Using AI for Better BI with Ajay Khanna from Tellius

DISCLAIMER: Below is an AI generated transcript. There could be a few typos but it should be at least 90% accurate. Watch video or listen to the podcast for the full experience!

Ajay Khanna 0:00
When we look at decision intelligence, the space where we combine machine learning AI with this natural language interface where business users of the analyst can get answers across terabytes of data, you know, understand not only what’s happening, but why things are changing and how you can improve outcome. That definition isn’t very clear. It’s not a clear cut define it, because it’s a new category in a way.

Alexander Ferguson 0:28
Welcome to UpTech Report. This is our applied tech series UpTech Report is sponsored by TeraLeap. Learn how to leverage the power of video at eraleap.io. Today, I’m joined by my guest, Ajay Khanna, who’s based in the Washington DC area, and he’s the CEO and founder at Tellius. Welcome, Ajay is good to have you on. Alexander. Great to be here today. Now, Tellius if I understand correctly as a business analytics platform, specifically designed for enterprises out there that are trying to get more data driven insights faster to be able to make better decisions, we all want to make better decisions. So that I get that correct, though, what you guys are focused on? Absolutely correct. Absolutely. So help me understand being okay, you’re focused on enterprises and then trying to make better decisions, what’s the real challenge them that you’re, you’re seeing that they’re facing and feeling every day?

Ajay Khanna 1:18
Now? That’s a great question. So the challenge, we see Alexandra that organizations are collecting when massive amount of data, but when we look at the amount of insights they are creating, there’s a massive insights gap there. And the reason we see this gap and the insights gap is there are two silos there we see the BI and the dashboarding silo. And then we see the machine learning and the AI silo. Now let’s talk about the BI and dashboarding You know, every organization use some kind of reporting mechanism. But But mostly organizations with the amount of data they are collecting have hit a wall with their current bi stack. And they want to also move beyond BI dashboard, not look at pretty pictures and reports not have high level KPIs, but be able to drill down and understand why things are changing. And that’s something a team, you will hear from me a lot, you know, this, I’m a big fan of Simon Sinek. Start with why. And that’s actually kind of the core theme for us in Telia is, you know, as a business we are doing but also, not only with the technology side, but as a company on so we really believe in that. So organizations want to understand by and, and that’s becoming really, really challenging with the current bi technologies. Now, on the other side of the spectrum, you see machine learning AI, and you see, you hear a lot of that dog. But those technologies are still not accessible to the business teams, they are still used by the advanced users who understand what those algorithms are. So that’s kind of leading to where the business uses business analysts are not able to leverage these machine learning AI technologies to get those insights from the data. And that’s the problem, we see

Alexander Ferguson 2:52
this terminology, say siloed it’s like they cannot connect it up to this point. You say bi business intelligence dashboard of pretty numbers, but they can’t get any further than just seeing the top numbers, then the machine learning which gives you all the interesting insights. But it requires a data analysts or data scientists to be able to to work with that. You’re trying to bridge that gap because that of that challenge, is that what you’re focused on?

Ajay Khanna 3:21
That is absolutely correct. We’re trying to bridge the gap. And the way we see that if you’re able to bridge the bridge that gap, I think there’s lots of things will happen. First of all, that if you see if you hear about these large numbers out there that 80% of the MLA projects fail, right. And the reason they fail, isn’t that, that, you know, they were developed, they were not the right, technical resources, but more importantly, what business teams and the data seems tangency. What are they aligned or not. And that’s a big challenge, you see. So if you bridge the gap, then what happens is, you can bring in the domain expertise of the business teams, they understand their data, they understand the problem they are trying to solve. And then you utilize machine learning and AI to do what they do best. and machine learning is really good at churning data, figuring it out the key contributors. And when you combine these things together, now the magic happens. And that’s kind of where we see a huge, huge gap. And that’s what Belize is focused on. Oh, gee,

Alexander Ferguson 4:17
do you just really enjoy enterprise data? You’re like, Oh, I wake up every morning. I’m like, Yes, I can’t wait to help people understand their data more, is what drives you.

