#463 From Corporate to Startup: How DXwand is Driving AI-Powered Business Growth from MENA to the World

Episode 463

[00:00:00] 

Mehmet: Hello and welcome back to a episode of the CTO Show with Mehmet today. I'm very pleased joining me, the founders of DXwand, both Ahmed and Ahmed, the way I love to do it. Guys, as I was explaining to you, I keep it to you to just, [00:01:00] you know, introduce yourselves, like tell us a little bit more about you, your journey, and of course we're gonna later on talk about DXwand more in details.

So, I dunno, Ahmed, you. And Ahmed can start. So, so, so I gotta call Ahmed the CEO and Ahmad, the Ct O so we can, uh, we can, uh, you know, make the difference between you. So Ahmad, the CEO if I start with you. 

Ahmed Mahmoud: Okay. Um, thank you. Thank you Mehmet. For, uh, this show, um, ah, Mahmoud, uh, DXwand founder, um, father of five four, uh, amazing daghters and one son.

And I come from technology background and then enterprise sales. I worked for around 21 years of my career, uh, in several corporates. I ended in Microsoft for seven years and then, uh, discovered some gap, worked on it, and then started the X one. 

Mehmet: Great. And thank you for being, [00:02:00] uh, with us here today. So Ahmad, the CTO, now it's your turn.

Ahmed Elshahawy: Yeah, my name is Ahmed Elshahawy. Um, started my career 20 years ago in IBM, so I was doing, uh, solution architecture, delivery management, and pre-sales across different practices. I. Um, that was in Egypt until 2015 where I moved to Malaysia and, uh, joined, uh, also, uh, leadership, um, roles in, uh, data and analytics, uh, especially in Ernst and Young and other startups.

And a couple of years ago I joined DXwand and it was, I was lucky enough to join in the beginning of the, uh, advancement of the generative AI technology to build our platform orchestra. Yeah. 

Mehmet: Great. And thank you again, both of you for being here with me today. Uh, we share like similar background also as well, so I've worked like with couple of like the big names, uh, in sales and pre-sales.

So, uh, [00:03:00] we, we share, we share a lot and I think this will be a great discussion. Now, anytime I have with me founder or founders, uh, of a company or a startup, the question I'm curious about is what kind of, you know. Missings or gaps, let's say, uh, in the market you spotted and then you said, you know what?

Like no one is solving this. Um, no one is maybe talking much about it, and now it's the time to start this company. So what was the story for you to start the X one? Like, uh, and I would love to hear it like from both of you, uh, honestly. So, uh, maybe Ahmed, you know, you, you start first and then, uh, yeah. You can hear it from Ahmed.

Ahmed Mahmoud: Well, um, started in 2017. Uh, by the way, whatever we used to do in that time is not actually what we're doing now. Yeah, we pivoted a lot. Um, but at that time, uh, I was a Microsoft and, uh, the [00:04:00] conversational AI started to be commercial and available to several organizations, uh, while they were all excited.

But the support, even the basic support. Of, um, the regional requirements, which part of it is, uh, the Arabic language and then the direct support and then the culture of support was not there at all. And that, um, provided for them, like, uh, feeling like they are left behind, but it chose to me that there is a gap here.

So they started to validate that, um. I joined, um, I, I come from corporates work, so I didn't work in startups before, so I, I'd like always to learn. So I joined, uh, an accelerator and, uh, where I started to learn how to turn an idea into a business where we started actually to, to, um, uh, focus on smaller businesses at that time, uh, [00:05:00] to help them to get such technology in the region for the languages that, uh, or total dialects that's spoken in this region to maximize sales and customer retention.

Which we started to sign and then I got the leap of faith and resigned. But at that time, this is what we were focused on is, uh, uh, providing, uh, conversation AI solutions for specifically small businesses. And we end up today having egen tech platform, orchestra or enterprises, corporates, and also sometimes small businesses.

Mehmet: Great. Uh, Ahmad for you, like, you know, like maybe you joined later as I understood, but, uh, what attracted you to to, to join DXwand. 

Ahmed Elshahawy: Yeah, it's, it's, uh, the generative AI advancement, right? It's, um, for me, um, I've been working in predictive analytics, um, uh, which relates to, uh, data science and ai, but, uh, it's a, it's a whole new giant, the [00:06:00] generative AI paradigm, and I love building things.

So, uh. X one was trying to start, uh, a new paradigm in generative AI and just had a great chance to, to excel in the early wave and resolve the real problem of generative AI at that time where they, even the frontier models, was not performing as expected for most of the clients. Just getting a demo to work or a POC was a nightmare.

In the early days. Now it's, uh, maybe it's, uh, it's, it's different. But to build an entire, uh, production grade, uh, enterprise level working capability with generative IH is a new technology. It's very exciting. And, uh, just, I, I, I, I took up the, the role and we, we built, we built something amazing. 

