Hermes Frangoudis (00:07)
Welcome to the combo AI world podcast where we interview the teams and builders forging ahead the future of conversational AI. I'm really excited today. have Ram from Colaberry. Thanks for joining me today, Ram.

Ram Katamaraja (00:21)
Hermes and everyone pleasure to be here.

Hermes Frangoudis (00:24)
So let's get right into it. I always like to start off with a little bit of an origin story. So what led you to create Cola Berry and why?

Ram Katamaraja (00:32)
I would say Colaberry is more of an accidental venture. ⁓ So I was working in the tech space, doing all kinds of cool stuff in the enterprise space.

business process, re-engineering, enterprise service, buses, et cetera. However, around 2010 timeframe, I saw an opportunity to train military veterans and help them transition into good paying jobs. So I took that as a, opportunity as a responsible citizen and started training. that became like so successful and popular. was forced to create a company and yeah, so that's the origin story. ⁓ From the beginning we have been doing.

work, amazing work with like a lot of Fortune 100 companies in agriculture, utility, IO tech, and some healthcare finance industry space. And at the same time, as a way for us to contribute back, we have been training military veterans and then historically marginalized people in the tech skills and helping them transition into the

good paying job. So it's been like a wonderful journey.

Hermes Frangoudis (01:47)
That sounds amazing. Helping people kind of better themselves and get ahead in life is huge and very important, I think, in this world, right? But that's not the only cool things you've done. You won an award, right, for AI for the Betterment of Humanity. Can you tell a little bit about that project and the competition?

Ram Katamaraja (02:06)
Absolutely. So, you know, we're always obsessed with the business with a purpose. ⁓ So, ⁓ and that purpose, know, entrepreneurship is hard, like starting a business and growing it is hard and you need a purpose to keep it going. And for us, it's more than money. Of course, like we do a lot of work with large enterprises building like amazing, cool, futuristic.

data science, data engineering, AI infrastructures. But at the same time, as part of our give back, are working with individuals who need like pathways to transition to the tech jobs. And that job was not easy. And we as technologists like thought about how can we use technology, data science, machine learning algorithms to help people skill better and argument them to be prepared for work on day zero. Right.

So as part of that, we developed technologies using what we now call conversational AI, where we work with text based interactions and providing them real time analytics about their learning and also capturing their voices and videos and giving them like real time feedback about their communication skills so that it helps them get better. You know, one example, if I have to put it like whenever we are preparing for interviews, they advise us to go and speak in a mirror.

Right. But a mirror doesn't give us feedback. However, we thought like, okay, what if we can capture the video and give them feedback? So that was like a kind of like change the game. And some of these innovations like literally helped us to move like baristas to data engineers, customer support people to like data analysts and basically like Uber drivers to reporting engineers. So it was like wild, pretty interesting, interesting changes because

We were not only using ⁓ skills to succeed hard skills, but we were providing them ⁓ the communication and collaboration skills using technology and help them like get it like in a much faster way by giving them objective feedback. technology, when we present like back in 2018, you know, it's not a long time ago, but also a long time ago, right?

The world was talking about the future of work and how AI is coming and how it's going to disrupt the entire workforce. So we presented our solution of how we can quickly re-skill and up-skill, how we can use technology to better humans. So that's when we received these really two cool awards. One is from MIT itself called MIT Solve. We're being recognized as a solver in the.

work of the future arena. also, we also received like, which is like a big surprise for me, AI for betterment of humanity prize. So that was like, wow, this is so cool. Yeah. So it's been like an interesting ride doing some cool things using tech.

Hermes Frangoudis (05:06)
That sounds like such a wild ride, but you touched upon some like really interesting pieces there that I'd love to dive into. And some of it is around the, that idea that, you know, the workforce is going to be obsolete thanks to AI and how do you, you upscale and bring people, like you said, from Uber driver or barista to data engineer, right? Connect people to those jobs that will help them become better in life. And in that sense, what do you see as the most critical skill gap in that?

piece of the workforce now.

Ram Katamaraja (05:37)
That's a very tough question to answer. So we thought we had the answers, but when AI became reality, it's a completely different world, right? And we can see the disruption that is happening. However, the way I am observing is the AI is moving like really fast, but at the same time, organizations are moving slow.

Right. So the adoption of AI has been slow and deliberate. And I think probably it's because how fast it is moving. So it has not matured. So the way ⁓ I perceive it is like, like any other technology that has evolved in the past, either you take going back 20, 30 years internet to mobile adoption, to a big data adoption, all these things went through a maturity model.