Ajay Khanna 4:29
Yeah, I would say it’s a it certainly is certainly an evolution and journey where we know at some point, it started with more kind of this this aspect of bringing in automation. So it started with the automation was an evolution and journey. I don’t think if I if I rewind back, you know, 15 or 20 years back that, you know, I didn’t even know what data analytics mean or heard of that term. But it started with this desire of automation. And I think this one thing which always stuck with me is that that you know, there’s there’s two scenarios, either you automate Something or your job is going to be automated. So you pick one, right? So, so so so I prefer to pick the first one, like, let’s automate things rather than automating our jobs. So, so the way it is that, and that’s when it started, but then eventually, as we were looking at automation, I come from a telecom background. And, and we were actually doing all this analysis for telecom data. And it was a very niche problem. At that point, we were solving that, you know, how do we, you know, look at all the data and figure out which area knew we need to focus on in telecom scenario, working with companies like at&t, T Mobile, Verizon, and that then are going to give this sort of this idea is like, Oh, my God, this is just not Telecom. But you know, if you expand it to different industries, you know, there is there is certain opportunity to unleash power of data. And I think if I look at the last few years, yeah, that’s certainly something which, which excites me get it get me out of the bed and said, you know, what we are we are sitting on this massive amount of business data. Now, how do we how do we unleash the power of this data? Right? How do we get the business opportunities out of the data, you know, in a much more easier way than we do today.

Alexander Ferguson 6:09
Your history just mentioned, starting in Telecom, you were actually handling the data working, and your previous company is one of the founding members, right as it became the CTO, is that correct?

Ajay Khanna 6:22
That is correct. So so my previous company, so my journey was that this is my second startup, the previous company, was named cell site and cell site was when we started it, it started as a services company, we This was 2004 or five timeframe. You know, we did not know where to raise money, how to raise money, none of that. None of us was rich. So we like Oh, how do we, how do we make this happen? And then we ended up doing services. So we said, okay, let’s go to, you know, work with T mobile’s horizons, at&t is they have some projects, we took over some projects. And then eventually, like, I have this always this, this, this kind of product, guy, you know, Product Manager, product guy in me always like I want to always build products. So we started with that we got some money from the services and generated enough revenue. And then we said, oh, let’s, let’s build a product. And then we ended up building a product, which we deployed with the at&t s and t mobile’s which was very much focused in the telecom space, it was more helping them optimize their their network quality. And but that’s kind of where you’re this, this whole thing, when we also saw a gap, which because we were building this path platform, and you be like, Oh, my god, there’s nothing out there. They are these pretty dashboards. And then you got these data science tools, which, you know, none of us are most of the team members did not know how to operate them. So we ended up doing a lot of SQL Python stuff, and the put all this all together, and then you know, build something. And then we like, it’s not scalable, right? We were like, Okay, this works for a certain use case. And also, so we figured out something where it’s like, Okay, this works, we understand, you know, we can automate some of that stuff. And that, combined with the fact that iPhone was launched, launched at that point, which the whole world benefited, and we benefited now, because when iPhone launched and at&t s and t mobile’s, they couldn’t hire enough people to manage those networks. So they ended up using their platform. So that was kind of evolution where we like, okay, so there is, there is a huge scope for automation, like what we had built was a very niche, you know, domain specific thing. And then machine learning was coming to life, and was becoming more accessible. So we were like, okay, so if we were to expand it, if you embed machine learning into an a bi, you know, combined bi and AI together, you know, then you can actually apply this aspect to more industries.

Alexander Ferguson 8:48
So automation has always been a focus for you, Allah you already said either be automated or automate other things. I’ll choose to automate other things. Where did then the business intelligence side it was it right around this area where where you were helping as a services, then you built a product and you’re realizing providing business intelligence? There’s a big need for it.

Ajay Khanna 9:10
Exactly. So what was going on? Was that the no way you can take the previous company was very custom bi for telecom analysis, you can think of that right. So it was a because what is bi business intelligence, right? So business intelligence, we want to get some information from the data presented in some way to the to the team. So that’s kind of the the objective at that point was but we also were looking at that going a little bit beyond bi meaning I would call bi a lot of reports and dashboards, but then you have to understand, you know, why things are changing. So the so we were like, Okay, why a network quality dropped last week. You know why? This customer complaints have gone up by 50% last month, and all these questions, how are we answering it? We were actually, you know, created some business logic using SQL Python. And then we embedded that to the knife UI, and then, you know, heading out to the users. But then that going to, wasn’t the idea, which worked fine for some use cases, but was not a scalable idea. We were actually expanding outside us, we were actually expanding in Latin America, we were going to Asia Pacific. And all these areas. When we went there, we saw that the logic which worked fine in LA, and er, didn’t really work in in Johannesburg, South Africa, and then like, didn’t work in, in, in in Dubai. So what that meant was that you need machine learning, which can learn from your data, and can tune the rules, then provide an output, which is more tuned to the current scenarios, also,

