Transcript 0:00 Starting a company is a race in mental health disguised as a career path. People who give you bullshit answers about augmenting only and all that, yes, it is augmenting, but it's going to replace jobs. 0:09 We should definitely be very comfortable with that idea. Bezos famously said, "Your margin is my opportunity," and we famously say, "Your attrition is our opportunity." 0:17 Company that comes in and says, you know, "We're 97% accurate in Arabic," or whatever, those companies should shoot themselves in the face. 0:23 Media headlines are great, but if you really know the technology, you know that English doesn't have that high of an accuracy rate for you to have that in Arabic. 0:29 Voice, it is the best form of communication, and I would argue it's not just the first one that we learn or we've learned as a human race, but it's also going to be the last interface that we actually use. 0:40 The next best thing is gonna be telepathy. If you want to play with, in the big leagues and you want to go where Ronaldo is, a lot of people look at me and they say, "Are you guys a Saudi company?" 0:48 Our company, a MENA-based company built with love in MENA. Welcome back to the Forward Start podcast. This is Jamie Lane, your host, and we have an absolutely fantastic episode for you today. Why is it so fantastic? 1:00 Well, essentially, our guest is just incredibly open, transparent, and not shy of a contrarian or controversial take, which always helps. 1:06 The guest in question is Fouad Jeries, he's the co-founder of Maxam, a MENA-based AI-powered cloud communications platform. 1:12 Now, over the course of the episode, we get into why Fouad thinks his competitors are actually his best salespeople, why most Arabic accuracy benchmarks are complete BS, and why Maxam has stayed deliberately quiet on fundraising while quietly building one of the region's most capital-efficient businesses. 1:27 This is a fantastic one. You'll want to listen to it in full, but as always, before we get into it... 1:32 Warren Buffett once said the stock market is designed to transfer money from the active to the patient, and Lebanese American mathematician Nassim Nicholas Taleb reminded us that you cannot predict, you can only get ready. 1:43 Those two insights define today's investing environment. You don't win by guessing the next move. You win by preparing, cutting through the noise, and letting discipline compound over time. 1:51 That's exactly what Sarwa delivers. Sarwa's modern platform is grounded in simplicity and lets you trade stocks, ETFs, crypto, options, and high-yield savings all in one app. 1:59 They keep your fees low so more of your money goes to work, with minimums as low as $500. Sarwa is loved by experts, not just beginners. 2:06 Whether you're a first-timer, an active investor, or a day trader, you'll join hundreds of thousands of users who choose Sarwa to grow their wealth. 2:13 As a Forward Start listener, you get $200 when you open a new account with code FORWARDSTART. All the details are in the description. And if you're chatting to the team, please do tell them that JB sent you. 2:22 So what are you waiting for? Get started in minutes and take control of your money your way. Sarwa is regulated by the ADGM Financial Services Regulatory Authority. 2:30 The information shared in this segment is for educational purposes only and should not be considered financial advice. Investing involves risk, and past performance is not indicative of future results. 2:39 Fouad, very welcome to the podcast. Thank you. 2:41 Very excited to be here, and thank you for l- figuring out the logistics on this as well, because I know I've put you through the ringer as far as you've come directly from the air force, especially for this. 2:50 There's no other reason you're in the, the Dubai offices. Not at all. For you, anything, man. Exactly. Don't worry about it. There you go. 2:54 But I'd love to start, and I didn't realize the backstory to Maxam and how it actually kickstarted. So I really enjoyed that and the research, and I'd love to get it from the horse's mouth, as it were. Sure. 3:03 Can you tell me about Kashbana? Basha. B- have I, have I done it wrong? [laughs] Kash Basha. Kash Basha. Jeez. So the word Basha is originally a Turkish name. 3:12 It's Pasha, but we don't have, obviously, Ps in, in the Arabic language, so it turns into Pasha, and that is usually used with, you know, saying to somebody, "Oh, you're a Basha, you're a boss." Really? 3:20 So the cash boss is basically the, the service that we built. Oh, fantastic. Yeah. So the genesis of, of Maxam happened when we really felt the pain. 3:26 Sinan, my co-founder, and I have been working together for over a decade now. Lasted longer than most marriages, as, as I would say. Yeah. But, but we've done a series of startups together, and we were in payments. 3:35 We started looking at the region, obviously, seeing what problems exist in this part of the world, and in a region that is very, uh, border, you know, uh- Yeah. Fragmented... 3:42 heavy and fragmented, you wanna take a look at what kind of, uh, things that, uh, you wanna erase between borders and, and- Mm-hmm... 3:48 big businesses are in payments, are in logistics, and it just so happens that also, you know, e-commerce is a big part of that. So we started building a payment solution. 3:55 That payment solution evolved into something that was more into logistics. Yeah. 3:59 After that, we took a look at customs problems that were there, and then we realized that 80% of people who want to buy anything online actually want it from an international website and not a local one. Mm-hmm. 4:07 So obviously, all these, you know, it's a big mess and big difficulties in people getting to buy stuff. 4:12 We built a service called Kash Basha, and that took some of these technologies that we built, and it was really a, a kind of an experiment. We were a bit frustrated. You know, we built a lot of great stuff. 4:21 We went to the souqs at the time. Yes. We went to all these other players, and they're like, "You know, uh, we'll put it into our pipeline. We'll see you in six months." 4:27 And we wanted to eat our own dog food and get something up and running. And so with some of the facts that we knew and some of the difficulties that we knew we can solve, we built this service. 4:34 And very shortly after, became one of the heaviest trafficked e-commerce websites in Jordan. Basically, we built it on the, the shoulders of Amazon- Mm-hmm... 4:40 and one of the giants in this world, and we allowed anyone to buy from Amazon, get it locally to their doorstep with the final price right there. 4:46 No surprises in shipping, no surprises in customs, and you can pay in cash, which is the preferred- Yes... you know, form of payment at the time. 4:52 After that, we had more customers in six different markets at the time, so we expanded the business. But they needed to call us, and we needed to call them. 4:58 One of the essential things in every business, not just customer service, and Amazon is probably the quintessential company in the customer service world. You know, you want to be able to delight your customers. Mm-hmm. 5:07 And doing that over the phone, over the, over voice communication is one of the best ways of doing it. 5:11 There was no solution for us as a small to medium-sized company at the time to be able to speak to those, these customers, get local numbers in these different markets that we have, them call us locally and us call them locally, be able to record it, unless we went to a legacy system like Cisco or Avaya, any one of these guys. 5:27 Yes. We were quoted about $170,000 at the time- Oh, jeez... 5:29 to kind of set up this, and it's gonna take months of setups 'cause you have to have local data centers, local registrations before that, obviously, in every single country. 5:36 You have to integrate into the telco infrastructure there, put it over the cloud, have a system that can actually accept it, and then you can route the calls as you wish, either to a remote team or a local team. 5:45 And that was just- Headaches everywhere... super head- super, super... Big pain. Mm-hmm. And so we took that pain and realized that there's a massive opportunity here. 5:52 The region and the fragmentation of the region is actually an opportunity because of the regulation that's, that's present in this part of the world, because of these difficultiesA lot of the international players that are all billion-dollar players in the space don't come to our, this part of the world. 6:05 Why is that? Well, first of all, because of the, the difficulties. Complexity. The, the complexity. But also, I mean, to them, the Arab world is a rounding error in terms- Sure... of them doing it. 6:13 So we talk about lo- localization. Mm. When you talk about AI startups that are looking after dialects and all these different things, like the dialect problem is a symptom. It's not a disease. Yes. 6:22 The disease is Silicon Valley. The disease is, for lack of a better expression, Wall Street- Mm-hmm... who isn't investing in this part of the world because it wasn't one of the big major markets. Yep. 6:30 And so actually building products from the market to the market and being able to service these customers was not something that they did. 6:35 And we thought that we can take it onto ourselves to be able to build something that solves our solution. We talked to a bunch of our friends who are in the startup community. Mm-hmm. 6:43 They came back to us saying, "Hey, we want the solution and we wanna use it." Alhibbi was one of our first customers, Tamatem, all these startups that were in the same facility- Yeah... 6:50 that were in the business park in Jordan at the time, and we just tried out stuff and they caught onto it. Their customers were delighted. Mm-hmm. 6:57 The voice, you know, communication was much better than most other means of communication. And we started expanding that. 2019, we decided to split off things. We bought out our investors in the previous business. Mm. 7:07 There was an acquisition for KashBasha and the asset of KashBasha by Dubai Group that then took it to Africa. We're very happy with that. It's a great success story for us. Awesome. 