Right. So there are going to be AI first companies, just like we had internet first companies. are going to have AI first companies that are going to disrupt, but a majority of the world is run by brick and mortar like industries, agriculture, manufacturing, utilities. Right. So these industries are going to exist. And as these, these industries have to move deliberately in a very thoughtful fashion. And as part of that, the

Reskilling and upskilling is like a very important concept that is on everybody's mind. Right. So the way I am pursue, it's going to be like multiface. So one is there is a productivity gain tools and then there is an enterprise tools. Right. So what that means is if you look at the technology as such that we experienced today. So we have like the word processors and Excel and PowerPoints of the world, which were the productivity gain tools.

right? So the chat GPTs of the world are such productivity gain tools. So basically they allow to do more with less, right? So that's one side. what I'm saying is like everybody must get trained in this fundamental productivity increasing tools, meaning everybody needs to become what you call

Hermes Frangoudis (07:45)
⁓ a prompt engineer. ⁓

Ram Katamaraja (07:47)
So, yeah. So, and then you have this whole idea of enterprises moving and adapting AI, right? For enterprises, just not about costs, basically reduction of the cost and increasing the bottom line, but also they need to come up with new products, new architectures, new solutions, right? So this is where ⁓ I think the agentic AI that is evolving is having the way for it. So that's the future. And in that future.

What skills are required are still need to be seen. And in general, the humans are going to do more high value work and humans are going to be more, more orchestrators of the business, monitoring things, making the, and making decisions based on the, what, ⁓ the agents are like doing. So what that means is we have to learn to work with the bots. We have to learn to work with agents.

And this whole skill development is going to be like, we don't know what, how it's going to evolve, but it's something that we have to look forward to. Like one exciting thing that I am working on is this particular fall semester, I'm working with a group of MIT Sloan students who are trying to figure out, do research on, okay, what does this future of learning and skill development looks like? And conceptually, I feel that there is a room for what I call like learning ops.

Right? So have DevOps, have like FinOps, all these things, right? So these, these ops have changed how we develop software or how we manage revenue, how we manage finances, et cetera. Similarly in the learning space on the skill developments case, the skills are disappearing faster than ever. Right? So previously you learn a skill, you could probably work five to 10 years. Now it's probably, it's going to, it's going to basically disappear like really fast.

But at the same time, the need for re-skilling is also growing, right? It's huge. So like now, how do you do that with the same amount of resources? Right? So I think that's where I feel that the whole concept of learning ops could be an interesting thing to experiment with. So yeah, so that's one exciting thing that I'm working on this fall. So we'll unveil what we find out about your question that you asked. ⁓

At high level, I don't know the answer, but we'll figure it out.

Hermes Frangoudis (10:16)
No, that's exciting to hear that you're actually researching it and thinking through and trying to find what those answers are. What are those skill gaps? Because like you said, they just don't exist. I'm sure people are curious because as AI eats old jobs, it creates opportunity for new jobs. Right. And that's how technology has always been as one piece of technology makes this really labor intensive task infinitely easier and easier to scale. You now need to learn how to use that piece of technology to do everything.

at a bigger scale. So everything is that corporate buzzword of upskilling at scale. Like you said, the world runs on a lot more brick and mortar and traditional industry than just digital industry, right? Like the digital world is changing constantly and has been since its inception. Brick and mortar, those sort of more traditional industries, enterprise where you're investing on a 10 year timeline.

20 year timeline, right? That's not one year like most AI companies are working on.

Ram Katamaraja (11:19)
That's right. Yeah. So for me, what, what I observing is for the first time, these brick and mortar companies are asking, can AI help them do better and faster? So if you just look at AI itself right now, there's a framework called WINS framework, W-I-N-S that was, up in HBR, a Harvard Business Review. It's like words, images, numbers, and sounds. Wherever these are, there is immediate disruption. Right?

We see that happening, right? So either it's marketing or whether it's legal, whether it's like, I don't know, finance, like wherever these things are, there is immediate disruption that the chat GPTs are like the LLMs of the world has created, right? And so, ⁓ so that's, that's right now happening, but the more exciting pieces as like I said, we work with this boots on the ground type of industries, right? So we can like, let's say you want to set up like a new industrial plan.

