Is There Anything Gen AI can’t do (Yet)!

Is There Anything Gen AI can’t do (Yet)!

In this episode, we dive into the game-changing world of Generative AI in talent acquisition with Sukumar Rajagopal, Founder & CEO of Tiny Magiq, and a seasoned veteran of the IT services industry. The episode unpacks the many ways Gen AI is transforming recruitment—from automating repetitive tasks to making smarter hiring decisions. We tackle some of the biggest myths about AI, including fears of AI replacing human recruiters and concerns about bias. Sukumar also shares his vision for the future of AI-driven recruitment, highlighting the evolving role of human recruiters and how organizations can foster a successful partnership between AI and people. 

Highlights

Imagine a world where AI isn’t just a “smart intern,” but an “Einstein at your desk.” Instead of merely automating routine tasks through prompt engineering.

Agentic AI uses knowledge engineering techniques to drive more disruptive ideas and opportunities. In recruitment, this could mean AI tools that autonomously source, screen, and even hire candidates based on predefined criteria and real-time insights, all with minimal human oversight.

Gen AI has made its mark in the recruitment sector with automation of tasks like resume screening or interview scheduling; and these applications can now be defined as Everyday Gen AI. Game-Changing Gen A takes things a step further. It reimagines the entire recruitment lifecycle by autonomously managing every step, from sourcing candidates to signing 

Myth #1 – Generative AI can replace human recruiters – While generative AI is transforming many aspects of recruitment, the idea that it will completely replace human recruiters is an overstatement. AI excels in handling repetitive tasks, such as parsing resumes or scheduling interviews, but it lacks the nuanced judgment, empathy, and relational skills that human recruiters bring to the table.

Myth #2 – Generative AI will lead to massive job losses in the recruitment industry and beyond – Rapid Gen AI adoption will drive higher demand for skilled professionals who can leverage and manage these AI tools effectively. The surge in investments and advancements in generative AI is expected to create numerous new job roles, especially in developing, deploying, and optimizing these technologies. As companies grapple with tech debt and growing demands, the need for human expertise in utilizing AI effectively will only grow. 

Myth 3 – Bias in AI is a real concern, but it stems from the data and algorithms rather than the technology itself. AI systems reflect the biases present in the data they are trained on. If historical data reflects discriminatory practices, the AI can unintentionally perpetuate those biases.

Whether you’re a recruiter looking to stay ahead of the curve or just curious about the future of work, read the blog version of this episode that is packed with insights into what Gen AI can—and can’t—do in talent acquisition and the innovations shaping the future of hiring. You can also stream the full episode here 

So is there anything Gen AI can’t do? Let’s find out. 

Sat J: Let’s start today’s conversation with a quick introduction of everything that you do at Tiny magiq, Sukumar.

Sukumar Rajagopal: Yeah, it’s interesting, right? So I’ve always had a challenge in explaining what we do. And the biggest challenge I faced was I was talking to a petrol bunk owner, he asked me, Hey, what do you do? I was like, how am I going to explain this to this person? Then I thought for 30 seconds and then I finally came up with an idea. I asked the person, Hey, do you use Uber and Ola to travel to places in a cab? He said, yeah. Then I asked, do you remember how it was when we had to hail cabs the old way— find an auto or cab? He said, yes. I said this is great. So can you see the transformation? 

So this is what we help large companies do in every system, process, function that they have in their enterprise. Bring disruptive innovation of that kind using digital and now generative AI to solve problems in an entirely different way. This is what we do for a living. 

Sat J: Fantastic. I’m sure the petrol bunk guy would have understood very well. 

Sukumar Rajagopal: Yeah, in fact, he came back with a lot of digital things he was doing. He was, he gave me a speech for half an hour on all the digital things he was doing. He was saying, Hey, I run the biggest petrol bunk in Chennai. There is a queue of trucks, which wait for three hours to get into his petrol bunk. Because he said, Hey, I monitor the quality, it is very precise, and while they are waiting, they have a relaxation space. They can send money to their family in the village. They can do all these digital transactions. I was amazed, a petrol bunk could think like that. So, I was happy that my message got through to him, and he was already doing disruptive innovation in the petrol bunk space.

Sat J: There’s a lot of buzz about GenAI and what it can achieve. It is capable of a lot of wondrous things from producing amazing content to automating repetitive tasks to conducting comprehensive research or even diagnosing medical conditions. 

What I want to ask you is what’s your take on the current state of Gen AI and are we close to realizing its full potential especially in the realm of talent?

