecting data from all the you know, companies that we spoke to, not on the, of course, not on the podcast, but user interviews. And we have some really, really good design partners, very big companies onboarded, that are helping us to do the right engineering, to build the right co pilots for customer success, and these are the problems that we you know, on this interview, I was just like thinking like, this is, this is something that you can also do and that can elevate your customer success. Let's say your success manager picks up every day and he has, like, four or five next big sections against some critical events that happen with those customers that might churn or might give you an opportunity to expand them next month or next quarter or next year. So you gotta make sure you stay on top of it, and AI is going to do all of that for you. Just have some triggers for you to review, and you just make, you
Alexandre Duffaut 34:12
know that clearly makes sense, because, like, you know, when you are, like, you have a growth like, I think we were, like, 20% of growth, like, months of amounts. I know there are crazy, crazy numbers now in the US, but still, like, I think we get like, 300 new, new users every day. So when you have this like, and you need to track down manually, or at least to do a lot of manual works, like, you need to get the right tools to help you. If I had to do a list of all the tools we're using in the company, like that's getting crazy, and I think the more AI will come, maybe the more efficient they will become. Because, as I said, there are manual data entry tools, still, all the ones that we're using, but when they will be automated, agent based, or at least, yeah, automated, that will be just amazing. Yeah. I
Adil Saleh 34:58
mean, the biggest goal with the. US, is it, I think you you might be thinking along the same line, same line as that, because you are a founder too, and you're also building a co pilot. So the only goal is right now, is to at least like, cut 50, 60% of their time and make them smarter, make them operate smarter. So, and that is why we call them design partners. We work with them closely. We make save their 3040, 50% time, get the CO pilots being better and better with every month, every quarter, and then come to a point where we can replace maybe 10 or 15 or 20% completely, and then rest week we keep human in the loop. And let's say previously, there's a big company, I cannot say the name like but they are close to a billion dollar company, and they're Head of Customer Success, working pretty closely with us, and now they're thinking that they need, like, 30% less people to do the same, same job, and those 70% people that they're working not 100, they're doing like, 150 160% smarter work, and they're yielding more revenue. They're retaining more customers. They're expanding more customers, they're serving more customers in a better way, and that's
Alexandre Duffaut 36:03
obvious when you have like, the data and the trigger at the moment where you have like, let's say, You know what a customer is thinking while before you didn't like, if like, you know, when you have like, 1000s of customers, you cannot be behind each of them, like, it's impossible. So having like, I don't remember the English word like for this, but like, you know, you have like, very obvious points, but sometimes you have like, Shadow points or shadow or shadow events that that are difficult to understand and to see. So I guess that's where you're going. And you collect all this, you show it in the right moment.
Adil Saleh 36:38
The biggest thing with the CS is prioritization. Like how they need to prioritize their task, what needs their attention the most every single day. Because a lot of these companies in the mid size, they have like one CSM had like 7080, 100, 150, 200 customers to look after, like you mentioned, so and and they cannot put on dedicated account managers for those because they not paying much. But maybe a year later, two years, they'll they're going to be paying like, three, 4x more. But you never know, because you're not tracking them, you're just treating them as small customers. So there's a lot to explore in our industry, and it's so interesting to know that how you're evolving in your industries. Because just like in your example, in your industry, let's say, if I'm a hiring manager, if I was interviewing 15 candidates to one qualified now, with the help of AI, with all the contextual information and insights that I have, I will be interviewing five or six people, but those will be so relevant, this will save my 70% and at the at the end of the day, it will save me from the times when I used to say, Hey, I've interviewed 12. They're not so good. But since I don't have more in the pipeline, just like, push it. Push like these two more. Now you're compromising on the quality you got my point. So you cannot put a price on that in your space, right? If you get a bad hire in a company and they realize it six months later, or they tried fixing it a year or two, took them a year to you cannot put a price on that, right? So I
Alexandre Duffaut 37:58
think there is actually like, I think it's in companies at 32k around recruitment, when you do like, a mistake in the recruitment and only can cost up to 32k and then I guess for you in like, the CS market, it's clearly depends on the ticket size that you have, but it can like, for us, it can be like, obviously, something very, very, Very bad. And so, yeah, critical that I can change completely the course of of the way things are going. I think there is a moment, you know, in the evolution of a startup where you you go with the guts. I mean, that's good, yeah, yeah, by instinct, and you talk to the people, you talk to the customer. And that creates your instinct base. But there is a moment where you cannot be everywhere. You cannot, like, be enough, always in touch with the ground, like you try to, but you have your instincts still, but you need data to help, like, yeah, collect and, like, make the right decision. So, yeah, absolutely, absolutely.