Alexander Ferguson 10:45
integrating machine learning into this model than was that a simple process? It

Ajay Khanna 10:50
was actually, it’s all of those things, which they never turn out what you originally expected, which actually, I always say gifts is always a good thing that that niceness really works for you. So you know, we went to the first company was actually a pretty good exit, we actually ended up growing from few of us in 2004, to we were about 400 of us doing 100 million in revenue in 2014. When we got approached by a company named amdocs, we ended up getting acquired by amdocs. I don’t know if you know, amdocs is in about a five to 6 billion company publicly traded company, and they wanted to move or expand into the network analytic side. So so I was with amdocs for a year. And then I was like, Okay, so, you know, let’s see what would be new, what what to do. And I think that the enterpreneur, bug was coming back, and it’s like, oh, I need to do something, I need to build some product. And with that, we start looking at is, this idea is like, Okay, so what we need to do now is, is, is looking at these two, three areas, one of thing was that bi really hasn’t changed in the last two decades. So are one of the driver force was that the, the dashboarding, and its interface hasn’t really changed. So it used to be dashboards, on desktop became dashboards on clouds, you know, client server then became dashboards on cloud crystal dashboard. Right? So what we’re thinking is, how do you create a revolutionary new UI? Like, you can think of an iPhone for analytics, like kind of the idea, like how do you create an iPhone for analytics, right? And for us, that was that you have to provide like a Google like search interface. So what we thought is like, okay, we need a really like a Google for your business data, right? So you have a search interface, which we use for everything we do today, right? We, we want to book a flight, we go online, and you know, do something and then get a flight booked in a matter of a minute or two, you want to buy a pair of shoes, you can you can do that in a matter of a few minutes. But it’s not the same for your business data. All right, if you have a business question, you can go and type a question saying, Okay, tell me, you know, what happened to my campaign conversion last week in California, you know, for social media. Right, right. You can type a question today, which is actually sounds pretty strange, you know, especially now we’re in brain 21. When we were starting to think, in 20 2016, we like, okay, there should be a Google like search interface. And, you know, for the data, so we started working on that I get, and at that point, none of that existed, like for your business data. And especially when you have terabytes of data in your system coming from disparate sources. And you want to provide this interface, which is very different than a typical search. Because typical search, you’re just looking for the documents. In this case, when you ask a question, there’s going to be a SQL query underneath, it’s going to go and query the data data chart on the fly. And it’s all happening in like, one to two seconds. So a chart gets created on the fly. So we started looking at it, we started thinking, Okay, how do we how do we find the right people and ended up going to experimentation, like any other startup would be ended up finding some individuals do a lot of that experimentation ended up building that interface, which was it took us almost two, two and a half years. You bootstrap this first portion yourself? Yes. So one, one thing we had miners that are By the way, no previous company was an interesting story that we, we we never raised money. We never raised money in a previous company,

Alexander Ferguson 14:19
customer customer funded, it was all like that i

Ajay Khanna 14:23
can i doing that. So so it actually so it’s very interesting. When you talk to the CEOs or founders, or some of the team members, their normal kind of resume would be a waste so much money, right? So when I say in our previous company, we did not raise any money. But we actually had almost 980x, almost 1,000x ROI on the initial investment over the nine year window. So but it was fun. It was fun. We made a lot of mistakes. So we know what mistakes not to do in the new venture just make new mistakes there. But, but it was actually fun. But what ended up bootstrapping that for two and a half years, so I put in some of the money went through from my previous venture. And then actually I pulled in my other business partner from my previous company, because he had, he had, he was almost retiring he was he’s about like eight to 10 years older than me. And he, you know, he’s like, almost like a mentor and kind of big brother to me. So I pulled him in, and like I said, he put in some money. So we did not raise VC money, till last institutional VC money till last December, which was almost, I would say, less than five years, almost five years into this. So we spent three years building the platform, and then two, two and a half years in the market, get those initial customers and we had some good names. And then we went out on a summary that there was a reason for it, and we could have raised some money. But we knew that if we were to raise money, that that would actually inhibit us to build that solid foundation we wanted to build, because, you know, once you get the VC money, then you, you know, you have to that you call the cycle, that 18 month fundraising cycle, and it’s hard to get in, you know, go and completely disrupt that. So it took us three years to build that platform. But we are so glad we did that. Because this space is is a noisy space, and you have to build a solid foundation, if you want to go and cause a disruption. You mentioned,

Alexander Ferguson 16:28
you learned a lot in your first venture. So you were could not then do those same issues only new, new, new, new challenges at this time. If you had to think of one lesson learned from your previous venture that really helped you then execute, tell us even faster, what would what would be that?

