What's Up with Tech?
Tech Transformation with Evan Kirstel: A podcast exploring the latest trends and innovations in the tech industry, and how businesses can leverage them for growth, diving into the world of B2B, discussing strategies, trends, and sharing insights from industry leaders!
With over three decades in telecom and IT, I've mastered the art of transforming social media into a dynamic platform for audience engagement, community building, and establishing thought leadership. My approach isn't about personal brand promotion but about delivering educational and informative content to cultivate a sustainable, long-term business presence. I am the leading content creator in areas like Enterprise AI, UCaaS, CPaaS, CCaaS, Cloud, Telecom, 5G and more!
What's Up with Tech?
AI is Not What You Think: Insights From a Global Tech Leader
Interested in being a guest? Email us at admin@evankirstel.com
Pattern recognition masquerading as intelligence - that's how Zensar Technologies CEO Manish Tandon provocatively reframes today's AI landscape. With refreshing pragmatism, he cuts through the hype to reveal why most enterprise AI initiatives struggle to gain traction.
The fundamental problem? We're approaching AI deployment all wrong. "It's not the AI engine that needs to be made a hero," Manish explains, "it's the people who are using that engine who have to be made heroes." This simple yet profound insight explains why many employees resist adoption - they perceive AI as a threat rather than an amplifier of their capabilities. When organizations reposition AI as a tool that helps workers deliver better results faster, adoption naturally accelerates.
Through real-world examples like Zensar's "AI Engineering Buddy" and specialized healthcare applications, Manish demonstrates how successful implementation depends on creating "pull" rather than "push" adoption. Starting small with dedicated teams tackling specific problems, showcasing early wins, and focusing on significant efficiency gains builds momentum organically. The most revelatory insight? AI currently sits in enterprises' "cost reduction bucket" rather than "revenue generation bucket" - and until that changes, we won't see explosive growth in adoption similar to what e-commerce sparked in the digital revolution.
Ready to transform your approach to AI implementation? Start by rethinking how you position these tools to your team. Focus on empowering your people, not replacing them, and create opportunities for small, visible wins that demonstrate real value. How might repositioning AI as a capability amplifier rather than a replacement change adoption rates in your organization?
Growth marketing tips & tech insights from those who’ve done it.
Listen on: Apple Podcasts Spotify
More at https://linktr.ee/EvanKirstel
Hey everybody, really excited for this chat today with a global business transformation powerhouse at Zenzar, with Manish. How are you? I'm doing very good, ivan. How are you in the enterprise? What's working, what's not working and the future.
Speaker 2:Before that, maybe introduce yourself and Zensar and your mission within the company. Yeah, I think I'm a CEO of Zensar Technologies 11,000 people globally, about $650 million, 50 million revenues listed on the Indian Stock Exchange. Primary markets are US, uk, europe and, interestingly, south Africa, and I have been CEO for about two years. At Zensar. We specialize, we are a different sort of technology services company, primarily because we just don't focus on engineering, but we also focus on experience and engagement and how you use technology to influence all these three things together, which is experience, engineering and engagement. So a different sort of a company, agile, fast-moving, customer-focused company. And you know our vision and mission is, our purpose is really to make sure that we are delivering collaboratively with our employees, customers, society at large, to better futures of mankind, so to say.
Speaker 1:I love that. I love the mission and you've seen a lot from your time at Infosys to now leading Zensar. You've seen a lot of projects go right.
Speaker 2:And enough projects go wrong also.
Speaker 1:Well, that's what's a little more interesting to discuss is where are companies going wrong when it comes to rolling out AI to their employees or even their customers? What are some of the pitfalls, roadblocks that you're seeing?
Speaker 2:Yeah, I think. First of all, ivan, I have a slightly different view. I think we as an industry have done a great disservice to calling this thing as AI. So I'll start with there From there. See, essentially what we have is a great tool set for pattern recognition. We pattern recognition on the spoken word, the written word, the videos, any sort of content, and the ability to look at these patterns and predict what is going to come after that. Whether you can, because we are calling it AI or artificial intelligence, you know, just because you're calling it, that a whole set of issues is arising which is actually preventing the industry from moving forward with this technology. So I would first of all say the biggest problem in adoption of AI in the industry is understanding what this technology is capable of and also understanding what this technology is not capable of.