7:14 And then the genesis of Maqsam came to be, and here we are. Was the 7:20 ambition always, okay, once you kind of discovered the size of the problem as related to what Maqsam was solving for, you thought, "This is what we're gonna dedicate our time to"? 7:27 Was there any temptation to continue both concurrently? Yeah. Let me tell you, first of all, in KashBasha, like I said, the chain of payments- Mm... 7:34 logistics, there's a lot of opportunities, a lot of billion-dollar opportunities- Yeah... 7:36 in fact in that, and we were one of the first people, by the way, before Tabby and Tamatem to introduce buy now, pay later in the region. Really? Yes. So on KashBasha, we actually loaned out a few million dollars- Mm... 7:45 from our own balance sheet to be able to prove that buy now, pay later works, and we were incorporated in the IFC Fintech Hive at the time, and we- Good kudos... 7:51 talked to DF- DFSA and all these, uh- Was there even licensing available for- There wasn't... 7:55 no, there isn't And, and still I think they told us that there wasn't and they didn't know how to license it, and so we'd come to this gray area with them. Sure. And banks weren't, like, open to it. 8:02 We had to have underwriting. The more money that you want to lend out- Obviously, yeah... obviously, that you need, you need funding for it. And so that wasn't available. The small, you know, $100, $200 tickets- Yes... 8:10 weren't an opportunity that they thought about. They wanted to do car loans and home loans. And so it took maybe someone like Houssam Arab- Yes... or Tabby to come in and build that. Absolutely. 8:18 But even if you ask Houssam, maybe one of the first iterations of buy now, pay later in the region was when I showed it to him at his office- That's crazy... uh, when he was back at, at Nabchi, yeah. 8:25 We're very good friends. They were also one of our first customers at, at Maqsam as well. That's great. 8:29 Uh, so yeah, it's a small world and it's a very close-knit community that, that came together, but that is a billion-dollar opportunity, right? We said we, we tried this out. We had a product called Jaib. 8:38 J- J- J-A-I-B, which means pocket in Arabic and also about nine other languages. Yeah. 8:42 And then we went into actually trying to build out something in the logistics space, and we had the telecommunication, you know, contact c- center as a service type of solution. Mm. 8:50 And we had to make a choice, and we're like, "Listen, investors look at us. We have too many things that we're running at the same time. We do have a profitable business, got us where we are. 8:56 We do have an exit," all that, all that stuff. Yes. "But is this significant enough to just focus on everything together and try and sell different companies, uh, on, on trying to use this tool set? 9:04 Or do we focus on one thing and we try to make it massive?" Yeah. And we know communication is one of those areas where, like logistics, like payments- Mm... if you erase the borders, lots of magic can happen. 9:14 We were also very intrigued because Sinan and I worked at international companies, so- Yeah... I, my background is in data mining. Originally, I worked as director of data mining at a company called Open Insights. 9:22 He was at Google for many years, and we realized that the value of data and being able to train on a language like Arabic, which is a low resource language- Mm-hmm... in this part of the world, is super, super valuable. 9:32 And with all the international companies not looking at the- Yes... Arab world in that way and giving it the light of day, we thought that we can come in and do something that's very special there. Build our own ASR. 9:40 This is pre obviously the world that we live in today with, with AI becoming- Of course... you know, the next generation of change that's happening. But we thought that we can build our own speech-to-text analytics. 9:49 We can find insights that are jewels for businesses. Mm. Turn the profits, the c- cost center to most companies into a profit center. Mm. 9:55 And that was the main goal, and we thought that if we can have this d- distribution, we can build all the layers that are separately sold in the US- Yes... 10:02 by different companies like Twilio and all these different players over there. 10:05 We can probably have a turnkey solution for the first time where a business is incorporated, and next day they can have telephony across the region and can actually expand their businesses. 10:12 So we are an economic engine for a lot of these companies in, and in many different ways, either on the customer service side or on the sales side. Mm. So we decided to spin off and, and, you know, jump- And-... 10:20 into the deep end with Maqsam... haven't looked back since. How? What, sorry? Haven't looked back since. Haven't looked back since, no. Been, been going at it since 2019 now. 10:26 Can you talk to me about your relationship with Sinan then? So you mentioned you've been working together for quite some time. Yeah. 10:32 That's unusual in many respects that a co-founder relationship would persist for, for that length. Can you talk maybe about when you first met him, how that relationship has developed over time? Sure. 10:42 Why you work so well together? Because we're very different. I think that's, that's part of it. But let me take a step back. So Sinan was a founding engineer at Jawaker.com. 10:49 It was the biggest car gaming website- Oh, yes, yes, yes... in the region. Sold for $220 million. Huge success story. I was part of a company when I moved back from the States at the time in the early 2010s called D1G. 10:58 Mm. Spelled D-1-G. It was basically a rich media website for entertainment content, and entertainment content, even on YouTube at the time, was not in Arabic. Yeah. Mm. 11:05 It was very dispersed, and obviously we know the problem with the richness of the Arabic- Yes... content on the internet, still below about 3%, I think, to 5% is what most studies say. 11:13 And so we used to have lunch together every day in the cafe that was underneath our building. Mm-hmm. Coincidentally, our, Maqsam's office is right across the street from that building right now. 11:20 But we used to have lunch every day because some of the engineers at D1G went to school with Sinan, and then we became very close friends. 11:27 I think what's special about this relationship that we have is that it's not about us being alike. It's more about us being different, as I said. Mm. I'm very much a machine gun, and Sinan is very much a sniper. Okay. 11:37 The approach that we have to how we see problems, how we think about different things, how we share the responsibility of building a company together is something that I think completes each other. 11:46 You know, it's very complimentary. I can definitely tell you that, you know, starting a company is, is maybe a race in, you know, mental health and management- Yes... disguised as a career path. Mm-hmm. 11:56 And so when we think about things, we disagree a lot with each other, but we know that we have to disagree and commit, and sometimes it's about-Not about who's right, but whose turn is it is right now- Mm-hmm... 12:06 to be right, depending on the situation that we're in. Context. Yeah. And so we know so many things about each other. We've become great, great friends, brothers. Mm-hmm. 12:12 And we see a long-term, you know, development in us building more, more products and more companies. And he knows more than maybe, you know, my fiance or my mom knows- [laughs]... about me. 12:20 And we've been through a lot together, whether it's on the mental side or on the personal side. Mm-hmm. You, you mentioned there decision-making. I'm curious about that. 12:26 Is there any times where you've vehemently disagreed with each other? Oh, yeah. All the time. Drives me crazy. [laughs] So I mean, some things happen on the direction of the product sometimes, right? Yes. 12:36 So, like, this is off the top of my head. So when we're on different sides of a spectrum, I'm the type that wants to jump out of the plane. 12:44 He is the type that maybe wants to do the logical wise thing, is like, keep everything as it is and then make, you know, certain- Yeah... surgical enhancements that can actually get us to the next phase. 12:53 So I'm always thinking about, all right, so we are in a world where we're building... Let me take a step back. There's two worlds within every single company, right? Mm-hmm. 13:01 There's the world that you're building today and the world that you need to prepare for in the future. Yes. Right? And the same thing goes from when you kind of go for funding, right? 13:09 You're pitching the present case- Mm-hmm... and you're promising a delightful future that's about to come around. The difference between those two points is usually your integrity as a, as a founder o- overall- Yes... 13:18 and what you fight for and what you try, try and do. 13:20 But usually companies, in my view, who don't do well long term are the ones who stay with the former state, which is trying to build for today, trying to make things work and all that, and don't actually take risks that are a bit more calculated. 13:31 So we built different functions inside the, the company, whether it's a skunk works type of department where we can actually experiment with things and do things, but we don't want to have it, you know, radically take over the rest of the, uh- Yep... 13:42 of the conformity that we have within the, the system. 13:45 And so we, we kind of mixed and matched on different things, even if we had disagreements to kind of fulfill this journey that we're all going together and really, uh, uh, you know, allow for some level of not just differences, but experimentation, risk-taking in calculated ways that I think that my preference is more, you know, to be a bit more radical. 