Sukumar Rajagopal: I have never been excited about technology to this extent. I mean, the last big wave I came across was the e-business wave and I managed to jump onto that, but Generative AI is something that I have not seen in my 36 years experience. And I think there is a reason why people are not able to appreciate it and I think it goes to mindsets which we work a lot. And I’ve been articulating this only recently because I’ve been working in AI for the last 10 years through the CMI algo labs initiative as well as through Tiny magiq and the last year and a half whatever through Generative AI. So let me ask you a question, Satish.

So do you think Generative AI is a Smart Intern or is it an Einstein under your desk? Which side of the spectrum do you fall on?

Sat J: If you had asked me this question six months ago, I would have said a smart intern, but right now, some of the transformative things it is doing I would say it is slowly becoming an Einstein at my desk.

Sukumar Rajagopal: Yeah. So that perspective that you have, Satish, if you start to look at problems from that lens, you are likely to come out with more disruptive ideas. This is what I found. And I am surprised when big name experts in the world of generative AI say that it’s a smart intern. Now, there is an interesting conundrum or interesting insight that we found. If you merely prompted it using prompt engineering techniques, you will be in the world of smart interns. So at Tiny magiq, we developed this strategy called expertise augmented generation or a different way of describing what the GPT should do using knowledge engineering techniques. So if you use those, you will get access to the Einstein at your desk type of ideas and opportunities.

So in the world of talent, what I want to do is like, this whole world is becoming agentic and your audience definitely has heard this. What if there is an agent? This is what I imagine— what if there is an agent or a service which uses generative AI and tells me, “Sukumar Rajagopal, I know I have been studying your company, you are my client. I believe that you need three software engineers with these capabilities like tomorrow. I have done all the hiring, fact checking everything and tomorrow they will show up at your office.”

Imagine such a thing happening and I am taking software engineers, but why can’t it be done with any skill? So it does the whole thing, and I will not participate in the process at all. You do the screening, you pick up the best candidates, I will even give you the empowerment to hire them for a certain budget I can give you. So you do the whole thing. Imagine something like that happening. Now, why I say this is important is, I read somewhere that India is going to need another 5 million engineers with the latest and greatest generative AI skills. So, where are we going to find those people? I believe the engineering curriculum graduates about half a million of them. I don’t know if that statistic is correct. So we need 5 million. We need a broader pool of people. How do we get that talent? How do we train them in all these skills? And how do we put them to use? So this is my perspective on talent and generative AI.

Sat J: So in your experience, Sukumar advising companies especially on their disruptive, innovative journey, do you have an example of an innovative company that is revolutionizing the way talent acquisition is done? And again, right now, for the focus of this discussion, I’m saying talent acquisition, but it could be any other talent process where they have used gen AI and truly revolutionized the way they’re doing things.

Sukumar Rajagopal: I have done HR innovations. For example, we recently did for a large company, reimagining their intranet. So, this is what we imagined and we were able to prototype this also for this client. Imagine that I am an associate and an employee in a company. I go and say, hey, I want to file an expense claim. It automatically does the whole thing. I want to organize a trip to the airport. I have to go somewhere here. It books the tickets, it organizes everything. And it knows everything about me and it does that. So imagine an intranet to have a set of agents, which is the agentic intranet, if you want to think of it like that.

It does everything based mostly autonomously with just a simple check with the human just to make sure it is doing the correct things. Everything is agentic. So instead of apps, we have agents. If you remember the one Cognizant initiative, we built these small apps that will do, people have to still invoke them and do it.

Instead of that, imagine an agent which does that. So I want to do X, for example, this agent that I talked about in the previous question. I will just tell the hiring agent, hey, go to work. That’s all I will tell it and it will automatically determine what my needs are, hire the people. Sign the contract and make sure they show up tomorrow in my office. This part that I’m talking about I have not seen any company do but one of my clients is investigating the service called noon.ai Which is claiming to do autonomous hiring. I have not used them myself. But I went to their website and checked them out. It seems very interesting. If they do what they claim to be doing, it seems like a tremendous innovation.

Sat J: We keep hearing AI will automate everything. AI will put millions of people out of work. AI can do any task that is imaginable. And some of the examples that you gave kind of makes it seem like that. The hype around AI has given rise to numerous myths, both online and within organizations. Some fear AI is taking over the world, while others dismiss it as just another buzzword. Like every five years in the IT world, there is one buzzword. So this is just another buzzword which will die. But the truth probably lies somewhere in the middle.

I have a set of myths that I would like you to debunk for us, especially in the context of the talent industry.

So myth one – Generative AI can replace human recruiters. What do you think about it? 