Speaker 1:Right.
Speaker 2:And we are going into, you know, artificial general intelligence and suddenly everyone will, humans will be replaced and, you know, all sorts of things will happen. That is not, that is not really the case, at least it is not going to happen in the next decade or so. Maybe it will happen after that, I don't know. Uh, so I would say the biggest problem is thinking of this technology as artificial intelligence or intelligence and hence limiting the use cases and limiting the possibilities to which you can apply this technology to. So that is the biggest hindrance to adoption, and maybe you've not heard this from others, but that is what I personally feel as I look at all the hype around this technology.
Speaker 1:Wow, it's quite a pragmatic take. You're not just pumping the sunshine pump here. That's interesting. I've read you also see a lot of AI projects failing, not because of the tech inherently, but because they're not built for the people using them. Talk about that user experience, employee experience, customer experience. What's missing there?
Speaker 2:Well, I think the first thing, that mistake that people are making, is they're making the AI engine the hero. It's not the AI engine that needs to be engine the hero. It's not the AI engine that needs to be made a hero, it's the people who are using that engine who have to be made heroes. So that is the first issue that I would say that if you go and tell a person that hey, start using this, and, by the way, out of the thousand of you, maybe a hundred will be left after this, okay, then you know that adoption is not going to happen. But, on the other hand, if you tell them that you know, by using this technology, you can deliver, answer your customers' questions in five minutes instead of 15 minutes, that is positioning the technology as amplifying human experience, not reducing human experience, right. So I think that is one of the key things, that that is. That is the problem. The second, from a human element, the problem is that you know you have this. What do you call a new job description, called prompt engineers. Now, a prompt engineer is not, it's a specialized role. Unless, until, how can you even talk about AGI, ai et cetera if you can't even have a human-to-human kind of conversation with the tool.
Speaker 2:So I think, as far as the human element of this technology is concerned, those are the two things that are extremely, extremely important. And the final thing is the technology today doesn't have the intentionality or actionability right, so it can tell you do this, but it cannot go and do the thing itself. And that is why I feel that the first, uh, big use cases of this will come either through agentic ai or actually through robotics, and you will be surprised to. You will be surprised to see the development, um, that this technology will have, the, the development that this technology will have, the impact that this technology will have on robotics stuff, and we are not seeing it here, but if you look at what China is doing with this technology in robotics, it's eye-opening really, and that brings us to that intentionality or actionability of the technology overall. So I would say that those are the three big things that are hindering the adoption of this technology.
Speaker 1:Oh, great insight. So I, as a content creator, broadcaster, podcaster, I mean I'm using AI tools all day long. I'm a tech geek, but I don't think I'm the norm, for obvious reasons. A recent Pew study showed that most workers don't use AI much at all really in their day-to-day jobs. Why do? You think that, why is adoption still so low? Why is adoption still so low?
Speaker 2:Well, I think, first of all, what has been presented is a very, very high level or a very initial version of the technology. What has been presented is you know, if you think about integrated circuits or chips today and you look at the transistor that was originally done here in Bell Labs in New Jersey, there's a huge difference there From a technology perspective. What has been presented just now is really that old. That initial transistor it is not the integrated circuit or chips that we are accustomed to, and when that happens, people play around with this technology and see where the sweet spot is and where they can utilize it. So we are still in that process as of now.
Speaker 2:And you know a lot of good things. The most cases of this technology, at least in the enterprises, is being used more as a personal productivity tool than anything else, right? So, yeah, I get like 100 documents every day to read it. I'll dump five documents and summarize these. Then I'll decide whether it's worth reading or or not. So, uh, you know those are the kind of use cases, the personal productivity kind of use cases, and that doesn't excite the enterprise to invest too much, because you know I can improve productivity of each person by 5%, 10%. That doesn't mean much.