14:02 I'm curious, you were very quick to pivot or lean heavily into to AI. Was this one of these projects that you had running concurrently in the background? Never pivot, by the way. 14:13 So AI was built from the, the genesis- Session... from the, from the inception of Nuf- Nufsm. So when we were thinking about data- Mm-hmm... 14:19 we weren't thinking about, you know, just collecting the data or harvesting the data. We're thinking about what could we do with the data, what, what insights could- Always structuring it... Absolutely. 14:25 So when the AI agent future that we are, you know, presently- Yes... building right now came about, it's an extension of what we were doing. 14:32 We're simply bringing the AI that was already in the transcription of calls, and the analysis of calls, and the categorization of calls, and the summarization- Mm-hmm... 14:38 all the magic that we had in the back end, we're just bringing it to the forefront of the customer- Sure... right now in our conversation. So our sales pitch before was, "Hey, come to us. 14:45 Three minutes, we're gonna connect you to the world, and we're gonna allow you to grow your business, and you can connect it to your CRM and do all this magic stuff." 14:50 Now the pitch is, "We are a labor as a service company." Mm-hmm. 14:54 And labor as a service is what the attrition that you're experiencing within these very heavy call centers and, and, and tough environments is the opportunity that we, that we make. 15:03 I mean, Jeff Bezos famously said, you know, "Your margin is my opportunity." Mm-hmm. And we famously say, hopefully fame- hopefully famously say, "Your attrition is our opportunity." 15:11 And I believe that there's a much more functional way to running businesses with these technologies. Mm-hmm. 15:16 The thing is, I, I feel like a lot of people look at AI as if it's a, something that's replacing jobs and, and truly it is. 15:22 There's no denying that, and I think people who give you bullshit answers about, you know, it augmenting only and all that, yes, it is augmenting, but it's going to replace jobs. In the long term, yeah. 15:29 We should, we should definitely be very comfortable with that idea. Mm-hmm. 15:32 I don't see why there is a somehow a moral obligation on AI companies to kind of save the market from this, from this happening all, all of a sudden. Mm-hmm. 15:41 You didn't hear that when people, you know, I don't know, like ba- bank tellers when, when e- electronic banks- Yes... came online. Mm-hmm. Bank tellers went out, went out of jobs. Nobody said anything. 15:49 When Uber came to, to be, the taxi dispatchers went out of, out of business or, or the job was eliminated. 15:55 In, in very stressful environments like assembly lines for, for car, you know, manufacturers or car factories, they were, they were replaced with robots. Nobody said anything about that. 16:03 So why is there a particular moral obligation around AI, I think? I think it's inflated by the media quite a bit. But I definitely do feel that it's important for us to realize that this is a 16:14 replacement for bad jobs, not just any job, right? Yes. And it's an upskilling of the human race over, over time. 16:22 And I think this is super important for us to kind of realize and, and kind of ponder on a little bit, because being in a call center eight hours a day with complaints coming through your ears is terrible for your mental state. 16:31 It's... You know, some of these call centers have 10% per month attrition and, and- Yeah... turnover- Yeah... of, of the employees. And that just isn't an environment that, that should remain for a very long time. 16:42 And I think it's gonna be, you know, about, about 2030 is when about 40 to 70% of all the activities inside a call center will be automated. So we're close to getting there. Mm-hmm. 16:53 With relation to the extent to which maybe the media has exacerbated things, I think maybe it's now that it's encroaching on the knowledge economy more than it has in the past. And to your point, for more, 17:03 I don't want to say mundane, but more repetitive jobs- Sure... you know, there, there isn't the same level of concern. But when, you know- Yeah... when ChatGPT can write as well as a, a journalist- Yeah... 17:11 or when it can do these different things. So when we talk about call centers then, like what level... 17:16 And potentially there's an adverse effect regionally in particular in that we do have a lot of outsourced call support here as well. I think of Egypt in particular as maybe a, a hub as far as that's concerned. Yeah. 17:26 As much as it's not your moral obligation or imperative to, to have to be concerned about what we do, but do you still think about it, I suppose- Oh, yeah, we think about it all the time... as much? Mm-hmm. 17:36 We think about it all the time, and we think about how businesses, you know, or, or, or human labor in particular will be dedicated towards more sophisticated type of jobs and, and not completely eliminated- Mm-hmm... 17:46 uh, and still augmented by AI, as I said earlier. And we think about it in terms of, all right, if this change happens far too quickly- Yes... 17:51 I do think that there are entities and structures in governments, uh- Mm-hmm... regulation, what have you, that should, you know, speak up and, and kind of take action. 18:00 Do you think governments are asleep at the wheel?I think they will not evolve as quickly as the technology that we're seeing has evolved. I mean, ChatGPT was introduced in 2022, and now look at where we are. Mm. 18:09 Uh, I still think we haven't seen the full agentic AI promise that we've, uh- Yes... we've seen. 18:14 I think, yeah, now the closest thing that I've seen over the past month is per- perhaps Claude Bot, uh- I was just gonna say MozBot. MozBot now, obviously. Of course, yeah, yeah. Yeah, yeah. I haven't got it though. 18:20 I don't know why they have an obsession with lobster or, you know- [laughs]... claws or, or whatever with this guy. Yes. But yeah, this guy, uh, Peter Steinberg, I think his name is, uh, you know- It's crazy. 18:27 Did you see him on TVPN? No, he did an interview. No, I didn't. Worth checking out, like half an hour. Guy is crazy. Yeah. But it's insane. But, but what he's delivered is something that I think is the big promise of AI. 18:35 It's a glimpse into the future, isn't it? Yeah, yeah, yeah. Yeah. 'Cause, 'cause everybody got excited, and the hype is, is huge. Insane. This is terrible for, I think, for markets in the short term. Yes. 18:42 But, uh, back to the governments, usually what we see is that obviously technology's gonna be fast, fast-paced- Yeah... and, and, and exceed the, the pace of regulation. 18:50 But, but in this particular case, I think it's gonna be much faster than anything that, that we've seen before in terms of what it can take over. So Claude Bot, for example- Mm... 18:56 has not just i- i- implications on the, on the, the labor side- Mm-hmm... but also on the security side. Yes. You're putting this onto your own personal device. It's running natively on your device, does amazing things. 19:07 Mm-hmm. But imagine me through your WhatsApp application on your laptop here, I send you a, a prompt that is an injection, and I can overtake your machine at any point in time. So many vulnerabilities. And that's... 19:16 Yeah. Yeah. Lot of vul- vulnerabilities, especially if you don't know what the hell you're doing. Yes. 19:19 And so a lot of people today are excited about this because, yes, hype and all that, and people are getting Mac Minis and putting Claude Bot on it- [laughs]... 19:25 or, or MozBot or what have you, and it's, it's really leading to a lot of vulnerabilities and access to things that people didn't want. 19:30 And I think we're gonna live in a world where there is gonna be a dark time where, uh, the naivete, I think, of the general public and, you know, the lack of regulation will, will unfortunately not meet in a place that is gonna be very friendly. 19:42 What does your own AI use look like? Obviously, when the hype of ChatGPT started out, I was using ChatGPT quite, quite frequently. I've since moved to m- more towards Claude. So have I. 19:51 Uh, Claude is definitely, I think, superior on many fronts. I would tell you I've played around with obviously a lot of open source models- Yeah... 19:57 so OpenAI's OSS and different things of that nature, but, but Claude has a reasoning, I think, that is much more impressive. 20:04 And even me with the mental conversations, you know, having d- different conversations and trying to really test out and try and challenge the, the bot, it acts very close, obviously, to, to human behavior. 20:14 We all know that these LLMs are simply, you know, stochastic parrots, right? Yes. And, and they're predicting the next word, the next letter, the next, uh, you know, response that should- Mm... sh- should happen. But 20:26 the illusion that we have, I think, as, uh, as humans, is that we build relationships that are intimate with these types of tools. Mm-hmm. And that has never been introduced before, I think, in, in humanity. 20:35 The, the relationship I have with Google or even Yahoo back in the day was, was, you know, just me getting, uh, entertained. Yes. The relationship with Google was me going to... similar to me going to the library. 20:45 Mm-hmm. They're introducing things to me, and I'm making my own deductions about what that's about. 20:48 But the opportunity, danger, and relationship that's quite intimate with these LLMs is that I'm speaking to them, and they're becoming my friends, and they're becoming my... 20:57 the tool that I'm using, and they're becoming my opinion, if you, if you will. Interesting. So I'm asking particular things, and it's giving me a direct answer, and that's becoming, uh, you know, the source of truth. 21:07 So this can be super dangerous, I think, in the future- Mm... when you're talking about political opinions, you know, the directions. Sure. 21:13 Even preserving culture, I think, that's going to be done in LLMs rather than just books and, and the minds of people. 