Mehmet: That's great.

Now, um, one thing, uh, Ahmed you mentioned about the pivot, and I think this is something, uh, common, right? Uh, and [00:07:00] it's not, it's not like something wrong, actually, I tell sometimes people, like, if you are not seeing results and you keep doing what you're doing, you are probably on the wrong path. So pivoting, uh, is, is like something, uh.

In my opinion, very natural. Especially when a technology like generative AI comes into the picture and it changed everything. Now I am in what I'm interested, you know, to hear from you, Ahmed, like the moment you started to see like, yeah, like. This generative AI is really something which is not like we will pass quickly.

No, it's, it's a new, I would say, path for companies to follow. Like how this has affected, you know, maybe the company mission, maybe also your roadmaps, like I would love to hear from both of you, but Ahmed, because you know, you founded the company, you know, when I would not say like AI wasn't there. AI has been always with us, but I mean this rapid evolution, so [00:08:00] how this affected.

You know, your vision and mission for the company? 

Ahmed Mahmoud: Yeah, sure. Um, so, so D one was always thinking as a mission that we want to empower businesses regardless of their size to grow and we need to be a growth partner. So this is, uh, I would say a mission that we worked on. Uh, for quite far, and we didn't change it, I would say, per se, but we, we, we have seen the impact of the hype as a change of our paradigm that everyone would want and will use AI eventually.

And I. The businesses need to see how that would benefit them. And then we get some more focus on how we can enable the businesses to grow with AI and how to make this artificial intelligence that is very complicated, whatever needed to get this really into a business scenario. How to [00:09:00] make that. Easy.

How to make that enabled for any business of any size to be able to empowered and to grow. So we, I would, I wouldn't say we changed the mission, but we mm-hmm. Tweak a bit to, to get a bold message out. AI will be there for any business all the time from now. 

Mehmet: Absolutely. Ahmed, uh, for you as a CTO, um. You know, you've been working in data analytics and predictive, uh, AI and all this.

So for you, the moment you saw the new wave, I would say of, of Gen ai, like what was your reaction and how you started to think, okay, oh my God, like, like this is, we can apply this use case here and it will. Solve these issues for us. I would love to hear, you know, the, the technical, the, sorry, the technology side of, of, of the story.

You know, when it comes to, you know, the rise of gen AI and, uh, shaping, you know, the company and where you are today. 

Ahmed Elshahawy: Yeah, first things where we [00:10:00] just, uh, uh, started trying the technology and see the, the, trying to reach the limits on the technology and extending the use cases from the common, uh, chat bot scenarios and see what else we can do.

Just trying to, uh, to understand the boundaries of, of the technology and, and, and, and to keep pushing there, um, so we can, uh, resolve real world problems rather than. Sticking to, uh, legacy technology and that moment I just starting using at that point. What the GPT-3 0.5, the model is not that advanced, but just the, the concept of providing a, a prompt to a model, which can do a lot of stuff, including.

Coding and building, uh, even diagrams and, um, uh, uh, business flow optimization, basic stuff. It was just testing, uh, the limits of this model. At that time, it was very promising until shared. GPT-4, [00:11:00] uh, uh, was released and we a huge hike in performance, which then opened the doors to. Huge use cases. And then I discussed with, uh, Ahmed and uh, uh, and, um, the team in, uh, DXwand mg the, uh, co-founder as well, that what are the potential use cases we can cover.

And DXwand was working on, um, uh, technology of knowledge discovery. Prior to Genet V and it was very technically challenging to get it right with Genet v. We, it's just going to an entirely new paradigm where you can present the real world. Um. Knowledge in, um, uh, uh, like, um, knowledge graphs and combining this with the intelligence of the large language models.

And that's just opening the limits, uh, uh, to, uh, to the sky where you can actually combine the data, the [00:12:00] organization with the intelligence of the model to cover many, many use cases. So we took it from there to build a platform that we can actually cover, uh, significant. Critical use cases in the industry starting from.

Customer service, uh, chat bot, uh, sales assistant, internal employee support, even document parsing and intelligence, and, uh, validating legal documents. Uh, doing a lot of use cases for the enterprises. And surprisingly, it was very, very, um, uh, robust compared to the, um, uh, the infancy stage of the technology area.

Typically, when a new technology arises. The early versions is, is very primitive. Uh, just after strategy four, everything was skyrocketing and frontier models was just getting better and cheaper every day, opening the skies to a lot of use cases. So since then we have been working with lots of corporations on different use cases.

Mehmet: Great. Now [00:13:00] talking about use cases, and maybe Ahmad, you can, because you started it like. 2017, the company. And of course, like you, you pivoted later, but, and there must be like some sweet spots I would say. Which, uh, you know, when we talk to founders, usually you go and try different. Approaches. Right? And you try like different ICPs, let's put it this way.