So you need to do engineering design, right? Previously you have like people like doing the whole engineering designs. And now the questions that they're asking is, Hey, previously we produced two options. Can we produce 10 options using AI? Yeah, you could do that. Right. And similarly, let's, let's just take your home, right? I mean, if you are in New England, like you are in United States, you have like so many windows and doors in your house and we have to refresh the house like every 20, 30 years, right?

So, and then somebody who comes to your home and takes the measurements of your house and like sends it out to the backend manufacturer, they do a custom manufacturing, right? Somebody has to manually sit and like validate all these things. And just imagine how many errors the humans who are taking the measures could do, right? So yeah, I could fix those engineering designs and make these things better and efficient, right? And so from...

creating like new designs to just changing this entire workflows to creating like more and more options in terms of creativity. also very interesting thing, we work in this subsurface oil and gas space. We could use AI to create new eco-friendly materials, like new components using this AI. So then we can create more sustainable drilling options, more sustainable, I don't know.

I mean, close, I want to get...

Hermes Frangoudis (13:45)
Well, what is the animal materials?

Ram Katamaraja (13:48)
Yeah. So, so, so there is the option of this, our ability to have more sustainable and eco-friendly controlled creation of substances is going to be like a lot more. So that is like the most exciting part of the whole AI in a larger way. Right. Whereas if you take a step back, when we come to our day-to-day interactions, I basically, I would come back to this conversational AI piece because I feel that that is the new UX, right? Voice is the new UX.

It's not the HTML anymore, right? It's like the bot, like the text is the new UX. Like you just type it in and you have the voice is the new UX. Video is like the new UX. So our interactions with this technology itself is going to get disrupted. I'm like really excited to see that Agora is making strides in that space. So you have like the whole front end and then the back end like it's pretty exciting where the AI is taking us into the future.

Hermes Frangoudis (14:46)
Yeah, no, it's really exciting where we're seeing the trends and stuff like that. And you brought up a good point. Agora is upskilling our network essentially. Like we focus traditionally on human to human interaction, but now there's a lot of opportunity between human and computer. So as you said, voice and video is the new interface and power the best voice and video infrastructure in the world. I also want to hear about what you're building with voice and video because

I know you're building some really cool stuff on bringing this all together. You're using Agora, but can you maybe walk us through some of the awesome pieces of software and systems you're putting together?

Ram Katamaraja (15:23)
So I like Agora for its ability, its enterprise-grade ability to stream billions of minutes. So it's an enterprise-grade infrastructure that has been tried and tested. And then when it's related to the conversational AI, there are many tools out there. Some of the notable ones are related. Eleven Labs is a pioneer. And OpenAI, obviously, you have it.

Microsoft Azure voice libraries are there and there are many other conversational AI platforms that are coming up, which is like super exciting in this space. However, what I like about Agora is it's a tried and tested ⁓ enterprise-grade framework on top of it. And it is being built in a way that we can integrate with various systems. I mean, if you don't mind, I can quickly share something.

Hermes Frangoudis (16:12)
It's built like Lego blocks, gives you all the control while taking away all the headaches, right?

Ram Katamaraja (16:18)
Yes. Do you see my screen? Okay. All right. So what you're seeing here is like traditional, ⁓ agora framework, right? So you have the, ⁓ agora platform have, sip web RTC conversations, and it has ability to integrate enterprise systems in CRM systems, call records, like everything. ⁓ and it can stream, it could do

all kinds of like really cool stuff related to the text, voice and video streaming, right? On top of it, the conversational AI is being built and this conversational AI is more like you said, it's a Lego block which allows to integrate like with any other systems. So for example, this is like a stack that I'm just showing. So for example, at Colaberry, we have the AI orchestration platform that has a way you can fine tune LLMs, you can fine tune

text to speech, translation services, MCP connectors, and all kinds of orchestrations and everything. So these architectures could be used to enable basically ⁓ Convoy AI. So let me just show you this particular flow diagram a little bit. So here, as you can see, so this is an app where ⁓ you have a real-time audio and real-time audio streaming can happen. And here in the backend,

You can integrate with agentic flows here. So for example, here you can integrate with 11 labs, text to speech, right? You can integrate with open AI whisper modules, or you can integrate with Microsoft modules, right? And in the backend for an enterprise, which is like, you have like a lot of custom intelligence inside the company, right? So for that, you may need like a large LLM models and you have custom databases. can build like rag, rag tools and everything, right?