21:20 And this is why I would assume, or, or, I mean, I would hope, not assume actually, I would hope the initiatives that are being done in the region- Yes... 21:28 whether they're Jais or whether they're the Alam, Allam, or whether they're any particular sovereign, uh, LLMs, and I think every country should actually experiment with this, is that the preservation of culture moving forward will happen through these types of, of reasoning models. 21:42 It's, it's, you know, humans traditionally have done that through speech at first. Yeah. Right? 21:47 So, so language, I'm forgetting the name of the, of the researcher right now off the top of my head, but language was first discovered about 500,000 to a million years ago. 21:55 Text was only discovered about 150,000 years ago, and, you know, other forms of communication, you know, only about... like keyboards have only been around for like 150 years. 22:03 And so, and so the evolution of whether it's voice or how culture is, is transferred over has, is, 22:10 you know, not changed for a very, very long time, but I think this is the first time where it's going to be done, uh, digitally through these particular mediums. A ton to unpack there. Can we start with voice? 22:18 Do you think that would be the more natural mode of communication with AI moving forward into the future? I, I find it myself, I use Whisper. I don't know if you're familiar with this. Of course. 22:28 I, I can't remember the last time I wrote anything. How do you use Whisper? For everything. So, so I mean, you mean through ChatGPT, you just use the voice, uh- Use it if I'm creating a- Yeah... 22:37 Google Doc, I will Whisper it. I will do the entire thing on that. Yeah. If I'm asking a prompt, even if it's on Claude, Gemini, ChatGPT, I will do it via Whisper. And I suppose it's all about context. Okay. 22:46 It's the amount of information that you can supply. Yeah. There's a tendency, I think, when you're typing, you're like, "Do this." Yeah. "Please." Yeah. I don't mind chatting. 22:54 I will provide as much context as is required. Yeah. I'm happy to chat. You know, the... I think the context window for Whisper is like six minutes. Yeah. I can't tell you the amount of times that I've exceeded that. 23:01 [laughs] Yeah, yeah, yeah. I think... No, I'll actually, I'll actually tell you my opinion on this. It's, it's if you look back into human nature- Mm... you identify your mother's voice within moments of being born. Mm. 23:13 You identify their face within, I think, weeks of, of being born. Voice is naturally something that humans have grav- gravitated towards. 23:21 It is the best form of communication, and I would argue it's not just the first one that we learn- Yeah... or we've learned as a human race, but it's also going to be the last interface- Mm... that we actually use. 23:30 The next best thing is gonna be telepathy, right? And that needs a whole different slew of things that we can, you know, dive in. Elon will figure that out at some point, right? Right, right, right. So the neural- Yes... 23:40 brain link, uh, is, is, is, is all into that. But arguably, the best interface is voice, just because it's the easiest way to express yourself. Yeah. 23:48 I mean, if you think if you're talking to your, you know, your life partner or anyone else- Mm-hmm... how many times over chat have you misunderstood each other and have to put in emojis or whatever? 23:55 The emotive- Many times. [laughs] Yeah, there she is. She's sitting here off camera. [laughs] But, uh, but yeah, you know, voice is the most emotive medium. Mm. 24:03 It is also the fastest medium to communicate what you want and potentially get an, uh, a, a result. For years, I think, you know... 24:10 Why, why are LLMs so important is because prior to LLMs coming into the world, we had to figure out-... machine speak. Mm-hmm. Right? We had to put in commands, prompts, code, what have you. 24:20 Now hu- the, the machines, you know, know the human speak. And so that is where this innovation is, is just so important. And we are expressing ourselves the exact way that we want. Mm. 24:30 And we're able to kind of do things much, much quicker through, through, through voice overall. And especially in our part of the world, that is the preferred, uh, medium. 24:38 There was a time, I think, where a lot of people saw that, you know, chat mess-messaging became a big deal- Mm-hmm... especially with businesses. So, so sometime where WhatsApp was introduced- Mm... 24:46 a lot of different businesses like Intercom and others- Yeah... 24:49 for example, came onto the scene, did very well on the chat interface with, with, with businesses, and thought that that was gonna be the end of call centers and, and what have you. 24:57 But lo and behold, here we are, companies like Intercom and many others are going back to voice. Yeah, Finn, I mean, they've- Exactly, exactly. 25:03 Because you need an omni-channel presence that is commanded by, by, by voice and, and that changes, I think, the game when it comes to customer service specifically, or even sales. 25:12 So yeah, voice is going to be king for a very, very long time. It's nice to have someone to shout at. Yeah, yeah, of course, of course. [laughs] Absolutely. Especially from a customer support perspective. Yeah. 25:19 I mean, it's important listening to Jamal from Governata, previously, previously with Jeeny in Saudi, and to his point, he said this was so important from a customer support perspective. Yeah. 25:26 The taxi drivers wanted to have someone that they could shout down the phone at, and maybe they didn't have that to the same extent maybe at that time with Uber or, or Careem, for example. Yeah. 25:33 Sovereign models, which we touched on there, they seem like something that's happening at the periphery to we're not really interacting with them. Yeah. Mm-hmm. I'd love to get your thoughts. 25:43 You mentioned that they're gonna be quite important, but am I going to be interacting with any of these models at any point? Am I gonna have a cause to do that? Yes. The answer is yes. I think in certain silos- Mm... 25:53 they won't be as popular as, you know, the, the frontier models that you- Sure... you'll be interacting with, for sure. But, um, if you talk to Dubai Police, I think that is already Jais that you're talking to. Okay. 26:03 If you are in Saudi Arabia talking to the Ministry of Interior, then some of the services are on Alam today. But even when you take a look at our company- Mm... 26:10 you know, sometimes we, we've developed our own, fine-tuned our own models, and many other businesses are doing the same. Yes. 26:15 But we do know where if you're building something that's differentiated, that's important- Mm... and where sometimes data sovereignty is, is a big deal and you have to deploy locally, that's also going to be important. 26:24 But we don't shy away from knowing that if there's something that's better on the shelves, make sure you implement it in the right way- Yes... and you don't kind of redesign the, the wheel, rebuild the wheel. 26:33 This is something that's available to everyone, and I think that's part of the reason why, you know, a lot of the, the models that are being built by, by these sovereign, you know, entities are not getting enough popularity is because the distribution is, has not been presented the same way. 26:45 They've been done top down by government, they haven't been put into applications- No... 26:49 they haven't been really deployed in a way that is, that is going to be touching every single citizen the same way these sovereign models will. Yeah. A dry press release is not quite the same as- Exactly... for sure. 26:58 Exactly. We've touched on models there, so it'd be remiss not to touch on your own models, which I know you've been developing for quite some time- Yeah... before it was popular in mainstream. Can, can you talk to that? 27:06 Can you talk to me about the process from when you first started training models? Absolutely. So, so first thing we did was a text to... S-sorry, speech-to-text model. Mm. 27:13 So we wanted to transcribe the voice calls that were happening on our platform and be able to analyze them, and obviously Arabic was really underperforming. Yes. 27:20 Even with open source models like Whisper and others, they were not trained on an Arabic data set. And so we had a golden data set that we had from our own business, so from CashBasha days and, and all that, and- Okay... 27:29 some other data sources that we had proprietary access to, and, and even the consent of some of- Yes... the customers that we work with, uh, to, to actually use some of that data and train it for them. 27:37 And we were able to build a golden data set, which took a huge team that we had to put in together and annotate and properly label all that together. 27:43 So we got to this golden data set, and we've been able to outperform some of the international companies at the, at, at Arabic language in particular. Can I ask for... 27:50 How long did that take, the process of cleaning it up, annotating? That took, uh, quite a while at that time. Yes. Obviously it's much easier to do this right now. Of course. Yeah. 27:56 Thankfully, we have, uh, uh, great advisors and early investors in our company who work at OpenAI and also speak Arabic and know the, the nuances of Arabic, and they were able to help us and guide our AI research team to be able to do that, and we put out some academic papers on that front that we're proud of. 28:09 But, um, you know, we were able to outperform these models, and I don't think it's difficult for anyone else to outperform these models. 28:15 These models, by the way, uh, you know, any company that comes in and says, you know, "We're 97% accurate in Arabic," and whatever, those companies should, you know, shoot themselves in the face because- Yeah. 28:23 What's the deal with benchmarking? Can we talk about this? Because- I think it's so immature. I think it's so immature that, you know, you come, you... I mean, media headlines are great. Yeah. 28:30 But, but I mean, if you, if you really know the technology, you know that, you know, English doesn't have that high of an accuracy rate for you to have a- Yes... have that in Arabic. 