Um, and always this is very healthy because as Steve Blank always say, we have to go out of the building and start to see where our messaging will be resonating. So for you am interested, you know, from early stages, from business perspective, like what is the moment that the customer say, Hey, hey, stop, stop.

Like. I need this now, like you are solving my biggest pain ever from business perspective. So I gotta start with you Ahmed, and then I gonna go to Ahmed, uh, from this technical perspective and ask him about, you know, what technical pains, uh, you might be also solving. So, but Ahmed, [00:14:00] let's start with the business use cases, like which industries and how this also evolved with time.

Ahmed Mahmoud: Sure. Well, as I told you, when we started, we were selling to tiny businesses. Uh, this moment comes with, uh, w when we started with, uh, what they call, uh, one person businesses like mm-hmm. A person, an expert on something like professional photographer or someone does custom cakes or whatever it is, kind of, uh, small, tiny businesses, is it's, it, it was that moment when, because.

The business is all about your expertise. So the time you spend to build the things is the time you get monetized. And then if you spend half of this time just to interact with clients for very trivial tasks, then your, uh, income will go down. So that was the, I would say the, the pivot, the solid one that we worked on very well.

And. [00:15:00] However, when, when you move forward into the scale of that, it was very early technology. So a, a lot of customers were very, I would say, um, visionary and, uh, risk takers. Let's do that. Others were shying away. And then there was a lot of educational activities that we need to do, which economically wouldn't make sense because the ticket size for a smaller business is very small.

So we had at that time to think about. Other use cases that we can do with the same technology or a bit of evolution to pivot for bigger businesses, which were, we started, uh, one of the cases that we, I would say, um, I'm very proud of and actually feel the impact that we have done is doing this for customer support, for our sophisticated cases, for financial services, mainly trading platforms.

We have, uh, several clients, uh, big clients in, um. Uh, in the region. They depend on that for now, four years on [00:16:00] providing support to traders on the market, sentiment details of each and every stock, uh, IPOs and so on. So that was a nightmare that impacted many businesses in this field of financial services on their productivity, efficiency ability actually to serve.

And we're very glad that we helped several companies of this. One of them is actually a four year client that evolved their business from 3000 customers to nearly a hundred thousand customers without adding a single agent to the customer service. And because they have a huge volume of conversations that are handled with our agents.

That was a great impact that I have seen with this ICP.

We started also pivoting more to see more of your cases. And then we're trying to diversify what we do. Uh, we're doing this for education, for example, uh, recently, the past week, uh, [00:17:00] ministry of Education in Kuwait launched their agent for students and teachers that helps them. Um, um, I would say Learn Britain have some personalized learning in their own pace, uh, and provide educational support.

Um, that was like. A great moment for us in seeing an impact on nationwide and education sector, which is something that impacts nations in general. Uh, so yeah, uh, I give you a lot of unstructured, but that's, uh, a lot of moments that happened there. 

Mehmet: Great. Ahmad, like what's, um, you know, from your.

Perspective, like I always say, the technologist has also his own view on how the technology that you offer might affect the business. Because, you know, and the main reason by the way, this show is called the CTO, 'cause I believe, you know, the CTO is the person who bridges the both sides, you know, the technology with the, uh, business and in, in this role.

So from [00:18:00] your perspective, when sometimes I'm sure like you go talk to customers also as well, so what you see like. You know, some of the use cases from technology perspective more than, you know, just, you know, business outcomes. 

Ahmed Elshahawy: Yeah. So, so, so basically we're trying to. To explore what is beyond standard chat bot, right?

So in the, in the initial days, so like we just nailed the chat chat bot, um, uh, for support and sales support, um, uh, the conversational AI itself, we just nailed this, um, very early, but we then there's significant use cases that was. Very compelling to explore, and that is working with documents and large knowledge bases, um, and how to build indexes, uh, in a smart way to get deeper insights from documents and knowledge that was not possible prior to, uh, Genive ai.[00:19:00] 

So in, in, in predictive analytics or any, uh, data analysis, um, technology, typically it's working with structured data, uh, and non-structured data as well, but not in an intelligent way. Um, so just getting the insights of the relationships between the data across thousands of documents, representing this in, um, uh, uh, a rationally, uh, sensible, uh, way, and combining this with the intelligence of the AI to, for use cases like you have like thousands of documents in your organization.

Do you want to stay compliant? Um, you don't want to have this, um, uh, discrepancy in compliance in any document and contract and employee contract, for example, uh, GDPR or any regulatory compliance, uh, that it takes like thousands of hours and investment from different, um, uh, like, uh, departments in any enterprise to get it done.