Ajay Khanna 16:48
Yeah, I would say that the one of the lesson which we learned is, which is aspect around, you know, this product market fit, you know, this takes just way longer than then then you ever expect. So that was kind of one of the lessons, which is what helped us where we can undo that, and which is where a lot of the first time startup founders have that challenge, because, you know, they tend to underestimate that they tend to underestimate that they like, Oh, yeah, we’re gonna build a product, it’s gonna, it’s gonna happen. So that is one of those, which is what helped us kind of plan through it, you know, make sure we have the right product, make sure we go through some of those initial beta customers. So that’s kind of one of the lesson what’s a tactic when you’re working on product market fit

Alexander Ferguson 17:33
that that has worked for you to then really drill in and find it?

Ajay Khanna 17:37
I would say that the I don’t think there’s one going on to you, it is certainly an iterative process, I would say it’s an iterative process, where I think the first thing I would say is that you got to start with some, some defined criteria and objectives, but then then can I’ll keep listening to the market and can achieve that. And that’s what we did, even even though we, we, you know, this was kind of the second company I was in, but we still made lots and lots of mistakes, we hadn’t, you know, unlocked all the formulas there. So what we realized was that, yeah, there was, there was a vision we had, but then we actually started going to the market and kind of testing that hypothesis. And even if some of those customers end up buying it, like, you know, that doesn’t really mean that you have a product market fit. Because I think the real inflection point happens, when, of course, you have, you know, a certain number of customers, but there is also clustering, start forming, because clustering in terms of the the inhibitor, very similar behavior. And then also you start seeing kind of the adoption of the product kind of taking off. And I think, when you see all of this happening, because earlier in an earlier time, we saw different companies adopting the product, but they were all very different use cases are so because we didn’t live in lies, because analytics, the biggest challenge you have is the there is no, there is no one use case, right? The Art of possibilities is up to your imagination, which makes it really hard and challenging. In this this area there.

Alexander Ferguson 19:11
This is something I’ve been talking to some other SAS founders of when you have a product that can be utilized in so many different ways and across industries. How do you build a common feature set a common ability that serves them all, without trying to go too far down one lane and alienating everyone else? How are you balancing that then?

Ajay Khanna 19:32
Yeah, no, I think that’s a that’s a great question. So in analytics, what happens is because that’s something which we are very clear that we won’t build something which is going to work for one or two customers or a few customers. So that’s something which you always have to have a strong discipline around it. And it’s always hard when you have a customer who is, you know, ready to, to pay you and then how do you draw the line and how do you convince that it’s always tricky. It’s never easy. It’s easier said than done. But I think that’s that Where are the keys that you have to know which direction you’re heading, you’re not have to know your North Pole, like you have to know where you’re heading. And once you have that, then yes, you can go a little bit like you can make some variation. Like, I think you have a core, I will call it there’s a core. And then then there’s kind of the adjacent like, and you’ll learn on the other senses, like you know which ones are resonating better than the other. And then there’s an area where you don’t want to go and you have to know, you have to really define that and then stay within that. What have you seen when it comes to analytics

Alexander Ferguson 20:33
and business intelligence, like the core use cases or abilities that everybody wants and needs.