28:37 So I mean, good on them if they were able to do it, but, but honestly, it's not about the, the accuracy of the speech-to-text models. Mm. 28:44 It's about what are the resolution rates that you have when servicing your customers, and that doesn't necessarily rely on accuracy that's very high- Mm-hmm... 28:51 in speech-to-text models, but it does actually rely on good training within the reasoning models that you have and the functions that go into the outputs of that reasoning. 28:59 So a lot of people look at the aesthetics of the model. So of course, any AI agent has three main components. The first one is speech to text. Mm. 29:07 The second one is the LLM, the reasoning side of things, and then the third one is text to speech, which you have this full voice model pipeline put together. 29:14 And to get a good result, it doesn't mean that I'm producing it in the, you know, the local accent. It doesn't mean that I'm having 100% accuracy on the, on the, the, the, the front or the top of the line. Mm. 29:25 It's really about what is happening on the reasoning state there that makes a huge difference there. 29:28 So when you take a look at companies, and even if you were to invest in a company today and what have you, where are your, uh, your data flywheels? Where are the proprietary data sets that you have? Mm. 29:37 Where are the integrations that you have that are, you know, not, not typical integrations? Are you using things that are off the shelf or not? That does not matter. Yes. 29:44 It really matters whether you're going to be a mature technology team that understands that resolution rates and being able to get to a low frustration experience with a customer is more important than anything else. 29:54 That's a good point. You can call- Yes. Yeah. You can call a bank here in the UAE. There's a popular one that everybody, you know, doesn't like their AI agent that they had for many, many years. Yeah. 30:01 And, you know, uh, you'd have a terrible experience because it wouldn't understand what you're saying, it wouldn't actually do what you want. 30:07 That has since changed quite a bit now with the advent of AI, but obviously there are a lot of traumas in the market too. Mm-hmm. 30:12 And a lot of companies on the other side are using that trauma-That's embedded into the minds of all these professionals to get these headlines and, and, and get in bed with, with some of these, the clients, all right? 30:22 But that's all good and dandy. They will sell. They will do very well. Of course. But then again, you're looking at a market where this is new technology. 30:28 There is a lot of- There's various aspects to this technology as well... skepticism. Well, yeah, yeah. But there's also a lot of skepticism- Yes, yes... right? 30:32 And I think why there's a lot of disappointments is because a lot of these companies are dreaming the big dream that- Yes... 30:37 has been promised, and the levels of accuracy that have been put into headlines, and then are very soon to be very disappointed very quickly. The reality doesn't quite match up. Exactly. 30:46 So that is, I think, one of the main things. Beyond that, uh, the companies themselves are not ready for this conversion that's happening, right? 30:52 So the companies themselves don't have the knowledge bases written in the proper ways that can be ingested by these models and be- Mm... 30:57 uh, the know- you know, they take the actions that they want or have the responses that they want. They focus on the aesthetics first, and then they don't get into the nitty-gritty. 31:03 Once they realize that this is a big undertaking, they're like, "Oh, oh, oh, wait, wait, wait. Keep the human servant, and we'll, we'll take... " So, so I don't think the AI change is gonna happen- Mm... 31:11 in the next year, for example, right? Yep. We are seeing today, according to Gartner, it's by 2020... In 2022, we had 21% automation. Yes. However, it'll go up to about s- 75% by 2030. 31:24 But yet again, this is in Western markets- Yeah... where language is not an issue, lots of competition, lots of distribution. Mm. Infrastructure's not an issue, right? 31:32 In this part of the world, you're dealing with a whole different animal. It's gonna take a little bit longer. There will be much more aggressive, many impressive things, but don't sell me on a demo. Yeah? 31:42 Sell me on something that's actually gonna be, you know, where we're positioned, for example, as a full stack solution. Right. 31:46 We're giving you the numbers, the contact center platform for your human agents and for your AI agents- Mm... and then the analysis on top of it that actually allows you to grow as a business. 31:55 And I think this labor as a service kind of- Angle... mentality- Yes... and angle that we have is what separates the, hopefully, the companies that outlast the potential bubbles that might happen or, or what have you. 32:04 I mean, a lot of these co- businesses, good on them that they're able to raise a lot of money. Yes. But, you know, the default state that they have if things go down is, you know, they don't have the customers- Mm-hmm... 32:14 and they're just building out right now. They're selling them on the dream. And number two, if there is a bubble, if it bursts, they'll default to bankruptcy. 32:20 We'll default to thousands of customers in tens of countries, thankfully, and, and hundreds of millions of calls being processed with AI embedded into that. Yes. 32:28 And potentially being able to outlive the doomsday scenario that many of them will be erased by. 32:32 And so, I mean, I'm sure you might ask me, like, "Do you believe that there's a bubble that's gonna happen or, or what have you?" We'll get to that at some point. Yeah, yeah, yeah. [laughs] Yeah, yeah. 32:39 But I mean, if I were to, to project my answer a little bit here- Go for it... uh, yeah, uh, I w- I would tell you that I, I do hope there is some level of correction- Check it... in, in the market. 32:49 I don't think we'll, we'll see a, a massive bubble that, that bursts the same way we saw earlier, I think, in the history of- Yeah... the technology industry. 32:55 But for sure, there should be some level of, of, I think, maturity that needs to happen, and the companies that are able, hopefully, to, to outlive that scenario, and hopefully we're one of them, will, will, will do well. 33:06 Eventually, the C-suites are gonna start asking for, "Where, where are these incredible results that you told me were- The market has spoke to them... for tells?" Yes. I mean, the market- Yes... 33:12 if you take a look at the US market, for example- Mm... there's a lot of- Yeah... skepticism around the hype that has been promised. Okay. Talk to me about resolutions. 33:20 How does the pre-agentic world compare to now the agentic world we live in as far as customer queries being resolved? 33:29 Is this like a step change in terms of improvement, or is it, okay, no, it's, it's a sort of cost arbitrage scenario whereby we can maintain the same level, but for much, much cheaper? 33:38 What, what are we talking about there? So the numbers from research that I've found at least, and again, these are Western numbers. Sure. We might, we might have to kind of discount them a little bit- Caviate, yes... 33:46 for, for the region. But it would cost a contact center call- Mm... in general, cost the company about six to $12 to resolve that call. Mm. 33:53 If you have a ticket, and especially if you're in government or enterprise, that goes up to $40. 33:57 So imagine $40 for you to be able to resolve one particular ticket, 'cause it's gonna take time to get, you know, calls back and forth. The human agent is there. The systems that they're using are there. Yeah. 34:07 The costs can pile up, but the company doesn't see it that way. Obviously, there's a call center that has to be there. Yes. But if you go into the nitty-gritty details, it's that expensive. Right. 34:14 When you're looking at an agentic world, you're talking about, about half a dollar to a dollar to do the same exact resolution. Significant. Absolutely significant. 34:22 And in terms of the cost structure, I mean, that's something that, you know, we would have to battle with every day because, you know, the, the... 34:29 It only makes business sense for, for people who are in the C-suite- Yeah... to, you know, change the orientation of what your sales and support people are doing to be able to save the company. 34:38 It's, I think, the responsibility of those individuals to repurpose a lot of those savings- Mm... into reskilling the- Yes... individuals that are in the company, being able to come up with different jobs. 34:48 A simple example would be if I was a call center agent- Mm-hmm... and I, you know, am eliminated, you can use me again for training more 'cause I'm a skilled call center agent. I've dealt with customers. 34:58 I know what the problems are. But you can s- uh, upskill me to be able to retrain some of the models that are specific to our, you know, type of business. 35:04 And, you know, these general templates that you find online with some companies, they don't actually fulfill everything that has to do with your particular business, and there are a lot of nuances that, you know, are, are involved in that. 35:14 And I don't think it's gonna be a massive issue where we see a huge change. Where big changes happen in, in, in reskilling that's, that's needed, for example- Yes... 35:20 if, I, I believe the statistic is, you know, the, the 40% of jobs around the world require you to drive a vehicle. Wow. 40%. 35:29 If that change happens very, very quickly, especially at the level of people working or education or, or what have you, of the level of society working in those jobs, I think you're gonna have a big problem, and that's when I think the, the governments will have to definitely step in. 35:42 It's a big market for Waymo. You're really reinforcing the centers. Waymo, Tesla, every one of these guys, right? Yes. So, so massive opportunities, and I think it's gonna be atypical solutions that bring this together. 