Right. So [00:20:00] getting the, the intelligence of Gent VI to resolve this was really. Mind blowing and it's alsos very interesting from technology point of view to get it done right. Building the, using the generative AI to extract the, uh, relation between content in wide range of documents and um, and just to use this intelligently on enterprise scale.

This was very fulfilling when we were building this. So document intelligence, as we call it, was. The most predominant tough therapy, uh, standard chatbots 

Mehmet: right now. One thing, and because just Ahmed mentioned about, you know, the, you know, the work in education and in different countries also as well. So we know in MENA region if we want just to look at the MENA region.

So we have like, mainly the native language is Arabic and of course like the business language is English, but within Arabic also as well. Like, and I've seen it both. We come all, all of us, we [00:21:00] come, you know, from this, uh. Uh, background where we have to respond to RFPs and we have to, you know, sometime the, the specs are written in Arabic, but again, you see the Arabic used is not always the same.

And when it comes, for example, to support agents, again, you have the dialect and you have all these changes. So I'm interested, you know, and both of you like, feel free, like. To, to answer this one because it's like not only technical, I think also from business perspective, it's important. So how, how, you know, this was a challenge for you, like, or is it like, uh, um, something that you knew before and then you designed and you decided also to go to market saying that, Hey, like we have this multilingual multi dialect capabilities in our product.

Ahmed Mahmoud: Well, uh, getting into multi dialect, uh, was something we much aware of before launching to lines because for example, I used to live in several Gulf [00:22:00] countries, so, uh, living there, it's, uh, more of a global community and you understand and see a lot of dialects going there. And if you are providing customer service to that, you don't expect people to call McDonald's and or contact McDonald and say.

In a formal Arabic how, what you want, so mm-hmm. It's an initial expectation that customers, to provide an intuitive experience, you need to speak and understand the language regardless what it is that those are using. And that was the first initial, uh, even at 2018, uh, I would say Quest that we put to differentiate what we offer and.

To work with that, uh, regarding the, uh, generative AI now helps a lot in this, but at that time it was very hard. So we had to work on several data sets to, uh, enable some labeling for our models to work well in several dialects. And from a product go to [00:23:00] market, we had to be a bit more focused on specific.

Uh, industries. So we provide high value, well-trained existing label data for customers to work on, especially in financial services and government. So we, from a go to market at the time, we thought of getting enough focus with a specific industries so we can provide an existing experience to clients to depend on, especially for Mulans.

And then later, Geneva had helped us a lot in, uh. Um, in providing this today with several, uh, clients and reasons. 

Mehmet: Great. Ahmed, like, do you have anything maybe from the technology perspective you can share regarding the multi dialect, uh, aspects specifically? Yeah. 

Ahmed Elshahawy: Yeah. JI, the large language models, the frontier ones, was really, uh, a great help to, to get this done in a more, uh, intelligent way.

Um. Honestly, it was a, a, a big hiking performance, [00:24:00] and even when we built our own small language model, we were able to train it on label data sets for different dialects in less, uh, with less cost than uh, uh, companies used to do with the legacy technology. So yeah, judge, I actually opened the doors for getting the dialect thing right in the, in the region.

Mehmet: Great. Now I want to ask you something, which is, um, maybe just to remove, I would say some confusions. Uh, so data analysis is changing and it's evolving, like. People talks about knowledge mining, right? Uh, so do, do you agree with this and this question for both of you and like how, you know, the data analysis actually is, is being reshaped currently with all the LMS and you know, the, uh, you know, [00:25:00] the additional technologies that we are seeing also blending blended within the LLM space itself.

Ahmed Mahmoud: Yeah, sure. Um, I will give a lot of room for however this some thing. Sure. I would mention we were working in delivering knowledge mining in 2020. And we were seeing that this is happening maybe, uh, because, uh, we don't have access to the massive compute or massive resources I would say that uh, Silicon Valley companies would have.

But we see that this is a solution for several problems. For example, the one I mentioned in financial services, we didn't provide a normal conversation. Ai, we provided a knowledge binding solution because you need a machine to understand. The stocks market and ability to understand, um, Egyptian person or a Lebanese person who talks about the stock with a name that never been mentioned in this.

So it was there in [00:26:00] company's DNA years ago. What happened even after with the generative ai, it made this even easier and gave us a lot of power to provide such components platform that we already have into a greater and faster results. So, but knowledge mining is something we have in DNAA long time ago.

I'm completely that, 

Mehmet: yeah. Ahed, do you want to add something on that? 

Ahmed Elshahawy: Yeah, I think the, um, the matter of having the, the enterprise data, whether it's in like an ERP system or, um, like knowledge management platform, SharePoint whatsoever, or documents or even on a portal website. And combining databases with systems and uh, uh, and files and documents as one rep pository to get your information right, uh, is really something that, um, wasn't, um, applicable before GTV accepting very.