So the Agora platform, by being the front end and having the ability to plug in any of these backend systems, allows to build an enterprise-grade text-to-speech, speech-to-text conversational AI platform. also, so any questions here?

Hermes Frangoudis (18:27)
In here, zoom out maybe a little bit, the person calls in. So this isn't like your traditional app. This is a dial-in number. So they're interacting with a traditional interface in terms of like the mode of connection, right?

Ram Katamaraja (18:42)
Yeah. So the beauty of Agora platform is it can work with both phone as well as well, we OIP, right? Regular telephone and we OIP. So you have a sweep switch, right? So, and with the switch switch, can basically stream the voice across the regular telephone line. But if like in this particular one, you could just have like a client, Android client or a Voip Client. So it allows like both of those, right?

So, and this whole Agora infrastructure is already there and it works at the enterprise grade. So all we are doing is making it AI enabled because previously these are not necessarily smart conversations. Now with integrating with AI platforms, they become smarter conversations.

Hermes Frangoudis (19:31)
Very interesting. You're using Agora as kind of like the underlying level of the streaming orchestration and then tuning it into your pieces that you've been building over the years that your enterprise's customers have been building over these years. The knowledge database, the knowledge base, their MCP servers, all these things that kind of can exist without voice, but now being able to bring it to this new interface sort of.

Ram Katamaraja (19:55)
That's correct. all these interfaces already exist in the internet 1.0 and 2.0 world. Like now, how do you upgrade to the AI world where it's smarter? So that's where the enterprise may already be using a good platform for their call centers ⁓ and all kinds of streaming needs that they have. And the calls may be already getting recorded, et cetera. Now we can.

Uh, we, we, uh, we can integrate them with the AI infrastructure and make the whole, uh, business smarter. And maybe they can even innovate like new products and services on top of it. So for example, they could have, uh, uh, they could have automated, I don't know, like if you take a healthcare industry, confirming appointments is a big thing. Following up with the people is a big thing, like prescription medications. All these are like usually instead of like humans following up or even robotic.

calls being made, like these can be like lot more humanistic empathetic calls because the AI interacts. It's just not delivering information. Yeah. It has the ability to interact so you can upgrade and make your, make the processes like lot more.

Hermes Frangoudis (21:06)
away from that traditional phone tree. Hi, we're calling to remind you about your insert, whatever press one. Right. So this is like having a natural conversation, which for most people is, all they really want to be able to do when they call in. Like, I don't think they really care that it's a person or not. It's I don't want to have to like change the way I interact and speak to accommodate their phone tree or their technology.

that just makes my life more difficult, right? So this is an ability to really bring that human interaction back into things.

Ram Katamaraja (21:43)
Yes, one, can bring the human interaction back, but also the technology is going to a place where you can record your voice and your nuances, right? Your voice nuances, highs and lows, and the AI can use that exact voice and nuances and show empathy while talking to the customers. And not only that, it can go beyond a transactional conversation, right? So for example, I don't know if I could play it, but...

happy to kind of like share in one of the conversations, there is a con, appointment confirmation that was happening for a, ⁓ a cardiovascular appointment. And the patient, it's, it went beyond appointment saying that, okay, hey, is there anything that I can help you with? And the patient suddenly goes, yeah, I'm having this issue. Like, can I talk about it? And then the system goes, okay. Let me relay this information to the doctor. I'm not authorized to kind of like.

help you with this, but let me inform you. So it can make more notes, right? Beyond like what a just bot would do because it's intelligent enough and it can be empathetic enough to respond to the customer needs. it allows us to build like a smarter systems that not only engage with our customers, but also are empathetic to our customer needs. That's the game changer.

Hermes Frangoudis (23:06)
huge.

The ability to be a more dynamic system, right? Even with traditional AI interactions, if you think about like text interaction with an AI, you're trying to one shot everything, right? Like, let me put everything I can into this perfect first message. But that's not how we interact. That's not how we think. similar with technology, if you're trying to always force people into these boxes to interact, it doesn't allow for

the same sort of dynamicness that you were just talking about. And that's important in so many professions like healthcare, where by putting the technology in, we remove some of that personal piece that can actually help the customer, the patient, so to speak. And being able to put that back in is huge. not just taking away, but putting back.