Sukumar Rajagopal: So if the autonomous agent has to be done, it has to do what human agents do. Now whenever I think about these problems, right, I try to partition this problem. Let’s assume that human recruiters recruit on a complexity level from low, medium, high. So think about the low complexity hiring that you do. That Gen AI should be able to do. Autonomous. With minimal human input, but for the complex recruiting challenges, which are very custom, which are very tricky. There’s a lot of parameters, those things it’ll take harder, but I will flip this problem around differently. Satish if you’ll allow me. And I want to stick to the corporate context because that’s what I’m familiar with. 

So I was the global CIO of Cognizant and head of innovation. I looked at the CIO surveys of the last 25 years. This is an analysis I was doing for a customer. Can you guess what is the number one item on that list? What do CIOs struggle with, have been struggling with for the last 25 years? There are actually two things. Let’s see if you can guess even one of them.

Sat J: Adoption of new tech and it’s a use case in their industry.

Sukumar Rajagopal: Yeah, correct. Since it’s a CIO survey, it’s worded differently. That is, the backlog, that is the business wants something to be built with new technology, whatever, but the IT department has no way of prioritizing that and that this list is growing so big that it’s like very, very big for most big companies. The other side of this equation is as people roll out various technologies the CIO’s Technical stack has a lot of what we call technology debt. We make a technical decision about tech debt. It’s called technology debt in the world of Technology, which is I have made a technology decision because I want to go to market quickly. I might not have made that decision correctly, which creates debt, which I have to remove to a more robust thing later on and I don’t have time for that. Okay. So now this is the state. This means one thing. We don’t have enough people to develop all these applications and requirements that internal and external customers have. If that is the case, why would jobs decrease and not increase? 

Imagine that I wear a magic wand, all the tech debt and all the backlog of all the CIOs is erased. Imagine being able to do that. It is possible by using Generative AI technologies by everybody. Because imagine that one person can do instead of X amount of work, 5X amount of work. If I am able to get that type of productivity out of all the people that I already have, and I’m just talking about the fortune 2000 CIOs. And if you look at the corporate world, which is very, very big, every company has the same problem, but at a smaller scale. So if anything, the jobs are going to increase because of this generative AI is making a lot of people invest in it. The discretionary investments are going in, which means they’ll hire more people.

In fact, if I were you, I would be very happy that the lot of the hiring is going to happen. It is probably going to go through the roof in the next 18 to 24 months as people figure out how to deploy this effectively, which is not very clear at this point. There are very few people with actual products and applications. We being a few of those. There are very few products and applications in generative AI. So, people are figuring this out. 

Sat J: So you’ve already debunked the second myth – whether we’re going to lose jobs or whether more jobs are going to come into the world and I think you’ve already debunked that. So the third myth that we keep hearing is that Gen AI is biased and discriminatory because it has been built by a set of people and people are inherently biased and discriminatory. Do you think this myth is true?

Sukumar Rajagopal: Yeah, it is true. For example, you can test this yourself. Go to chat GPT and ask, Hey, generate an image of a scientist. The chances that it will generate a white American looking scientist is very, very high. Okay. Now the point is when you do this transaction, do we know that it is biased? I mean, it’s not producing, why is it always producing a white American scientist? So we have to now instruct it to say, Hey, don’t do that, create a diverse thing.

So it is biased. And I think that that bias also exists in human beings. I am finding it very difficult to believe that human beings are somehow not biased. In fact, we are more biased and that bias is what is translating. I will give you a classic example of what happened in the recruiting industry. I don’t know whether this is common knowledge. You may have heard of it. Amazon. com rolled out a machine learning AI based recruiting system. It rejected all the women candidates. First, this is a very famous case. You can Google it and see it. Now, why did that happen? Because they fed it all the past data of the kind of candidates they recruited. It’s possible that there was bias in the data which was biased against women.

So whatever biases that we have displayed in the public domain or in the private documents, books, everything, all that it has absorbed. So, it is going to be biased. So, it’s up to us to see how to now think of bias as a risk. Which I believe human beings also do. I mean, how many cases have we heard that a deserving person was not picked because there was some bias against the person. So we, the best way to answer your question is bias is a significant risk. How do I mitigate that risk in my solution? This is something I hope all solution designers are thinking about and not just going with whatever the generative AI is producing. If we did that, I think we would be remiss.

Sat J: I want to talk a little bit about your vision of how the future of a talent acquisition system will be built by generative AI. And you gave an example of a situation that is not very far away where, you know, Sukumar wants three software developers with XYZ skills and then the next day they land up at your doorstep because the process is already done.