Like, um, limited way, I [00:27:00] would say. Uh, so, um, just asking like an inquiry about, um, a certain performance metric that combines the data from a database and from your ERP system and from a policy document and get an answer right to the user with a visualized chart, for example. This was impossible for Genet ai.

So when we blended the sources of information in the index or the knowledge graph mechanism and other, in this, uh, mechanism to represent the real data in, um, a well structured, uh, way using generative ai, it just opened the um, uh. The doors to different, uh, use cases, uh, that, that blends different, uh, technology aspects together.

And it was all about the accuracy and reliability of this system because it's really easy to get, uh, a small POC working, but to get it working right on a large scale for enterprise grade, [00:28:00] that was a completely different monster. 

Mehmet: Right. And you know, I always tell people, because we, we've mentioning a lot, you know, the.

GPT moment. And, um, you know, it, to me it looks like it's opened doors. As you said, Ahmed, like, it opened doors for a lot of use cases. And I always joke with some of my friends, I said like, you know, that is the moment when the genie went out of the bottle. So once, once it, it was like people were a little bit stuck.

What can we do, like how we can do this using different AI technologies? And then the GPT model came out and it changed everything Now. Uh, I'm, I'm a technology, um, uh, antis by, you know, since, since long time, since my childhood. Uh, but not, I like sometime to play the other role. So some people, you know, still, they have, I would call them concerns.

Sometimes they are like little bit, uh, you know, you know the, these people who like naysayers as they call them, right? Who [00:29:00] would tell you the. Bunch of the things that they heard or maybe they saw about ai and I won't now maybe, you know, I'm combining both together because I think they are somehow related.

So the first thing is about like, yeah, these LLMs, they're not reliable. They can hallucinate. Like even rack is not the solution for that. Uh, so, so what's your take on that? And I know like now the large language models are coming like so large that actually you don't need to go back and forth a lot with them.

And the second thing is regarding the security. Hey, like I, I, I am very, uh, let's say, uh, highly regulated. I would not allow, you know, data to go outside. I need to make sure that, you know, I have the privacy and all my data within. So, um. What you would say to people who might come with these objections and Ahmad both of you, like feel free from maybe business perspective and technology perspective to answer.

So Ahmad, I'll start with you. 

Ahmed Mahmoud: Yeah, sure. Uh, again, it's a more technical [00:30:00] part that, uh, CTO should in, I would say, uh, should contribute more. But from business perspective, we understand that those are problems that actually we work to, to fix and resolve and focus on with orchestra. So generative, I had come with a great, I would say potential, but also had created.

A lot of, I would say, noise where everyone can do anything and you cannot really get into what really worked for me as a customer. So we didn't want just to get into Journey ai. We actually took a pause. I. And started to to to study what the problems, really problems that customers were getting into. And thank you for mentioning the, these are two out of four main problems as a business.

What we decided that this is the direction for the company to resolve for businesses to help them adopt Journey bi. So we provide our solutions in complete offline mode where customers can get the benefits of our [00:31:00] AI agents completely under data centers if they are highly regulated or even hybrid, if they have some scenarios that can be outside or some scenarios.

That should stay in from hallucinations. Yeah. They exist and they will always happen. And that's why we had a lot of emphasis in our product to work on accuracy measurement and x-ray accuracy optimization to provide a more robust and I would say confident launch to our clients. But shall we can uh, I would say shed more light on this.

Mehmet: Yeah, sure. 

Ahmed Elshahawy: Yeah. So I think, uh, uh. The concern from, uh, the naysay is, uh, is, is a hundred percent valid. And yeah, that's why we are, that's the main focus of the company that we are tackling these kind of problems. I would say that, uh, the threshold comes always, whether this hallucination rate is acceptable from business point of view, or no.

Comparable to the, uh, error rate that comes from the human, uh, uh, [00:32:00] error, uh, factor as well, or no? The, the, the level of, um, uh, um, uh. Uh, sensitivity of the use case, uh, also is a factor there, and highly sensitive use cases are typically supervised and not fully automated. At least that's the boundary of the technology as, as per, uh, the technology advancement, uh, today.

I'm quite sure this will be more reliable as well in the future, but reaching 95, 90 7% accuracy in many use cases is acceptable. And most, uh, use cases, uh, especially if it was, uh, supervised and still brings business value. And ROI, uh, like tremendous business value and ROI. So we are, the challenge of the hallucination was very predominant in the early days.

And it's still predominant when you go to smaller models for cost efficiency or for, uh, regulatory requirements. Like, uh, [00:33:00] uh, you want to run your own model in your, uh, enterprise, uh, in air gap environment. So, uh, you'll be cost, uh, uh, conscious, and you don't, uh, you cannot host alike. The, the, the huge LLMs, even open source ones.