35:50 How are you going to do it? A lot of people talk about minimum income and, and- Yes... and, and things of- Yeah... global, global minimum- Universal... universal- Yeah... incomes. Yeah. 35:58 And I think that's gonna take a while to kind of- Flesh that one out... yeah, yeah. Yes. Or be effective even. Mm. We, we talked about cost there and, and cost of resolution. 36:07 It, it makes sense to talk about pricing at this point as well, because AI has precipitated outcome-based pricing like we haven't seen before. Yeah.Are you guys practicing out-outcome based pricing at the moment? 36:19 Can you talk to me about- Sure... how things have changed? Sure. I think that is an eventual future that, that we will mostly land at. Mm. 36:25 We have experimented with some models that, that we took a look with borrow numbers and that make sense on outcome-based pricing. Yes. It's similar to when you think about pay-per-click advertising. Yeah. Right? 36:35 So pay-per-click advertising, you're paying for per-per-performance. If I'm implementing an agent, I would definitely want to lower the barrier- Mm... 36:41 of, of entry into a company, have it live there, and if it doesn't solve the problem, then you don't pay me. Yep. If it does solve the problem, then you pay me very handsomely, and at scale, that is a lot of money. Mm. 36:50 For us, we are currently on the consumption pricing standard. Sure. Right? Yep. Which has been hereditary- De norm, yep, yep... in our business. 36:57 And I think in the place that we're in in the market, that is more understandable to a lot of customers than actually having outcome-based pricing. Mm. Because there is skepticism around the AI market overall. Yep. 37:06 I do know of other businesses that are doing package-based, not outcome pricing, based pricing, but actual package- Mm... 37:12 based pricing, where here's a number of AI minutes, and you can consume them over a different pri- piece of time. People understand that a bit more because they're used to the telco-based pricing- Yeah... uh, generally. 37:21 And I think going all the way to outcome-based pricing, you know- Maybe a small cold turkey... is-... too abstract at the moment- Too abstract, exactly... so you want to... Yes, yes. Exactly. Yeah, you want to- Yeah... 37:27 to simplify it to a certain extent. Yeah. Okay. Yeah. Let's talk about fundraising then. Sure. You don't make a big song and dance about funding, which is unusual regionally. Why are you guys so quiet? 37:38 We thought that, um, it wouldn't aid us in any particular way by, you know, announcing our numbers and how much we raised- Mm... 37:44 especially in a market that already has potential competitors that we've been able to, you know, outgrow and, and kind of get to a point of, you know, a, a good business. 37:51 A good business is a good, is, is a good business. We haven't raised a lot of money, I think- Yep... compared to a lot of businesses, thankfully. 37:56 We have achieved profitability, and we've been able to grow out the business and, and the metrics quite, quite a bit, and we're, we're lucky, I think, to be able to be in that position and choose who we wanna partner with and choose, you know, the, the position that we're, that we're in. 38:08 But, yeah, I th- I thought it would be anti-climatic to our overall story if we had just went out in, in the world and said, "Hey, we raised, you know, X hundred thousand dollars," or, or, you know, our first round- Mm... 38:19 a few million- Sure... our second round. And, and that's all we did. We only raised those, those two rounds, and, and, and, yeah, I mean, we didn't find the need for it. 38:26 I think there's no particular strategy that is beyond, you know what? Now let's get away from all this hype and, uh, keep our heads down and make sure that we build a great product that people actually need. 38:34 And businesses like us and like other mission critical type of, of infrastructure, they're not very sexy, you know, at the beginning. I think we're now much sexier- Yes... with the, with the advent- All of a sudden... 38:43 of AI and all that, all of a sudden. But I mean, if you take a look historically, yeah, you guys are doing call center stuff and whatever. That's, that's not super sexy. That's not like- Yeah... 38:50 buy now, pay later, we're giving you free money and, and, and all that stuff. And so we thought that let's keep it on the low key- Yeah... 38:56 and make sure that we can grow out the business and capture potential pieces of the market. But we continue to do that strategy. I don't know. Mm. 39:03 I think when you get to larger rounds of funding, it's the investors and the- Yes... 39:07 the, the PR glitz and glam that a lot of people wanna come to, but we would remain wanting to be a company that is, is gung ho on, on what we're doing today and trying to promise this future. Execute and just- Yeah. 39:19 You touched on competition there. It's becoming increasingly crowded. How do you think about competition? Our competitors are our greatest allies, actually. I'm so happy with my competitors. 39:28 They're our best s- our, our best salespeople. Um- Expand. 39:31 I mean, a lot of these companies end up doing such a terrible job on the infrastructure side and on the sales side and on the support side that a lot of their customers come to us because they want the solution- Mm... 39:40 but they're not happy with the peer s- well, well, the company that they're working with. I respect a lot. Don't get me wrong, I respect our competition. Sure. 39:46 And they only validate our market more fur- further and further, and depends on who we're competing with. Again, the stack that we- It's... Yes. It's a good point... that we, that we built- Yeah... 39:54 is, is four layered, right? Yeah. So you have the infrastructure, the contact center as a service, the analytics, and then the AI agent. So you'll find a lot now with the glitz and glam of, of AI agents- Yeah, yeah... 40:03 at the top, but not doing the rest. Mm. And a lot of these players are allies, as I said, because they integrate into our infrastructure. 40:09 They integrate into our systems, and when the agent doesn't know how to answer, well, you have to send the call to a human, and where does that happen? On a cloud platform like us. 40:17 So we have a lot of the, the local players now, whether in Saudi, the UAE, Egypt even, integrated into our platform and taking use of a lot of the infrastructure that we've built so they can give a better s- service to their customer, and we're happy to be a part of that critical infrastructure that makes their business work. 40:32 Mm. We don't necessarily wanna cut them off and say, "You know what? We're only exclusive. We only do things in house." Yep. And, and that's the strategy of, you know, other companies- Mm-hmm... 40:40 that we've seen that don't necessarily do well, uh, and we wanted to be able to kind of accommodate a better AI coverage and, and see whether, you know, our own models, hopefully at the one point in time, are preferential to customers that they would actually come over themselves. 40:53 I'm seeing a lot more voice orchestration platforms- Yeah... at the moment. How does that fit in with you guys? 41:00 How does that complement, or does it jar with, or w- I, I, I'm trying to figure this out, and I, I thought best, best to ask you. You, you've better, better idea than I do. Well, I'll, I'll tell you. 41:07 I mean, voice orchestration is something that everybody's gonna end up doing, right? Mm. So you're gonna have to orchestrate multiple agents together. 41:12 If you take a actual call center and how it actually lives today- Yeah... uh, you have different personalities, different people. 41:17 Somebody's named Fuad, somebody named Jamie, you know, you have Mohammed, whoever, and I think people are wanting that type of experience of variety. 41:23 Don't wanna talk to the same voice all the time when I call into the company, you know? So that distinct personality is- Exactly, exactly... enhanced. 41:27 But there might be preferences for me to deal with one particular, uh- Mm... person who is, like, my best friend within the company and, and service, servicing me or with all I want. 41:34 But, you know, those voice orchestration workflows and, and, and, and platforms are becoming clearly a way that you can expand and integrate a lot more of your, your services w- within companies. 41:43 So that's, that's something that I think is inevitable to happen, and you see a lot of different players coming about internationally that there's no barrier to entry, I think, into that market, uh, particularly. 41:53 A lot of people can just integrate that- Yeah... into your own workflow and, and, and make use of it. 41:56 Not much to say there, uh, honestly, and I think it's just inevitable that you have all these different components when you take a look at the topology of how an AI agent works, that you actually do need this multi-layered, you know, agentic kind of workflow that'll, that'll help you kind of resolve many things, especially when you have large knowledge bases. 42:11 Yes. You'd want multiple agents focusing on different areas of that, even if you hear the same voice on the other end. 42:16 But to be able to reduce latency, to be able to make sure that you're servicing customers even, even better, you wanna make sure that those agents have particular smaller sub chunks of those knowledge bases that they're able to, to kind of cater to. 42:27 How does the experience compare for a customer or a client coming to you where maybe their knowledge base isn't quite up to scratch, or it's not- Yeah... very well structured? Is that a much heavier lift?Absolutely. 42:39 It is a much heavier lift, but it depends. So if a customer is already using Maxum, for example- Mm-hmm... it's pretty much we've built a, a model- Yes... that already has all of your recordings in the cloud. 42:49 And so you press a button and it'll come up with the initial knowledge base that, that you need. And that is something that you, you can customize and change as needed. 