Ram Katamaraja (23:53)
Yeah, yeah, no, absolutely. So also another conversation where we enjoy like watching is like I said, we have like students that we train, right? So a lot of times people want questions to be answered. So we built chatbots while ago where they ask a question and we would go look at the repository and give them, okay, this is the feedback and like go through it, right? So now once we integrated with RAG, suddenly

We see that these interactions become like lot more smarter and the students, like we can watch the students feel as if they're talking to their actual mentor, a human mentor, right? So rather than a system itself, so, they would like, thank you so much for this previously send a document at a link. Now we are able to give them like exact answers that they are looking for. Right? So it's a, so that's the text conversation. Right? So it's.

It's wild. also we could see these systems getting like smarter and smarter. One as the LLMs get smarter and also as more and more conversations happen, these LLMs that we use like keeps getting fine tuned. And not only that, we always had this, what we call the thing, the summer sleep or something like that. The people like stop engaging, right? One of the biggest problems with the

distant education or like the online education or learning is people get disengaged, right? Once you disengage like a minute, that minute becomes an hour, that minute becomes like weeks, right? And then they're gone, right? So you could look at these patterns, right? As people, could, you could literally look at the pattern of engagement in the first week. And then like, can devise a plan around them, how to follow up and also do active follow-ups sending them.

active text and active WhatsApp messages and like interact with them as if we can assign like an AI mentor, right? So our AI tutor and they just love it because now they know that there's someone who they can talk, they can interact with, who not only clears the questions that they have, but also following up with them and ⁓ helping them out to kind of like stay ahead. it's a lot of, lot of experimentation that can be done. And not only that.

We can do this with not spending more, right? We had to hire like 5,200 mentors to kind of like support them, right? So we still want to hire them because human interaction, you cannot remove that, but almost 90 % of the work that these mentors are supposed to do, keeping these mentors do their work is also challenging, right? So they have a great heart, they wanna give it, but at the same time, it can be taxing, right? So now AI could do like a lot of their work.

and make them look smarter so they can focus on the human interactions and then the systems take care of all these things. A lot of exciting things in the conversational AI. I mean, at least, again, my prediction is that in the future, will have the UI, the entire UI UX is going to be driven by conversational. ⁓

Hermes Frangoudis (27:10)
I'm right there with you. think it's really just starting. We're just kind of like at the tip of the iceberg and the shift is already very, very visible in that sense. I do want to dig into... sorry. I didn't mean to...

Ram Katamaraja (27:21)
Yeah. So one, one area, like as of now we see, ⁓ I think it's also the nature of the job, job market at this time, which is like tough as we go through this transition. Right. So the general, there are AI screeners, AI recruiters that are out there. Everybody's using that. So the general expectation that they need to talk to a human is going to like becoming less and less. And then the acceptance of.

the AI agents out there, like working with them would increase and humans would like naturally like start using them.

Hermes Frangoudis (27:56)
No, no, I think you bring in a very interesting take there, not only on just like the recruitment part, but also the mentorship piece of it. Being a mentor to someone is very taxing, right? On your time, on your mental capacity to not only do your tasks and what you're supposed to do, but also supporting someone else and supporting their growth into doing what you do, right?

And part of that for the mentee is just being able to check in and be like, I'm on the right path or not. But that, that level of disruption also is sometimes very difficult for someone offering mentorship. So the ability for them to kind of clone pieces of themselves into this and use those agents to better facilitate and support the growth of many more mentees, I guess you'd call them. Is that kind of like what you guys are working on to facilitate?

Ram Katamaraja (28:48)
Yeah, yeah. No, so I call it like the mentor, mentoring fatigue, right? So you have, you want to give back, you have a great heart, but at the same time, you end up having like certain expectations on the other side. So, and the other side may or may not respond in the similar way because everybody responds in a little different way, right? Before, so the mentor fatigue sets in, right? Which I could see when you're talking about it, right? You're like, I want to mentor, but I don't know. It's so difficult to do it.

Right. So yeah, we are automating like lot of these things so that we can, so on the flip side, the mentees need this, the time from the mentors and the check-ins and the guidance, et cetera, et cetera. Right. So we see that I can look at the data and like, Hey, you okay? I haven't seen you in the last couple of days and what's happening, everything. Okay. Right. So, ⁓ like in your language, right. If you want that agent to represent you.

could send it. So that's the beauty of

Hermes Frangoudis (29:49)
You also brought up an interesting piece of the hiring process, the recruitment process is like, I don't even know how to describe that labyrinth right now. It's a maze because there's a mix of AI agents doing screening, humans doing screening, humans sending their own agents for applying, right? So it's like agents interviewing agents, which is a whole different conversation, I think. But the ability for like a company to...

deploy an agent that can kind of do this initial screening and figure out all of these baseline elements. It doesn't really take the human aspect out of it because they still have to go through a human, right? This is recruitment of people to people type stuff. So how do you see that kind of like furthering the current level of disruption? you see it really changing anytime soon?