Speaker 2:This technology is not cheap. You know, something like $30. I have heard two thousand dollars a month, all sorts of uh, all sorts of numbers. So this technology, uh, is not a cheap play thing. So I would say that is uh. That is one uh. One reason, and second reason is for enterprises. There are technologies that impact revenues and there are technologies that impact cost. Currently, this technology has been put in impacting cost bucket much more than impacting revenue bucket bucket. Much more than impacting revenue bucket the day this technology becomes more impactful and positively influencing the revenues of enterprises.
Speaker 2:that is the time when real adoption is going to happen, Because that is the time business will really spend money to get this thing going. And we have seen this with digital technologies overall, where all this e-commerce and so on, it created entire new revenue streams for companies. And that is not what is happening with AI just yet.
Speaker 1:Wow, such a great insight. I haven't heard that before. So I've spent 30 years in enterprises I work for Intel and Philips and Oracle and you were just sort of handed over a suite of applications and tools to use whether they were good, bad, middling, and you were kind of stuck with them. I think we've come a long way in terms of usability and employee experience with these tools, both from the vendors and the companies themselves, but still, how do we get employees into the AI conversations early so they're not just, you know, handed a bunch of tools but really can be involved in the process instead of just being told to use them?
Speaker 2:Yeah. So, interestingly, you know there are multiple ways of getting something adopted and I believe that with newer technologies, you have to show early successes. It is not important to cover 100% of your employees, because you can always do it by push, but the real reason is, the real thing is how do you create a pull for the technology? So I'll just take example of Vinsa, right? I said you know, I gave the team 10 problems and I said go and figure out how you can use AI to solve these problems. Okay, we created a sort of agile squads for each of those problems and then they started getting into details, looking at what are the tools available, what is the problem, how can we solve it, etc. They came up with proof of concepts and said okay, after some proof of concept, we said no, the technology is not ready for prime time. So I'll give you an example.
Speaker 2:We do a lot of work for our clients on engineering, which is basically migrations. You are in the technology sector, so migrate from A to B or change the code base from good old COBOL to Java or whatever right. I said look, suppose we were to use AI for doing this. How could you go about it? And that led to us creating something called the AI Engineering Buddy, to us creating something called the AI engineering buddy. So we said the best way of doing it is, you know, trying to look at the entire process and see what all we can automate. So I said don't look at it. You know, you feed a pig at one end and a sausage comes out at the other end.
Speaker 1:That's not going to happen.
Speaker 2:But you know, let's try and solve this, this problem, and the team came up with some fantastic uh approaches using yeah net net. I would say that would deliver about 40 percent uh efficiency benefits, uh which I am happy to give to my clients. Actually.
Speaker 1:Wow.
Speaker 2:Right. So, similarly, I said, okay, this content creation, you know. So, as I said, we work with experience, and experience is basically design and engagement is basically marketing and services. So content is key for us. So I said, okay, let's see how we can accelerate the content creation process using the app. And the team came up with a lot of things. So we created another thing called experience buddy using those concepts. Similarly, we created something called data buddy using those concepts. Similarly, we created something called data buddy using those concepts. So it is just my approach is, you know, let's get the people working on small groups of people working on a problem, make that a success, show it out there and then the adoption you will create the pull of adoption rather than push.
Speaker 1:Wow, brilliant, brilliant approach, and you've worked across so many different industries healthcare, finance, manufacturing. I'm really interested in healthcare. How can we make healthcare do a better job of helping patients and employees through AI? How can we make it useful and patient-friendly, and are you doing any specific work in this area at the moment?
Speaker 2:Yeah, lots of work is being done in this area. First of all, we have taken this engineering body and data body and all those things and created technology and vertical specific versions of these, because these are generic bodies, right. So how do we apply it to this industry? How do we apply it to that industry? So we have done that kind of thing. But very interestingly, ivan, since you have been in the industry, guess who was the first real AI? What was the first real AI tool which was much hyped and talked about? It might have been the MIT chatbot, which was much hyped and talked about Watson.