All that needs to be done is deliver the people at your doorstep. Now, do you really think that something like this will become the core of future recruiting or a future talent system or is this still several years away because many of our listeners are recruiters, or talent acquisition specialists.

Some of them are really worried about what’s going to happen to their jobs? What is it that they need to do to prepare themselves to embrace this new change and continue to flourish in the talent acquisition space?

Sukumar Rajagopal: Yeah, I think this is a very interesting and important question, and the way I think about it is, this HBR wrote a very popular article about a year or so back. They called it everyday AI versus game changing AI, right? In Tiny Magiq, we work on game changing AI, but I’ll tell you what the perspective is. Let’s take the recruiting journey, as a recruiter would see it.

So, first they create the job description and publicize it through the various channels. Now creating the job description can be automated or made better. Okay, now resumes may start coming through the various channels. Now you have to scan the resumes and make a short list— that can be automated. Now from the shortlist you have to do the screening interview first to make sure you got the correct candidate that can be automated.

Then you present these selected candidates to your client. Then they’ll interview. That can be automated. Then you will make an offer letter to them that can be automated. Like that, every step of the recruiting chain can be automated. But if you did that, I believe you will be only in everyday AI.

Because there will be impedance gaps between each of these steps, because it’s possible. And I don’t know how the people are organized, but let’s say there is a screening person, there is a short listing person, there are various kinds of tasks. So the job will wait in between these inboxes and outboxes. This is what I see will happen. Instead of that, this is what I was proposing early on— take the universe of hiring requests you get, sort them into simple medium complex requests. For the simple ones, can I do the whole thing using generative AI? End to end, if I’m able to do that, let’s say you are hiring, recruiting people are able to hire, let’s say 1000 people a year. I’m just making it up. I don’t have any private knowledge of your enterprise. What if through this strategy they can hire 10, 000 people a year? You are happy, your clients are happy, everybody is happy, everybody makes more money. What is there not to like in this system? 

This is what we need. I believe more jobs, especially in the technology world, will get created more and more. And we have seen this in every big technology wave. Now the people that are not going to embrace this will lose their jobs. This is the part I want to mention. Like I often think about what happened to all the people that rode the horse driven carriages. Did they lose their job? Yeah, ones that didn’t learn driving, probably. 

In fact, I have seen in my own life experience, I met one person, this is a single data point, but you can keep this in your head. This is reasonably a senior citizen. This person was talking about what happened to that person 20 years back when computers first came. This person was a new typist and stenographer. Okay. But he found that he could not learn computers. He tried, but he could not. So he became a messenger boy for the company which he was working for. So this can happen. I’m not saying this will happen, but these are the kinds of things that can happen because we are not keeping pace with what is happening in the world.

So I want to be a technology optimist. I will say you will get more jobs, not less jobs. . But if you don’t use AI at all, you don’t accelerate what you do, some other recruiter who can use generative AI and do their job faster will get more business. And this is true of any technology. Why is this specific about generative AI? All the typists who stuck to their typewriters and didn’t move to word processors on the computers lost out. Because typewriting you have to type, or you made one mistake, you have to scratch the whole thing and redo. Right. Those kinds of problems will start to come with what I believe. 

Sat J: Now for the rapid fire segment …

Imagine a day without Gen AI in your work life, what would be your biggest challenge? 

Sukumar Rajagopal: … I don’t know how to answer because it’s like, I’m using Gen AI so much every day. It’s like being without a mobile phone. I can’t imagine how to lead that life. Although people talk about digital detox, maybe there’ll be a Gen AI detox that people will promote. I don’t know. But I am running a query at least every hour. I get some idea, I ask it, so I am not able to imagine life without them. I especially love chat GPT it’s just phenomenal. The GPT, once it reaches 4.0, it’s l dramatically good. 

So you can imagine a day without Gen AI in your life. 

Sukumar Rajagopal: I cannot.

What are your top three Gen AI tools or investments in your arsenal that you use? One you already named ChatGPT. Are there others? 

Sukumar Rajagopal: Yeah, chatGPT plus subscription I think is a mandatory thing that’s what I’m saying. Without that, life was very different. So I will give you an example. This may be a non-corporate example, but I will give it anyway, right? Last night, this was about midnight, and no one asked me why I am so weird like this at Midnight. I was like, okay, I want to plot that trajectory- this is some research I’m doing, so I wanna plot that trajectory of Venus. Okay, and you can Google this later, Venus for the trajectory when it’s painted, it looks like a flower. It’s called a pentagram, it’s like a flower. It’s a complex trajectory and it is beautiful. Okay. So I went to chatGPT and said, Hey, can you generate me a p5. js code, this is a tool I often use p5. js. which will generate this Venus trajectory. Five minutes later I have a code. I cut and pasted it into the p5.js editor. It starts by painting the Venus pentagram. 