Ram Katamaraja (30:39)
It's a wild west out there. ⁓ So, on one hand, people think that this is the easiest place to deploy AI, right? And on the other side, it is like the, one of the most difficult spaces where you can have like a bot or have a conversation because regardless of how smart it is, the bot and how empathetic it could be, it can only go on until a certain loops of the recruitment, right? So, so it could be.

Hermes Frangoudis (30:42)
Totally.

Ram Katamaraja (31:09)
the initial screens or it could be like the onboarding, where, how do I put it? So wherever there is more transactionality to the conversation, right? That's where the AI could be like more effective, right? Because both sides could understand that it's a way to stuff each other's time, right? However, when you're talking about like the interviewing and you're talking about, I don't know, salary negotiations and all these things, it definitely,

requires involvement of humans because if not, it's, yeah, I could promise something. You could hack the bot to promise you something and you are.

Hermes Frangoudis (31:46)
Totally

works in a completely opposite way. Hey, I renegotiated your package. You're only down 60 % of your previous balance. Yeah. It's like, what?

Ram Katamaraja (32:00)
So yeah, so it's, I think it's still a wild west out there, right? We're all trying it out. Like what the thing that I'm trying to bring that up is there is a general sense of us being open to engaging the bots, right? So that's the, again, like you said, bots interviewing bots and we having our own bots, they having their bots, but the beauty of this is like, okay, we are using this tech.

So that's the most interesting part rather than the chaos that it is creating at this time and how this evolves into the future. At MIT, there is this MIT Media Lab. There is a professor ⁓ who is ⁓ basically popularizing the decentralized AI agents concept. So on one hand, we are talking about this AGI superintelligence and that would basically control us.

And then here they talk about the concept, well, everyone of us have these, I don't know, a hundred agents representing like various things and they are going to negotiate on our behalf, right? So, and they're going to make sure like our vested interest is being taken care of and like approve, deny things that doesn't align with us. ⁓ that seems to be like an interesting path to that would naturally evolve into like.

we don't want to be controlled by overlaps, right? So we want to put our agents out there. So yeah, so it's a decentralized distributed AI, just like everybody is carrying a mobile phone on that phone, everybody is probably carrying their agents, or they exist in the ether, like representing us. So yeah, that's the future.

Hermes Frangoudis (33:43)
That's a very interesting concept. Like, Hey, I'm on the job hunt. And you brought up a good point, like the person fit and the job fit. It's not always like the description that makes like that final match. Like it's that first conversation asking certain questions and that takes time. And so for both ends of it, that's probably like one of the most time consuming pieces is the lowest level because you have to do a breadth of it.

Right. Before you can kind of cut that out and go to the next level, go to the next level in advance. So if I have my agent, you have your agent and your agent talks to my agent and says, Hey, I hear you're hiring. What's the role about? Oh, this is the role. Okay. Well, I like to do this or I'm specializing in this. And they realize, Oh yeah, actually you'd probably be a good fit or no, maybe you'd be a better fit for like this other role we have internally. And it, and it changes a lot of that dynamic.

in a way that saves everyone time and kind of gets you to where you need to be. We'll see. Yeah. I'll always just have a mystic view on things.

Ram Katamaraja (34:42)
We'll see.

So, but one thing if we have to look from what is happening, right on the ground. So the idea that there are going to be humans and agents like working together is a reality, right? And in an enterprise hierarchy, right now the hierarchy is you have all humans in there. So now in the future, there are going to be AI agents that are to be part of it. So they're going to be part of the team with certain skills.

Right. For example, if you just take like a very simple customer service call, basically it's a theme of people and there are like AI screeners, right? Or you have to make a follow-ups and there are AI follow-ups. So these agents are going to be part of the organizational infrastructure. Right. So this is going to, hope this is going to increase the customer experience. And also this is going to reduce the cost with which the businesses can.

do the business. However, the businesses cannot be stagnant, right? They have to grow. They have to innovate new products, new services, right? So this is where more and more new opportunities that we don't see now will emerge. like a decade ago, they don't know, like I didn't know what an influencer job would be, right? Now they are like the stars, right?