So hallucination rates becomes, uh, a problem. And actually what we focus on, uh, to, to get this uh, done right is several aspects and, well, no, that means retrieval, augmented generation. This is like. The go-to design pattern for most of the, uh, use cases that we have worked with. And it works really well in most of the scenarios, but if you nail it down in the proper way, and the key element to get the rag retrieval augmented generation working right, is the data representation layer.

So if you get the data presented. Without ambiguity and you have, um, uh, a decent and a smart search system to get the proper data and curate the data in a proper way, uh, [00:34:00] using different technologies, indexing technologies to the, uh, large language model, which is. Fine tuned to a spec, special use cases as well.

Then you close the gap of accuracy. So, uh, search and, uh, uh, data curation and the large language model, uh, performance itself are all, uh, optimized for certain use cases. So, so you get the, the accuracy, uh, mark, uh, to the highest that it can be. And, um, when we, when we started in the early days, just getting the simplest case to work right.

It was like impossible to get this done without automated evaluation for the performance of the model. And that was the early investment that we did before even going into the, uh, advanced use cases because the thing that you cannot measure, you cannot advance it. So we built a very sophisticated evaluation system that.

We can measure the advancement that we do with the accuracy, uh, enhancement that we do. And we built [00:35:00] upon this, uh, to have even offline, uh, uh, small language models performing with very high accuracy. Uh. Because we are able to evaluate it automatically and rapidly and apply different optimization mechanism until you reach the acceptable point.

Uh, so definitely yes, it is, uh, a concern, but it's getting better and better every day. The, whether open source models, a reasoning models, uh, like deep seek, we have all seen the, the advancement, uh, did by Deepsea on. Technology and cost point, uh, of view every day we get this, uh, better. So, uh, so it's going to, to reach a very acceptable mark, I would say.

Mehmet: Yeah, it's getting better and better and a very fast paced also as well. To your point, Ahmed, like, because I think. You know when, when Chad GPT Day, they put, uh, the website live and I remember what that was, November, 2023 and just, I look now like we are just April, 2025. By the [00:36:00] time, you know, we are recording this episode and people, they had high expectations like the technology should work out because, you know, but the.

The amount of time spent to fix all the issues we saw at the beginning. I think it's, it's like a record because, and this is just my 2 cents over there. And yeah, to your point, like things are getting even better and better now. One thing, and I op and I keep like the, the, the discussion like, uh, you know, fast, uh, this time with you both agent tech, ai, so everything agents, everyone is talking about that.

So from your perspective, how do you envision this, you know, with. What you do at the X one? 

Ahmed Mahmoud: Well, um, we, I understand it's a high board. I would just, one thing I would say here is focus on real problems to be solved that really requires autonomous, um, software, whatever we name it, egen, whatever [00:37:00] that is really needed for that.

So this is what we are focused on and this is what I believe the world should focus on, not just getting this brightening. 

Mehmet: Yeah. How about as a CTO? Like how, how do you, you know, see this? Yeah. Uh, hype. Some people they say hype. Some people they say, no, this is the future. So as a technologist, how do you see it?

Ahmed Elshahawy: Yeah, actually, I would say that the agents is definitely, has a, a strong mark in the technology and in the industry. But like, ah said, right. Whether your use case requires an agenting behavior or not, this is very subjective and beyond the buzz, right? It's, uh, many use cases doesn't require this, but it's the same thing that happened with the, the, the spike of degenerative ai in the beginning.

It just opened the door for different use cases. We can have like collaborative reasoning between different agents to provide a level of. Uh, um, uh, reasoning and, uh, and use cases, [00:38:00] um, intelligence that is not possible to do it without the Egen framework. 

Mehmet: Absolutely, and you know what I tell people sometimes, like, like, guys, think about it.

Automation has been with us for long time, so actually you are trying to. Gives this autonomy to the agent. Once it trigger an action, it receive the, you know, the, the result or the feedback to go and do the other action. So you have like, kind of like autonomy plus automation, I would say. Like it can self decide which route to go.

But again, it's powered by automation, which has been with us for a long time. You know, like even before, um, the, the chat GPT start to appear, I was also telling people about like how they can automate tasks and, you know, use these automation tools available, you know, everywhere. Yeah. But AI actually accelerated, I would say the adoption, which is good.

Like, it's good for all of us as businesses. And you know, also, you know, for companies like yours, uh, as well. What is next for the [00:39:00] X one? Like what? Something exciting you can both of you share with us? 

Ahmed Mahmoud: Well, uh, back to where it started. So, uh, maybe at the time when we started working with smaller businesses wasn't, uh, maybe the right time.

Maybe we. Still didn't crack how to work with this ICP. But today with the ability, with, with the ability of technology, with our access to resources, we also started to see how we can empower small businesses to reach that in a very economical way, which we found the formula for. So I would say, yeah, we're back to where we started, but with a better proposition.