42:57 We're building also a training ground within our platform that can actually allow you to change different things over time and kind of speak to, to chat with the model itself and what have you. 43:05 But, but really I'm very proud to say that we've had some examples where, you know, this one-click knowledge base generation, the customer only had to change one line out of pages upon pages of a knowledge base. 43:16 They only had to change one line, and, and they were able to use it automatically in their, in their own, in their own AI agent. Huh. So, so why do I go back to saying, you know, what's the maturity of the product? 43:25 Where are your data moats? Where are the, the, the flywheels? 43:28 How are you building applications on top of it that actually lower this barrier of entry into customers, I think is what's gonna allow for more distribution and a smaller lead time to getting to market. 43:36 There are a lot of solutions I think that can help, but mostly I'm surprised of the amount of, or the lack, rather, of documentation- Mm... 43:44 inside companies and even government entities that absolutely do not have any documentation about how they work. 43:51 And so this is an opportunity for a lot of businesses, I think, to kind of do some spring cleaning on the inside and kind of up their game, make sure that they have the right documentation and the right knowledge bases because it's gonna really, really change the way that they work. 44:04 You've been at the coal face of model developments over the course of the last number of years. Has Maxum benefited from those step changes over time? What is the ceiling that we can reach? 44:15 Within our company or within models in general? Both. Your company first. So we focus very clearly at customer service- Yes... and outreach, right? So we're not trying to do everything under the sun. 44:25 We wanna make sure we're, we're, we're focused on those. To be able to predict what a customer wants before they even call in or they've- Mm. And that, I think that would be the ultimate evolution of our product. 44:33 We wanna make sure that, you know, you don't have to kind of realize that you need something from your bank or you might like this stock or what have you for you to buy it, or, or if you have a problem, we wanna be able to predict those. 44:43 Predictive. So if you're talking about predictive analysis within- Mm... 44:46 LLMs and reasoning, if we know enough about your behavior and that's fed into the models, we initially think that that's gonna be the, the, the way of the future. Globally, I think, you know, there- When's AGI? Yeah. 44:56 I think that's really far out, much further than, than, than what, you know, the, the headlines will say. A lot of... I mean, you hear 2029- Yeah... 45:04 being, being that, that number, but w- I think what is AGI is, is the big question. So do we think that models are gonna come up with theoretical, you know, scientific theories- Yes. 45:12 Innovation in science and biotechnology. Yes... I think very, very clearly humans have an ability that- Mm-hmm... this technology, even if it seems- Yes... human-like, just does not have. Mm. 45:21 I think being able to string multiple ideas together to come up with a net new concept- Novel that... Yes... novel idea is something that they won't, uh, come up with. 45:29 I mean, do you think a model will come up with the, you know, theory of relativity- Mm... you know, out of, out of the blank? No. It's, it's been trained on a certain dataset. 45:37 Uh, you know, if that da- dataset is not novel, that's a whole different problem. Yes. But if it's on dataset, it's gonna predict the next page of that, uh, dataset. Mm-hmm. 45:45 And will that dataset actually have a whole new idea that wasn't introduced before? I think we're a bit further away from it doing that. And do we think it's gonna do it on its own, I think is what, what people expect. 45:56 So is AGI gonna come in and then we have this orb entity- Yes... that is going to all of a sudden spew out- Hal. Uh, yeah, Hal, for example, but spew out all these new concepts that we haven't seen before. 46:06 Uh, I don't know. 46:07 There's a lot of controversy within this AGI issue because if we are creating something that's gonna be smarter than us, then how's that any different from, you know, a deity or a potential religious figure, for example? 46:18 I'm not getting into too controversial here maybe, but- [laughs] But, but seriously, I mean, what are the differences? It's gonna give us... It's all-knowing. It's gonna give us all the answers. 46:25 We're gonna start reacting to it as if this is a- Yes... yeah, a higher power. Yes. Doesn't have a physical form. It's- Well, yeah. Well, I mean, it does... it might have a physical form pretty soon. Yes. I mean, yeah. 46:36 You're seeing a lot of... Yeah. This is a good, this is a good point. Yes. We're all going to have helpers of some description, android or humanoids or, or, or whatever it, it may be. Yeah, or a girlfriend. 46:44 I think there's, I think there's, I think there's, there's some of those already, I think. Yeah, yeah. We, we, we certainly won't into... I, I, I don't want this flagged. Humans are weird, man. 46:49 Let's not get this flagged by YouTube as, as, as much as we can. I think we'll, I think we'll avoid that. You, you've been increasingly spending time in, in Saudi as well. How's that going for you? Fantastic. 46:59 I live in Saudi. Oh, you're full-time in Saudi now? Yeah, yeah. Okay. Yeah, in Riyadh. And I travel throughout the region. 47:03 We have offices in, in multiple countries, but, but Saudi is obviously, you know, where Ronaldo lives. So, uh- That's a great point... 47:10 so if you wanna, if you wanna play with, in the big leagues, you wanna go where Ronaldo is, and I think- Totally fair... that's, that's where it is in, in different industries. Say no, say no more. Yeah. 47:17 But, but Riyadh has been a amazing, you know, place to watch because- Yeah... I first went to Riyadh in 2010. Okay. 47:24 And it was a very different beast, you know, a sleeping giant, as many, many people will say, for many, many years. And now things have changed drastically, whether it's on the cultural side- Yeah... or, or otherwise. 47:33 But opening up to the rest of the world and seeing the influx of attention has been, has been really, really amazing to see. The vision that... 47:40 The vision 2030 or the vision of His Royal Highness Mohammed bin Salman overall, and also being able to actually have proper, for example, AI infrastructure in the country. Yes. 47:49 I think there is a lack of places around the world that have the political power, the financial capabilities, and also the space to be able to build so much. It's, it's happening at light speed. 48:00 I think, you know, I, I, I just have my fingers crossed that everything goes well in the region and, and, and they achieve what they're looking for. 48:06 And I do think that a lot of things are happening ahead of schedule in terms of the achievements of the 2030 vision and, and, and, and what have you, and they are putting the mon- their money where their mouth is. Yes. 48:15 I, I also do believe that they will need more help. So, so the world is coming to Riyadh- Mm... uh, in a very n- very new way, and there's a concentration of attention and, and financial gain from, from being there. 48:26 Of course. I never thought I would move to Riyadh, you know, f-Back in my 2010- Sure... uh, brain, especially full-time. But there's a wonderful li- well, life there with a great bustling startup scene. 48:36 And, and Saudi entrepreneurs are absolutely amazing. They've done great work. They are learning super, super quick. Not that we come with any... We're all learning together. Of course. 48:46 You know, obviously, I mean, we don't have any particular advantage over anyone, me being Jordanian originally. 48:49 But, yeah, I, I think there's a lot that has to change, I think, in terms of the, the venture capital side of, of, of Riyadh. And, you know, in Arabic, venture capital is Yes... it's 49:09 just the mentality that's, that's there, and the overall level of competition and the way that the mindset is, is, is put together that... 49:16 Or the mental model of actually being an entrepreneur is, is, is, yeah, super important. 49:20 Is there anything to that point that you'd like to receive more funding or that you'd like more investors to take a punt on that you think maybe there's reluctance to at the moment or- To us personally? Yeah. No. 49:31 No, just, just generally. Sorry, just generally as, as, as an ecosystem. I think as an ecosystem, a lot of people look at me and they say, "Are you guys a Saudi company?" And I think that's something that pisses me off. 49:43 This is a big thing, yes. I think it pisses me off- Yeah. I've talked to quite a few people about this... for all the right reasons. Yes. Right? 49:48 And, and it's not about Saudi, it's about, first of all, what makes you a Saudi company or not. 49:52 And I think if you take a look at generally in the world, if somebody starts a business in Europe and then moves to Boston or California or everywhere, you know, and they grow their business and they have most of their customers in a particular market, and, you know, they start hiring from that particular market, if the founder is not, you know, American- Yes... 50:09 does that not make it an American company? It's a great point. I, I think you see the same thing. I mean, just thinking of an Irish example- I mean, we have the majority-... Stripe or Intercom, like- Yeah, yeah. 50:15 Just to name a few of us. Exactly. Exactly. So for us, you know, we are clearly not Saudi founders- Mm-hmm... but very proud to be in Saudi Arabia, proud to be in the region. 