So the level of astronomy that is needed to do the calculations to put the trajectory will take a lot of time. So this idea that it’s a smart intern— disagree with it. Although I’m not sure how to prove this, every time I see something like this, I am like, wow. I mean, the level of knowledge it possesses in every domain, every industry. I’ve done so many queries with it. It’s able to create the disruptive idea that the original disruptor came up with. I have tested 500, I have like a list of 1000, I have got up to 500 and in every one of those cases, it’s able to do that.

I believe only an Einstein at your desk type of thing can do this. I have not come across a smart intern that will generate a disruptive idea. If you have come across any, let me know.

Is there one thing where you felt disappointed that Gen AI didn’t give you the response you required or it didn’t, no matter how advanced Gen AI gets, there are some tasks it just cannot do.

Sukumar Rajagopal: Not yet. In fact, I have a running challenge with all people. I’ll give it to all your audience also. Come up with something that you know that ChatGPT doesn’t know. Okay. And it, I, it’s not a trick question. Oh, I don’t know what I ate for breakfast. I’m not talking about that In your hiring domain, hiring is a science, right? There is concept, there’s methodologies, all of that. Can you come up with one methodology or concept in the world of hiring that chatGPT doesn’t know you can reach out to me and send it to me— if you are able to do that, I will buy you lunch or dinner in your favorite most expensive restaurant in your city.

Sat J: Wonderful. We’ll, take the challenge.

Sukumar Rajagopal: Let’s see if your audience can win this challenge. So I don’t see any tasks. So the short answer to your question is I don’t know what task it can’t do. 

Sat J: Is Gen AI the most pivotal change that has happened in the IT industry ever?

Sukumar Rajagopal: Yeah, I agree. So how do I give a short answer to this? So there was no technology which could create new things. Second— reasoning at this scale was not possible. So to bring a disruptive innovation to market, I identified eight skills or eight competencies. The GPT four class models score in the top one to top 10 percent of humanity and all those skills. Based on the tests that I have done. I have not done scientific tests, but I have done enough queries to be able to confidently say it fits. And some researchers have already proven that it is in the top 1 percent in originality. Researchers have already done that. So, despite all these eight skills, I don’t know if anyone has formally done this study. So, this is unprecedented. 

Sat J: Have you come across any individual who was anti Gen AI, who completely debunked the myth that Gen AI can do so many transformative things and you were able to transform this person’s perception through your work. And if so, you know, it’ll be great to hear that anecdotal input from you. It could be a friend. It could be somebody in the organization, some client of yours.

Sukumar Rajagopal: Yeah. Okay. So broadly, I generally face skepticism on this strategy that I am promoting called the expertise augmented generation. Okay, people have prompt engineering and they have something called retrieval augmented generation, which is ingesting all the databases and local content you have and incorporating that into the application.So this whole expertise augmented generation and the expert lens approach and the knowledge engineering approaches that I work on usually receive skepticism from clients. And it goes away the moment I do the first demo. So I tell them, okay, you are skeptical about it. I get it. I understand. I open one of the agents and let it solve the problem in their industry in like five minutes. Now they say, Oh, sure, interesting. How did you do that? Now, even after that, some people will remain skeptical, which is okay. So people who are amenable to reason, they can change their mindset. So there are two mindsets I look for. One, think that like you were saying, you are already on that side of Einstein at your desk.

So think from that perspective. Is it really Einstein at your desk? Maybe not. That’s just provocative and some say it is a wild exaggeration to make it understand the power, right? So think from that perspective. And use other new strategies that are emerging. Think that’s what I would do. Skepticism, I would say, is rampant because there is just too much hype. And whenever there is too much hype, the skeptical people are also on overdrive. 

As we navigate the rapidly evolving landscape of talent acquisition, it’s clear that generative AI holds immense promise—but only if we know how to harness it effectively. From transforming routine hiring tasks to enabling strategic, high-impact decisions, AI is not just a tool, but a game-changer. But remember, success lies in pairing AI’s capabilities with human insight and expertise. The future of recruitment is not AI vs. human—it’s AI and human, working side by side to drive meaningful change. 

If you found value in this blog, we’d love to hear your thoughts! Leave us a review or share it with your network on social media. For even more in-depth discussions and captivating insights, be sure to tune into Hireside Chats available on all your favorite podcast platforms. Subscribe to Hireside Chats today!