Ajay Khanna 20:43
And what we call this categorically intelligence we call this category is decision intelligence and the difference we see in decision intelligences. So bi bi is pretty well defined meant as a well defined meaning, you know, you want reporting and dashboarding, I think that’s kind of how bi has been synonymous. When we look at decision intelligence is a space where we combine machine learning AI, with this natural language interface, where business users of the analyst can, you know, get answers across, you know, terabytes of data, you know, understand not only what’s happening, but why things are changing, and how you can improve outcome. That definition isn’t very clear, it’s not a clear cut, define it, because it’s a new category in a way. So that is something which really excites us, because that’s kind of setting up a new category. And, and, and, and, and when we even talk to the analysts, which is a very interesting thing we hear is like, you know, what breed? Are you? Are you the BI breed or the data science creed? And and our typical answer is that you don’t have to fit into one of these breeds. Because there’s lots of things have happened from technology perspective. So this new category of decision intelligence, which is going to bridge the gap between the BI and data science, and when we are excited to be kind of setting the standard for that, like how that is going to be defined. So I think the organization still try to get wrap their head around what that is, which is a challenge in its own way. But it’s also I normally say I want to be be, you know, I don’t want to be one of those 50 companies trying to sell this, I want to be one of them who’s trying to convince the customer than users is a better way. And that’s a new category, which is, you know, its own challenge. But I think it’s exciting to be to be in that category.

Unknown Speaker 22:27
You

Alexander Ferguson 22:29
paint the picture of when it comes to new technology, a new way of doing things, the challenge of getting someone who’s used to doing it a certain way to make them realize there is a better way. How have you done your messaging? How have you been able to convince people or engage them to come away from what they had been doing, to seeing that there is a better way?

Ajay Khanna 22:55
Yeah, no, that’s a great point. And it’s not that we have figured out the magical answer for it. But I’m what has really helped us there is, is focus on the customer use case and problems. So when we talk about decision intelligence as a category, I think that’s one side of that positioning. But when you get down to the, you know, in the weeds, talking to the customers, they really don’t care what they really care about, you know, the problem they have and how we can solve that. And I think that’s a challenge I was talking to someone about kind of a challenge in general mln AI is facing while there’s a lot of potential is in how do you how do you connect that with the pain point business user or the analyst is having so what has helped us is focus on few industry verticals and the use cases to start with so when we look at the the the technology side, which is very infrastructure and industry elastic, but our go to market is not so go to market is very focused on let’s let’s make sure that we can 50 Industries we are focusing on like pharmaceutical Life Sciences, you know, consumer goods, and financial and even within that, like few certain business units. So what we go is when we go and pitch there, then we can actually just say how we are helping other pharmaceutical companies to do ABC, like, like, whatever that is. And that’s is kind of what’s working for us. And then they say, Okay, okay, yes, you have a new way of whatever problem they have, which is like they are looking at market access or rebate analysis, or you’re looking at improving the commercial effectiveness to launch the bugs in the market. So we approach it that way, which actually is much more effective. And actually the other side, the customer can understand it. And then we can we will keep expanding the library of those use cases as we grow and that there’s a lot of potential there.

Alexander Ferguson 24:46
The ability for a company to get better intelligence, I’m sure as everyone wants everyone’s be able to make better, better decisions. This technology, doing a Google search, one would say have insight Your company to figure out the answers that you want. is is is how is that changing and growing? Is it just like works perfectly all the time? Is it? Is it flawless? Or is it also still something that’s being developed and grown? And it’s still a future to get to? Putting my sales cap? Yes, it’s far less. It just works. Yeah, of course, all the time. Oh, it

Ajay Khanna 25:21
looks all right. Now, I think just that practically, the way it is that you still have to connect to different data sources, and data is never clean, all these typical challenges are out there. So the way it is that while our focus is to make it really self service, so people can just, you know, you know, plug this in with your data sources, we have provided the flexibility where you can actually go and customize some of these things. So I think that challenge, which is very interesting, we were talking to someone is what we focused on was, how do we combine this ease of use with the flexibility and that’s a very tricky thing in MLA, because either you get, let’s say, if you get a, you know, your iPhone, it has an embedded machine learning AI, but you can do customize it. And then on the other hand, you have these, these data science workbenches, which are very, very customizable. So how do you combine that, and that’s kind of the the focus we had is, when we go and connect to the organization to go and work with the organization, there is a lot of cleanup needs to happen clean up in terms of how they define the names of these columns, like so if there’s a data engineer, you know, they always like to put underscores there. It’s like, you know, revenue underscore 123 B. And then the business is in a state, oh, I don’t want to put an underscore and and I want to put, you know, normal word like, you know, this is my region, region revenue, or whatever revenue. So how do you map that, that interface, that’s one and second thing is going to setting up the data relationships is also key base, that you can set up how different things are linked to each other, because you businesses may have some better understanding on why they set up their data model that way. So there’s, there’s, there’s somewhat of that, which need some of that initial work, but it’s still pretty self service. Once you you bring the data in and you hook up those spikes, you could actually we have scenarios where you can get it up and running, you know, in matter of a day or two. So it’s it’s not that they the the effort wouldn’t be there. But yes, if you want to be very in you can do a lot, there’s a lot of customization people want to do. And that’s kind of where they can end up spending a lot of time.