So they are better than Hollywood are these football stars. So that's kind of like, so we don't know what the new jobs are going to be. And at the same time, even in the tech world, it's pretty exciting to see like some of the smartest individuals being treated like our NFL stars, getting like multi-million dollar checks. Right? So it was like, hopefully. Yeah.

So hopefully that would change like more people to pursue STEM education. Right. So, and, and, ⁓ mean, people, people are fascinated about sports because of this shiny checks, these stars are getting, and like these new big shiny checks that people are getting in the AI might, ⁓ might create like a complete new generation of like smart thinkers and the people who pursue STEM and

engineering and science and space, all kinds of things. I'm super excited about that. That just one piece is like super exciting for me. I'm like, wow, we like, know, growing up in India, like for us, it's always an academically accomplished person or somebody accomplished in a job is always looked like a superstar. Right. So when, and when I came to United States, it's like, what? We are not superstars.

Hopefully, hopefully that's gonna change with this whole big checks that Zuckerberg and others are doling out to this really, really smart people.

Hermes Frangoudis (37:39)
Yeah, it's wild to see the, like you said, the superstars of the tech world are now the data scientists, not necessarily the influencers or these people that have built quote unquote, followings. It's the people that can actually move the needle for these major businesses using AI, which is more of a call out to the younger generations is like figure out how to do this, figure out how to bring this together. The job future is changing. The landscape is always changing and maturing.

And I feel in the future, education is going to catch up. Some point has to, that's just like maybe Darwinism, right? Like he who fails to catch up will just be outperformed by someone else that can. And so it's cool to see schools like MIT are putting a lot into the research of this and figuring out like, what are these directions? Cause that's really going to guide a lot of the future understanding of like where these things are moving towards, right?

Ram Katamaraja (38:35)
Yeah, yeah, absolutely. Absolutely. I have young kids. this question is something that we parents like keep talking about all the time. Sorry. I mean, there's a tech talk, but you know, this is

Hermes Frangoudis (38:47)
I have two kids too, so it's definitely something to think about.

Ram Katamaraja (38:50)
Right. So like, what do you teach kids? what do you want kids to study? mean, it's, ⁓ yeah, I mean, they are going to go up in a world that's powered by AI, but going to the fundamentals of how, how we make progress. It's system, math, physics, right? You need like these fundamentals and you have to use the AI to create like a new materials of the world, new, better features, more.

sustainable systems, right? it's everything AI is the superpower that they can allow the future generations to do hopefully much better and sustainable innovations that are good for the world and good for the planet.

Hermes Frangoudis (39:32)
Yeah, no, it's true. You hope that this will move everything more towards a positive space. The more I interact with AI and these AI systems, like the more I realize, like you say, the fundamentals are still the most important piece of it, right? All of these function off of, ⁓ if you know certain fundamental understanding of things, you can stand so much taller with AI and do so much more with it.

But if you don't understand those things, you're going to ask kind of like these low level questions and get by and it's democratized that piece. But you're not really going to like progress forward in that sense because you don't have the strong enough foundation to interact at that different level. You're interacting at the base instead of deeper into it.

Ram Katamaraja (40:16)
Yeah, yeah, no, absolutely. So ⁓ I was having a conversation with my kids and the thought process is like, they are like with the idea, should we use AI or should we not use AI? And I was like, what do you think? Well, it might make us like not smart, right? I'm using diplomatic language. So my question is like, why, why do you think it would make you not smart? And they're like, it's giving us all the answers.

So then it was like a moment of, I don't know, like reflection and also thought for me. And then I was like, okay, hey, you work with Lego blocks, right? So they're like, yeah. So you, can you build with Lego blocks if you are not intentional and thoughtful, right? You build something which doesn't make any sense, but if you are intentional and thoughtful, you could build like amazing things, right? See, and why, how are you able to do that? It's just, these are the blocks that are available to you. Put them together. You can.

build stuff that you are intentional. It's important. So what's the difference? So one is mindless engagement. The other one is thoughtful builder mindset, right? So it's like, from my perspective, that is the most important thing. When we are engaging with AI, we need to have a builder mindset rather than a seeker mindset or like consumer mindset.

So that would make us as well as the next generations smarter because now they are using it to do something that is useful or powerful or whatever, like create something instead of just, okay, let me just type something and it will give me, you may announce it. Obviously you are just gonna become dependent on it and maybe not become smarter because it is this.

smart, our job is to teach ourselves as well as our kids to have the builder mindset so that they are building things for the future and AI works for them, not they working for the AI.