50:23 I think unfortunately over the years, again, back to the fragmentation of the region, there's an identity crisis in who owns what and, and what have you. Recently, there was a, a Nobel pr- Nobel Prize winner. Oh, yes. 50:35 Yes. Yes. Yes, yes, yes. And you had three or four countries- This is quite funny... fighting over him. Yes. I think Saudi won, but, but, uh- In the end. [laughs] Yeah, yeah. In, in the end. 50:42 But again, people were claiming left, right, and center whether this was Palestinian, Jordanian, American- Yeah... and, and Saudi. 50:48 So I think we should be proud of that diversification and take a look at the MENA region overall as a unified market. Mm-hmm. This is obviously me talking- Of course... 50:58 you know, unrealistically in, in present times and political, you know, situation over the, over the, over years. But I do hope that we can have a unified identity somehow. 51:07 Again, we speak the same language, different dialects, yes, I understand. Cultural differences, very slight. Sure. Mm-hmm. 51:12 But because I don't dress the same way doesn't mean that I'm not part of this and contributing heavily into this particular market. 51:19 And, and it makes me proud to, to, to be there, and I'm happy to be able to call our company a MENA-based company. Yes. Built with love in MENA. You'll see that on every- I have seen... webpage that we have. Yep. 51:28 And, and, and we believe that's the, the right message. I wanted to ask you about Jordan as well, because I hadn't realized how intimately involved with the startup ecosystem there you, you had been. 51:40 I was chatting to Ibrahim Manna from Bricks. Oh, yeah. And he mentioned you were quite an integral part of Tech Tuesdays in Amman- Yes... all the way back in the day. Can you- Yes, and so was he. Yes. 51:50 Uh, and, and many other people. So I had moved back to the region in the early 2010s. 51:54 I was in Boston, and every day you'd had an event that wa- you know, that brought together either startup ecosystem or a particular scientific ecosystem, or there was something exciting- Yeah... to do every single day. 52:03 And I just realized that this cross-pollination that was happening between all these groups was so valuable, that shaped my thinking, how I was exposed to different companies, how I was exposed to different founders, how some of their advice to me kind of gave me shortcuts into building- Mm... 52:16 a company and, and, you know, avoiding obstacles and what have you. When I came back to Jordan, you had great companies coming out of Jordan. 52:21 You know, technically I would, I would, I would maybe overstep the boundaries and say that, you know, Jordan was a hotbed for many, many years. Oh, absolutely. Uh, in many- Looking back, I mean, it's insane to see... 52:29 yeah, the Gulf talent and the engineering talent- Mm... in most telco companies or, or even startups today. Venture capital as well? I mean- Venture capital as well. Yeah. Yes, yes. You know, happened from, from Jordan. 52:37 Yeah. We had a great educational program that, you know, ga- gave a lot of engineering talent to a lot of these companies. And, you know, we had the first acquisitions. We still- Yes... 52:44 probably have, you know, the most significant acquisitions that happened actually started from, from, from Jordan. 52:49 And in the early 2010s, it was a very, very special time where there was a lot of excitement, a lot of new investors coming around, and we thought that it would be great to have a grassroots type of event happening. 52:57 So we called it Amman Tech Tuesdays. Happened on the first Tuesday of every month. And went on for 10 years from 2010 till 2020. Wow. Right where Corona came in. I mean- Something happened that year. Yeah. Yeah. 53:07 [laughs] Something happened that year. Exactly. But every year on our anniversary, we... It was obviously not sponsored. It was all grassroots. Mm. 53:14 We, you know, our first event had 400 people, then we grew up to about 1,000 people per event, and it was just people interested, enthusiastic about learning and, and being able to mingle. 53:22 And, you know, when I go into a lot of meetings today, people are like, "Oh, I attended Tech Tuesdays. I remember you from there." Or, you know, it was great for our exposure overall. 53:30 But what I wanted to, to say is that we got up to about 40,000 attendees- Wow... over, you know, aggregate over, over 10 years. And it became, like, one of the big pillars of, of, I think, Amman's tech scene- Mm... 53:40 uh, that people looked forward to, and I didn't expect it. Ibrahim was one of the people where when I put out on Twitter that, "Hey, I'm thinking about doing this thing," he showed up. 53:48 He was doing a bunch of different startups, some failed, some worked- Yes... and, and things o- of that nature. 53:51 I met a lot of other people that I wasn't exposed to, I think, before, having lived in the Amman maybe bubble that I was a part of, or even being in the States and coming back and just seeing how the internet was shaped in the minds of other people, whether it was cultural or technical or otherwise. 54:06 And I'm so proud of that being part of my journey, and it was completely extracurricular type of activity. See, see the legacy of that today. Uh, yeah, yeah. 54:14 I mean, that's- But I think it was important that, you know, we realized that neither I or Ibrahim or many of the other great people who were part of that, you know, non-for-profit activity- Yes... you know, owned it. 54:23 It was- Yeah... completely owned by the community. Completely. Everybody contributed. If I wasn't there, it still went on. And I think when youThink about startups and, and your worth in the business. 54:31 It's very important for you to be able to be completely useless at some point in time and, and have people take things on and get out of the way of, of how things can, can go. Still miss it. 54:39 A lot of people will say that, you know- Bring it back... we should start it back up. Yeah, but it's, it's, it's a lot to deal with. Sure. I think I've paid my dues on, you know- [laughs]... 54:45 things that it's for the community, and I think- Someone else's turn- Yeah... to take it on. And not, not just someone else's turn, but I think we're contributing more- Sure... Blaheem, I- Yeah. Yeah... 54:51 all, all these guys- Yeah... and I'm so proud of all these businesses, by being able to either invest in each other's companies- Yes... 54:57 or even just to have these companies grow and, and employ more people who came to those events, who, you know, were interested in, uh, in college at the time, and they just transferred over into, you know, whatever it is. 55:07 And I have a very nice photo of the home brew computer club in California. A black and white photo. 55:14 It's a auditorium with people who, you know, are very techy coming together on a frequent basis in California, you know, talking about all these computers at the time, personal PCs at the time, and a photo right next to it of Amatic Tuesdays. 55:25 They're almost identical. Right. And so, yeah, simple format, 20 minutes for people to give talks. There were three talks and then a panel, and coffee and tea, and, and off you go to the next month. It's fantastic. 55:34 I've seen... I don't know when it was. I went down an awful rabbit hole, and that's how I actually first found out that you were a part of it, was I saw, like, a photo, or it was a video from... I don't know- Oh, wow... 55:41 where I stumbled across this. Yes. I went down pretty deep. Oh, the dark web. [laughs] Something like that. Sure. What was... So was Twitter the, the main source of... 55:50 Was that the social network or thing you were happening? Yeah. Yeah. Twitter was just coming onto the scene. Yeah. And there was a group of people who were doing tweet-ups in, in Amman. Okay. 55:58 And, you know, there was a certain group who were just enthusiastic about social media at the time. Facebook was a completely different thing. 56:03 I think people who were on Facebook at the time and people who were on Twitter were not the same type of people. Different personalities. Yeah. Yes. Yeah. 56:09 So you'd find a bit more active activist types on, on, on, on Twitter, for example. Mm-hmm. People who were not there for entertainment, but- Yes... 56:16 are there to add something, you know, VCs and, and things of that nature. So yeah, that was the group that spun out many, many things actually in, in Amman, not just Amatic Tuesdays. Okay. Maybe a parting question then. 56:26 Very basic. What are we most excited about, considering we're recording in January, for 2026? What, what are we most excited about? So going back to, I think, Claude Bot- Okay... honestly. 56:36 I think it's one of the most exciting promises of what agentic AI could be. Mm-hmm. 56:40 And because it comes with so many challenges, I think on the political, on the security side, and what people might end up doing- Definitely... I think there's a lot of puzzles that need to be solved there. 56:49 I would say that being able to take that into a format that is actually something that your mother could use- Yes... right, would be something that would change her life without her even knowing this exists today. 56:59 I- Mm-hmm... would probably put money down that definitely my mother does not know anything beyond ChatGPT, does not even know what AI's capable of. 57:06 But once she sees something where she can actually have automations- Mm-hmm... 57:11 that do things for her, her groceries are automated, her, you know, different tasks are automated, maybe recording her favorite TV show automatically, and it just learns how to be her assistant and friend- Mm-hmm... 57:22 I think that'll change all of our lives and kind of augment where we're going. So I think that's a big part of what will change over the next phase, potentially this year and maybe part of next year. Love it. 57:32 Patriots fan? Yes. [laughs] Best of luck. This, this will, this will be aired- [laughs] The Super Bowl will have taken place- Yeah, yeah... by the time this comes out, for sure. Yeah. The best of luck. 57:39 I hope it goes well for you. Go Pats. Thanks so much, man. Thank you. Appreciate it. Really, really appreciate it. This was really fun. Okay. Appreciate it. Sick. Very good. Cheers. [laughs] [outro music]