Alexander Ferguson 27:26
If we paint the picture of the future of where we’re headed. using machine learning. I mean, what’s what are some of the barriers that need to be broken down, that needs to change for us to get to the world where it doesn’t matter how people form it name things, or whatever it just, it works, what needs to what’s the future look like? What do we need to change?

Ajay Khanna 27:47
Yeah, I would say the the first piece which we see is that the the cult cultural aspect, I think is is a big one right now. And which is where we go to the organizations, one of the challenge where you see one group really excited about machine learning AI, because you know, sometimes it’s a top down initiative, like CEOs, like oh, we gotta get machine learning here in the company. And then most of the, when you talk to these individuals, they’re like, I don’t know what that means. But we need machine learning AI, so AI, but then there’s other set of users who are like, you know, we don’t know, we don’t understand it, we have a list of 50 things we need to do on a daily, weekly basis. Leave us alone, right? So I think the cultural piece, meaning that kind of providing the guidance and knowledge around, like, why we’re doing what we’re doing, and how it’s going to help you and eliminate some of those fear they may have, like, their jobs will go away, or it’s too complicated. One of those scenarios could be that. So I think that I feel is there, I feel from a technology. But so this is more of a cultural, I think from a technology perspective, we are at an interesting inflection point with, you know, with the cloud data warehouses, like snowflakes, and redshifts of the world, you know, can store massive amount of data. And I think when we see these interfaces, new in professors, we call like, you know, this Google, like interfaces coming in, people are still not sure how that’s gonna work, or how to use that. I think that’s gonna I feel like it’s very similar, like Google Search were launched or search of different search engines were being launched in late 90s. When, you know, we had AOL then being them, then there was, you know, Yahoo, Google. And, and then people like, okay, you know, it makes sense. I think we’re very similar point in business data. business data is maybe slightly more complicated, because there’s more nuances there. But I think that’s the future. I mean, there’s no other way. Like, you can be relying on other analysts, or other advanced users to give you the answers to your basic questions like any business doesn’t make sense, like where we are right now.

Alexander Ferguson 29:50
Being able to get to this future and for you to continue to build obviously, having the right team is essential. And it sounds like you already from the beginning that you’re feeling on that, has it been a challenge to find good talent? Because I, I’ve heard this, that there is quite a battle for good AI and machine learning abilities and talents. Is that the case? Or is is there a bit of a big growth in the the pool of available talent for people in AI?

Ajay Khanna 30:19
Now, I think we I wouldn’t say we didn’t face that much. I think there’s a lot more hype around it then than actual reality on and I think people will also confuse about who they need, I think it’s really, really not clear to organizations who they need. So there is there is what we call as as machine learning AI engineers, and then you have applied ml. And if you have heard of that, but the point being that, when you if you want to try to build a new machine learning AI algorithm, yes, you need those individuals who have a PhD from Berkeley and all that stuff, right. But most of the organizations aren’t really doing that, because there’s a lot of open source components available, where you can actually stitch the existing pieces. So that’s what we call an applied amount. So how you apply these existing algorithms, which you can write with a couple of lines of code, it’s not that tricky, and then apply that. So and I think organizations are sometimes confused between the two, the lines aren’t very clear. So they end up hiding, sometimes these machine learning engineers, and make them do the other job, which then doesn’t make those MLSP is very happy and satisfied. So I think we didn’t face any such such issues, I think what we see is, what I feel is there’s going to be with the tools like le s and other companies who are innovating in this space. I mean, there’s a lot of other companies who are automating the work of creating ml models, I think the bigger need is going to be around the data engineers and the data experts. We call them data ninjas, the data ninjas, and the data ninjas are the ones who understand that domain knowledge, they have some SQL Python skills. And with the help of tools, like teleios, which is just automating some of the modeling piece, you can stitch these together that builder mentality, like, you know, like Legos, like you bring these together, and then create this piece, you can actually do that. I think it’s that’s where we see the future heading. And that’s how it’s going to be democratized. There’s no other way. This can be democratized and can be accessible, you know, in the organization’s

Alexander Ferguson 32:19
for everyone. Wow. I feel like your history and tracker as CTO and like, I feel like you were there in the weeds of it. How is your personal time changed? And how do you manage your time? I’m just curious, as a leader, for other leaves, like how are you a typical day? What does it look like for you?