Mehmet: Uh, any plans for expansions Ahmed? Like, uh, 

Ahmed Mahmoud: yeah. Uh, so we're expanding in, um. From Geographics, we started to examine expansion in, uh, in US and in uk. Uh, we already have some, uh, resellers and system integrators that are [00:40:00] helping it out on, uh, that it's, it's a very big noisy market. That's why we're very niche on the industries we are ex examining.

The second part is expanding our customer base from just enterprises into mid-market as well. So we're getting into the mid-market and we identified several problems that we can resolve with and we started to add and expand into that segment as well. And this is something we're launching very soon, actually, the first time we talk about it.

This is something we're launching, um, in next week, I guess, shall we? Right next week. Yeah, 

Mehmet: talk about, 

Ahmed Mahmoud: okay. We had evidence in this show. 

Ahmed Elshahawy: Alright. Not before. 

Mehmet: Yeah. So, uh, Ahad, like, um, yeah. So we have, now it's recorded. Just joking. Uh. From, from technology aspects Ahmed, like as a CTO, of course we talked a lot.

Ai. Ai, but like what [00:41:00] are other, I would say, emerging technologies that also you're excited about that might, you know, first be beneficial for the company, for DXwand and in general, like something which you are excited about, what you're seeing, uh, in the marketplace. 

Ahmed Elshahawy: Yeah. I would say that the, the AI technology is, uh.

We have a long wave adoption just to get the ROI of the existing technologies, even the early wave of the technology. Before, uh, actually envisioning new stuff. And this, I think one of the pitfalls of many startups will keep looking for just the next shining objects without realizing the full benefits of what you have in hand.

So we are, we decided to be ultra focused on the egen tech capabilities with document, uh, intelligence and knowledge mining. But the next focus area that we, we are, like Ahmed said, is to get this to a self-serve. Uh, mechanism that even [00:42:00] mid-market and medium-sized companies, business users can have the advancement of these technologies without getting into much of the technical details.

Mehmet: Great. Now the final thing I want to ask, and feel free both of you, you can answer. So being a company and Ed, you started this 2017. Uh. So, so you start the Yeah, I did. Yeah. Yeah. And you know, I know, matter of fact, because um, I witnessed, I tried myself, starting, especially in our region is not something easy, right?

So, but what I want you to share, like maybe something that can, I would say inspire fellow people who want, or they are dreaming of starting. And by the way. Some people, they get me wrong. I'm not against consumer businesses, but I would love to see more companies like us, uh, you know, being started and, you know, [00:43:00] scaling.

So what can you share about this journey that, uh, maybe I. One of our audience, or maybe someone who might discover us, you know, speaking about this in the MENA region, in the emerging markets, on the, you know, more wide, uh, uh, coverage that might benefit. So what you can share, Rahman. 

Ahmed Mahmoud: Okay. Inspiring people is not, uh, easy.

So, so yeah. Okay. In 2017, actually, 2018, when I started to resign, I wasn't a sales. With Microsoft, I had 200% attainment on my sales. It wasn't a sales organization. Uh, I had no problem. And, but I, the only thing is I just, I just felt it's not right. This is not what I want to do and what I can tell everyone that he might be inspired, I went through hell.

[00:44:00] Really? And that's really 

inspiring. 

Ahmed Mahmoud: Yeah, very inspiring. But guess what? If I go back and I never go back to anything, but if I would go back knowing what I went through, I will still do that because if you really have this call that you want to do something impactful and you know what you want to do, then this is where you should be and.

And I'm, again, I, I'm a father of five children. It's, it, it wasn't like someone just got graduated and I don't have a lot of responsibilities and yeah, I know much breaks. No, I had a lot of risks, but I'm telling you, if I go back seven years, I will do it again knowing what the hell I'm going into. And today I expect a lot of more health.

But the good thing is it's not about just getting into an exit, which everyone has entrepreneur dreams about, but really, and it's not a cliche statement, but. The journey that I went through is a great learning that I'm really proud of [00:45:00] and not proud of myself, proud of everything I went through, and that's what you would really get and you'll never get anywhere else.

Mehmet: That's inspiring. And you should be proud, Ahmed. But I want to hear from Ahmed leading tech teams in in MENA region, you know, and, and because it's, it's a team. It's a team game, right? So, um. And you need also to, to put this into the people you work with, you lead. So from your perspective, Ahmed, like reaching, you know, this stage, um, and encouraging more people in tech to take the same, I would say, steps of going out and start building things that can change people lives.

So what you can share, 

Ahmed Elshahawy: yeah, I would say that I was lucky to. To get in touch and work with then, and a wonderful team that they were really hungry for doing great stuff [00:46:00] and not of the type of, uh, like, um, standard employees, nine to five employees, which is not bad. That's fine. But most of the team in DXwand are really different part kind of, of, of people.