Adil Saleh 38:57
So, I mean, yeah. I mean, we had, like, final conversation started, and we went, like, so wide. So I wanted to also explore, like, what makes you super excited in this industry? Maybe product wise, maybe marketing wise, maybe any, any other wise in the year 2025, with given the AI and, of course, some Chinese llms. Of course, if you want to do at scale, the cost is going to be a big, competent so what makes you excited?
Alexandre Duffaut 39:23
Like, for in general, like, what's exciting me? Like, the way things are evolving so fast, to be honest, it's a bit scary, because, like, every morning you wake up, you're like, Oh my God, they're doing this. So, so you're like, How can I compete? How can I keep the speed and all but also that's that's as a founder, it's something that is super motivating, and I try to share that with all the team, to motivate them, to show them that we are in the race with, like, one of the most amazing startup around the world, and that's what it's super cool, like, you're competing with, like, obviously Silicon Valley start. Tips from, from France, like, from, like, even from the biggest like, sometimes we win against, like, a Microsoft co pilot, you know. So it's super challenging and and in in this moment, I think things are evolving so fast. But a good thing, it's not always the money that makes, that makes it all and that's it's what I love. Like, sometimes good decisions can make your product to become the best in the market. With absolutely has,
Adil Saleh 40:30
his has AI has actually get people on the same page, on same table of competition around about with people with maybe all you needed with AI is this domain experience, like you've been there, you live the problem, and that's it. Not a lot of engineering has been taken out. You don't, you don't need a lot of engineering stack, or really experience with the engineering to do the right product and sell the right product and and it is happening here too, but again, making sure you do it first, or you do it different. And in the AI, you have to do it both. So you have to do it first. You have to do it different too. So yeah, I mean, I wish really good luck, and I'm super excited to, you know, see how you, you know, evolve with this co pilot in the enterprise side of things. And by the way, how many customer success managers you have, like the in the sales and marketing and success. So
Alexandre Duffaut 41:21
I will be honest, we have like three people in the sales team, and until it started the beginning of this year, because before we were we were alone, so there was no sales team, basically. So I think we grown it like this, and now, now we're recruiting, so that's the that's the key. But now we have like three amazing people like working with us, and they are like, doing crazy, crazy stuff, perfect.
Adil Saleh 41:51
And they are also doing the account management, like those sales people. So yeah,
Adil Saleh 41:55
we have like one twin supports that are, like, not down comes in to the sales team, we have like, one CSM sorry. Account Manager always mixed up both, and then we have two account executives.
Adil Saleh 42:32
So it was really, really nice meeting. Yeah, it was. It was so good to know, you know, you're something is building big in the hiring space as well. And, you know, I've got, like, deep roots into into the, you know, into this SaaS and B to B SAS. And I've got, you know, should place in my heart for all those founders. Because, you know, I know it's not easy,
Alexandre Duffaut 42:53
same for me, like it's, I understand all of the founders, and I really love talking about that anyhow. So that was super cool.
Adil Saleh 43:00
Yeah, perfect. Have a good rest.
Alexandre Duffaut 43:02
Thank you very much. Thank you too. Say it Bye, bye.