Hermes Frangoudis (42:15)
No, it totally makes sense. And if you're using it as a learning buddy, a tool to make you smarter, it'll get you there a lot faster, right? Versus some of the hardships we had with just finding the answers to certain things, right? Now, it's not so much about the search. It's more so about, like you said, the purpose, the direction. This is the way I want to go. If I know I can find all this information,

How can I use this to put it together in a meaningful way to advance me more so than like, I got it to write me a story, right? Like I got it to do my homework. That's not going to help you in life. And the only person that those people are cheating are themselves. They're on a fast track to cheating themselves now. Maybe years ago they were trying to just write whatever to get the assignment done and it was apparent. Now it's the same mindset. Like, I'm just going to use the AI. You're cheating yourself out of an education.

at this point. And I think that's a big thing we need to be able to point out as parents.

Ram Katamaraja (43:15)
I think it's a transition period people need to learn. And also on the flip side, Ironman already taught us how to use Jarvis.

Hermes Frangoudis (43:23)
So true. We are coming up on almost an hour here. I'm loving the direction of the conversation, but I do want to be mindful of your time, you know, your CEO of a company and you got a lot to do. So I want to kind of wrap it up with this one wild card question I asked all my guests. If you weren't building Cola Berry and doing what you guys are doing, what would you be working on in the AI space?

What's another piece of the AI space that you find very interesting that you could see yourself working in?

Ram Katamaraja (43:55)
Yeah. So one space, it's, it's a common theme that I am seeing, which is as we interact with the enterprises, they talk about how to use AI. Right. So it's, it's still more like a technical guys, innovation guys. So they are like trying to do stuff, but yeah, I beyond the chat, ⁓ chat bots, can do like a lot more. ⁓ I would, ⁓ I would be.

trying to answer that question for enterprises. Okay. What can we do? Right. So, okay. Given an enterprise, can we diagnose it and what their business, what their processes are? And can we provide a guided recommendations on, these are all the various areas that how you can use AI. Right. So, and the way I would, there's an experiment that I'm doing like at this point of time, which I hope would be like the

future direction of Colaberry is ⁓ basically take a company name, scrape their websites and scrape their publicly available information, which gives us a very good idea of their products and resources, everything that they're doing. And then can we create like a roadmap for them to see how they can adapt AI in their enterprise for a short term, long term and midterm wins. and in the initial...

⁓ so the current who is on this call in the backstage, he's like doing that. So it's basically it's like both visualizations as well as the use cases. So those, those things are very, very engaging and for the people and suddenly it opens up their eyes, like what they can do. Because when you are in the vortex, you can't see it, right. But, ⁓ but, ⁓ but AI can like do an analysis. So that's, ⁓ and I would, ⁓ I would take it a little bit further.

And feed it all the, if it's a larger company, feed it all their past five, 10 years of the work that they did through all the projects that they did and extrapolate what their future is going to look like with AI. Where they can cut costs or where they can add agents, like sort of things. I would say AI discovery, right? So I would love to work on, I'm working on it and love to grow that AI discovery.

to like for enterprises because like large companies with large budgets could hire make-in-sees of the world like to do it for them but there's a lot of companies that that are like struggling and trying to figure out the path so yeah that's the a discovery is the something that i would love to do

Hermes Frangoudis (46:33)
huge

need for that. Especially for small to medium sized style enterprises, Like the big ones will always have that, like you said, the McKinsey and that white glove and it's a different experience for them. the smaller guys, can compete in the same way. And now this levels the playing field. And I'm looking forward to you guys building that out. We'll have to bring you back when you're at that stage, but hopefully before then, because I want to continue this conversation. I think you bring up some great, great topics.

Ram Katamaraja (46:51)
Absolutely.

Yes.

I enjoyed having the conversation with you Hermes.

Hermes Frangoudis (47:06)
Thank you so much, Ram. And thank you to everyone that listened in. For all of those on social media, please like, subscribe, follow the podcast, do that social thing, and we will catch you in the next one. Thanks for joining.

Ram Katamaraja (47:19)
Thanks for joining and watch out for Convo AI as the new UX.

Hermes Frangoudis (47:26)
what he said.

Ram Katamaraja (47:27)
Yes. Thank you. Bye.