They just, uh. Every time, um, they, they don't fail to inspire me actually how to get things done beyond my expectations. And I would say that believing in your team in a very pragmatic way, not like, um, just, um, uh, pure belief, but with some sort of evidence that the, the most effective team members can show.

Is the most inspiring thing that helps you to build new stuff and makes your work fulfilling [00:47:00] and makes your life easier to get things done in, uh, a different way. I will say that keeping this spirit of innovation, always keeping while being a hundred percent business driven, but keep keeping the level of innovation.

And novelty in your team to drive, um, uh, like passion and, uh, contribution, uh, in, in the product and technology, uh, uh, roadmap is always going to get people to ramp up their skills and reach a level that you have never dreamed of initially. And I, I, I believe we have actually over the, the last few years built a amazing, wonderful.

Skillset and the team that they are really, really inspired and passionate of building stuff in the X one. 

Mehmet: Uh, you know, both of you, you know, you uh, you shared a lot of, uh, uh, really, [00:48:00] really, it's inspiring and it's not like, uh, um, you know, it's not, I'm making fun, of course, of it. It's really inspiring and what I liked, you know, Ahmed, when you mentioned like about going back and starting again, and I think this is.

One of the traits of entrepreneurs like to accept that there will be, I would say, challenges. There will be some obstacles down the road and, you know, keep trying and never failing, which I always tell people maybe one of the things that we need, we need to encourage people to fail more. We need to encourage people to face obstacles.

We need to, not on purpose of course, but we need to let people hit walls sometimes and, you know, test this because it's not all the. And I loved when you said, you know, like, yeah, like I had this 200% attainment and you know, love, you know, life was, uh, you know, all rows and everything was fine, but it's the call inside you to go and build something.

And to do this, you need like the courage and you need also the [00:49:00] ecosystem and. I think you contributed in showing, you know, let's say the younger generation that this is, is doable. I'm very positive that, you know, we are gonna make it like even more easier for them over time because it's still very hard.

Let's admit it. And again, from technology perspective, Ahmed. Same. You know what, I have the same thoughts that, you know, with this passion about building something that, yeah. Like look, I, I built this and people are using it, right? And it's like a technology, piece of technology that we all together put and then see how it's affecting people's lives.

So really this is inspiring for me. Before we do the final, final closing, where guys, can people get in touch with you? 

Ahmed Mahmoud: Um, uh, I'm very much active in LinkedIn, which is not the most LinkedIn for the new generation, but yeah, this is where I, I, I exist more. I spend more time. 

Mehmet: Yeah. And Ed, you also, I believe on LinkedIn.

I [00:50:00] accept 

Ahmed Mahmoud: all invitations, everyone.

Ahmed Elshahawy: Yeah. LinkedIn is the main channel and. Definitely. Yeah. Feel free to, to connect, be 

Mehmet: yeah. Happy. Yeah, sure. And if someone interested, uh, you can go to the website, the x one.com. I'm gonna put that in the show notes. So both Ahmeds, thank you very, very much, you know, for being here with me today. I really enjoyed the conversation.

I love to hear, you know, these inspiring stories. I love to, uh, see. The technology, and again, guys in the B2C space, don't get me wrong, I, I love to have you also on the podcast, but I'm a B2B guy. I would say this is why, you know, I get excited when I talk to people who are working in B2B Tech and building things from scratch.

But yeah, don't get me wrong, go build things in B2C space also as well. So, and uh, you know, the conversation was really inspiring, I would say, and, uh, beneficial in my opinion for a lot of people in the region and abroad, I would say. And good luck, you know, in the [00:51:00] expansions, whether on the vertical perspective or geography perspective, I would say.

And as I said, the. Website, the links to both Ahmed's LinkedIn profile will be in the show notes. And this is how usually I end my podcast episode. This is for the audience. If you just discovered this podcast, thank you for passing by. I hope you enjoyed it. If you did, so, please don't forget to subscribe and share it with your friends and colleagues.

And if you're one of the people who keeps coming again and again, thank you for doing so because of all your supports, because of all what you do and you know, encouraging me to keep this podcast. So the podcast, and this is important. For me, it's about, you know, keeping it alive. So it's now trending in multiple countries, top 200 charts simultaneously this year, which is something didn't happen the past two years since I started.

So thank you very much. And also, we've been selected as the. One of the top 40 podcasts, most listened business podcasts in Dubai today. And for the sake of transparency, we are recording 15th of April. I just received [00:52:00] that, uh, the podcast was chosen as a one of the top a hundred, uh, globally podcast talking about future technology.

So thank you again for, you know, mentioning us and telling people about my podcast. And as I say, always stay tuned for in your episode very soon. Thank you. 

Thank you.