Ajay Khanna 32:35
Yeah, it’s pretty, pretty, pretty messy there. I don’t think I would say what certainly has helped in the last year, which I would say the pandemic, there were a lot of things, which weren’t good. But one of the things which was were there was this whole aspect of I would say, better productivity gain. And, and and better management, time management, which I think I certainly struggled with in the in the prior years. So I think the the aspect of just men, and as you get the to do lists, which I think I always had to do lists, but the stimulus would normally get hijacked or get distracted with the other other things which come to that, I think certainly has been better. So I know, we all are thinking, what does this new hybrid work is going to look like? Is it is it you know, two days a week? Is it three days a week? Is it one day a week? Is it no zero? I don’t know. I think I feel it’s going to be I think few few days a week, maybe maybe a couple of days a week. And I see that that hybrid, I’m actually very excited. And looking forward to that I think that hybrid, which, you know, certainly provides this balance of productivity and that interaction, can I get the right mix there?

Alexander Ferguson 33:50
I am curious, also of those who are going back to the office or not as the future that we’re in, everyone’s just remote. Wherever head looking at just the future for a second overall. What are you most excited about? In in, in, in the business intelligence area? For sure. But also, I’m curious for machine learning and where it’s headed and the opportunities it has?

Ajay Khanna 34:14
Yeah, I would say that I mean, we are very, very, very early stages of this fall, you know, machine learning AI, there’s a there’s a code, which I think Marc Andreessen used that, if you know, I guess about 10 to 15 years ago, he said that software is eating the word. But now like AI is eating the software. So and so what that means is that I think AI is going to be embedded in in everything we do, like pretty much software is right. So I and and I think with data actually, you know, just just looking at it like the amount of use cases and possibilities we are like not even 5% there in terms of you know, looking into that and say what’s the other possible so that’s what’s the alternative. The exciting part is as we progress further in like different use cases and what machine learning can do for for different activities there. I think we see a lot of companies coming in doing some innovative stuff there, which is going to really change the way we work and we access data and we access information and analyze information. I think that’s really changing. That’s really, really exciting. I think we, I would say, for the next four to five years, we will see a very rapid transformation happening, like there used to be this word, you know, data, digital transformation, I think the definition of the digital transformation is changing, and how AI is going to come in and, and change the way we work.

Alexander Ferguson 35:40
Do you think that this is only going to stay for enterprises and large organizations or will smaller businesses and be able to get a take advantage of machine learning and being able to have this business intelligence

Ajay Khanna 35:57
that would be our mission is that we want to democratize machine learning and AI. And he believes the best vehicle for democratizing ml AI is through the by putting it in the business, business analytics and business here BI tools. So we are actually working with some midsize companies who are actually starting to use machine learning and AI. And we also believe that’s the only way they can compete with these big giants, the Amazons and Googles, who, you know, have army of data scientists, I think the only way they can compete is by utilizing different companies, we’re innovating in that space, we absolutely see that we already have customers in that space. But I think they’re still in the early stages of this whole journey.

Alexander Ferguson 36:41
It’s all goes comes back to data. Right? Hey, how do people are they keeping good access of their data? And can they are they ready to be able to pull it in for machine learning? Take advantage of it?

Unknown Speaker 36:51
Yeah.

Alexander Ferguson 36:53
Well, I appreciate it for this this journey that you’ve been on and and also your passion, right? You already said like I don’t want to be automated. I want to automate things and help other people automate it. And then the this future that you paint, it’s it’s exciting to see where business intelligence will take us with machine learning. For those that want to learn more, whether you’re an enterprise or mid market or service and want to start looking at better ways to take advantage of your data, you can go over to Tellius.com That’s TELLIUS.COM Thank you so much, Ajay, it’s good to have you on. thank thank you, Alexander. It was a pleasure talking to you. And we’ll see you all on the next episode of UpTech Report. Have you seen a company using AI machine learning or other technology to transform the way we live, work and do business? Go to UpTech report.com. And let us know