Speaker 1:It might have been the MIT chatbot, the Eliza chatbot, I don't know. That goes way back to the 70s, so I'm not sure.
Speaker 2:I would say Watson, watson, if you remember. Yeah, they tried to. Again. The idea was one of the big sectors that they focused on was healthcare. But again, the problem with healthcare, with any AI, is that any AI will have an error rate. And what is the tolerance for an error rate in healthcare? Zero.
Speaker 1:Right.
Speaker 2:So, first of all, I would say, in healthcare, one should only be looking at I wouldn't say only primarily be looking at enhancement of individual productivity and process productivity, rather than, you know, replacing people. So, for example, we have done something. One big issue in healthcare and life sciences a very mundane issue is a lot of documentation has to be done. There are customer charts that. No, this is not the most sexy problem, but we said, okay, let's, this is a lower risk problem. So we took up something on adverse event reporting, for example, and we created a tool set on adverse event reporting. Similarly, again, we said created a tool set on adverse event reporting, okay. Similarly, again, we said we use computer vision and LLMs to do diagnostics or diagnostics recommendations. So, you know, you can feed it an x-ray and it will tell you based on that.
Speaker 2:Okay, I see some discrepancies here, there, et cetera, et cetera, but it is augmenting. The pathologist it is not augmenting, it's not saying that, oh, this is the thing. So it is trying to show. Whatever we are trying to do is how do we augment the human capability and efficiencies in that area, rather than trying to replace them? Because, as I said, even the best AI has significant error rates today. And healthcare you cannot afford any error rate.
Speaker 1:And, ironically, medical errors are prolific in the industry. It's actually shockingly one of the leading causes of death is medical errors by humans, and yet the irony is machines could do in some ways, perhaps much better, so there's a big disconnect there which needs to be fixed.
Speaker 2:How do we come? So? That's why the concept is more around how do we augment human capability and improve quality rather than, unfortunately, the conversation always goes into oh, this will replace me or this will replace that. That ain't going to happen, at least in healthcare. It's not going to happen. I can tell you very easily.
Speaker 1:Yes, indeed, you know. Final thought here fast-forwarding two, three, five years. What are you doing to get yourself and your team in a position to take advantage of this opportunity and also avoid disruption that might be coming for your business? Only the paranoid survive kind of idea. How do you see that unfolding for you and the team?
Speaker 2:That's a great question. I normally sleep very well, but sometimes, when I wake up, that is one question I have. So I would say, first of all, what my mandate to the team is that we need to bring in AI in every aspect that we are doing. So. You will not believe it, but I actually track that. If I am giving a customer proposal, how many of them have AI-related thinking into it? Not because every proposal should have AI, but I want to make sure that people are thinking my team is thinking that, okay, if I'm proposing this to my client, how can AI help in this? So the idea is that you have to.
Speaker 2:What I'm trying to do is get the entire team to look at all our processes and say how they are going to use AI in that process, right, For example, we have a methodology for implementing, say, Salesforce at a customer. Now I was going through some of the stuff that we were doing and I said, hey, we should actually come up with a Zensar specific methodology for Salesforce implementations.
Speaker 2:Use the AI? Okay, it will not. It doesn't mean that you know previously you were spending 100 months and 100 people doing it that it will be zero. No, but can we do it instead of 100 months? Can we do it in like 100 days? That should be possible, right? Instead of using 100 people, can we do it, uh, with 100 months? Right, that should be possible. So let's let's look at it more as a tool which can enable various parts of our business and make sure that we bring it in front of customers, because for us, customers, clients are everything. Clients are paying my salary, everyone else's salary, and the idea is that if you can influence, if I can influence my people's thinking this way, and if they can influence the client's thinking, then we hit a home run. They can influence the client's thinking, then we hit a home run. So that is where my thought process is on this.
Speaker 1:Well, definitely on that sort of mic drop moment. This discussion was a home run. I learned so much and I really look forward to following your progress onwards and upwards. Thanks so much.
Speaker 2:Thank you for having me.
Speaker 1:Ivan A pleasure, doc Likewise and thanks everyone for listening, watching and sharing. Take care.