Karthik Viswanathan edited audio for Fireside
===

[00:00:00] All right. Kartik, welcome to the show. Thanks for coming on, man. Good to see you again. Hey, Jeff. Pleasure to be here. Thanks for having me. Yeah, it's been a little bit, we had dinner in San Francisco back when Raj talked at that dinner in San Francisco. Yep. I haven't really had a chance to see you since.

You have a good holidays. Holidays were great. A lot of family time. Spent a lot of time in the Bay Area. Yeah. Stoked to have you all know, I mean, background kinda speaks for itself, but you founded RO and your CPTO over there, chief Digital Product Officer at Optum. Before that, years at at t Macy's, on and on.

I wanna get in and talk about this kind of idea of the chaos tax that happens due to the just sprawling systems everyone has at work. But can you just give us the insight into what brought you into product? I can tell you honestly, I didn't start out wanting to be in product. Uh, no one does. That is true, right?

So early in my career, I was building and scaling systems that worked from a technology [00:01:00] standpoint, but then when they hit the real world, it created a lot of friction, either because of adoption challenges or just because we never kept the consumer empathy in mind. And he created the technology. And somehow we just allowed bad code to be written on top of bad code, and users found a workaround and leaders kept asking why adoption was low, even though in theory the solution was correct.

Yeah. And that disconnect bothered me a lot. And right around the same time, I think product became a discipline where I could honestly sit at the intersection of technology, business strategy, and human behavior. Right. It was one of those unique roles where. I could think about influencing the why, the what we built and why it mattered to the larger ecosystem and how it can fit into somebody's day.

And then as I moved into larger organizations, I just saw how dangerous it can be when product decisions are made in isolation. 'cause as you get into larger organizations, not only does the product portfolio get a lot more complicated, but the number of business units making decisions gets [00:02:00] extremely convoluted as well.

So systems get bloated, workflows fracture. People end up compensating for just really bad design, but sheer effort, some of which I know Log Rocket does a really good job of really figuring out how to kind of surface for its users as well. But the chaos is real. My effort in product, and by the way, I still feel like I'm a student, is to really reduce the complexity at the source.

Ideally, we start off with the happy path in mind and then allow the right diversions and frontage roads, as I call it, to kind of take shape so we're not starting with complexity as a starting point. And that mindset is what ultimately led me to Teleo. I didn't really want to build another recruiting capability or ecosystem.

I wanna build a true operating system that not only respected human judgment, but also removed noise, clutter, blo, and made good decisions, easier to follow. And, and through and honestly, product has been the only discipline that has allowed me to do that at scale. One thing you have, I think really put a [00:03:00] nice point on, well you call it the chaos tax.

I've heard it called a bunch of other things, but like even this has come up when I've talked to product leaders and product teams across the country. Everyone has 5, 10, 15 different tools that they're going to pull different insights and different stuff from. And sure, best of breed's great in that you have the best thing that does every point, but it doesn't really help if you can't actually use every one of those to the peak ability.

Right. That's right. Let's dive in here. Like where'd this come from in your head? Like where'd you pull this from and what does it actually mean? I think the chaos tax was just an appreciation of the sheer amount of inefficiency that was there in the HR tech market as I did my user research and really figured out how to best define and also.

Assess market to product fit, and for the last decade, I honestly believe this unbundling that we've been trying to do and get into a level of specialization and depth, I think that's all well and good on paper. But I don't know in the real [00:04:00] world if it really came to play out the way I think people designed it.

Ultimately, we left it up to the humans to kind of become that integration layer or that arbitration layer to go figure out where do they go for what, to kind of get the best of all things they need. Because what we see is recruiters today aren't efficient or inhumane in the way they interact with candidates or clients because they lack tools.

There's actually a plethora of them. If you actually were to look at. The HR tech ecosystem, you would easily find like 300 different companies or tool sets that are out there. The problem is they're inefficient because they're juggling nine to 10 of them on a daily basis, and not to mention the fact that they're spending upwards of 400, $500 a month in subscription fees, although they're using.

Possibly 10, 20% of the overall capability set. Not a platform, but a, a tool or a system offers, right? So every context switch, every be it manual, copy paste, every reconciliation that they have to do is what leads to that inefficiency that I call the [00:05:00] chaos tax. It's like when you go to a happy hour or when you actually make a good barrel of bourbon, it's a devil's cut, right?

Like it always seems to just evaporate and then we wonder where that invisible time went and we just ends up crushing productivity and overall decision quality. Yeah, sure, you can have the best set of things, but. If you're blowing 40 to 60% of your time switching between platforms and trying to like figure out where it all goes together, you're not getting the best out of all of us.

You can have the best, it's like having 20 Ferraris and you're trying to commute to work, but you're saying I gotta, you know, drive a block and a half in each Ferrari. I think that's exactly right. And the reason I actually picked the bourbon example is where I maybe slightly disagree with you is when I've spoken to these recruiters, they actually reaffirm that wastage or that inefficiency as part of the process.

But yeah, McCarthy, that's. That's inefficiency you have to deal with if you wanna be an amazing recruiter. Well, that's not true. That is fundamentally not true. Why? I'm curious, like where is that coming from, do you think? The, like, because like the [00:06:00] time does make the bourbon good, we know that, but like, I think this is the only group of people I've ever heard of who say like, oh no, it's a benefit, not a bug, that I have all this wasted switching time.

Yeah, because I think that's the, the so-called luxury tax we put on the best of breed, quote unquote. Systems that are out there. Right? But the challenge is these so-called best of breed tags become a liability. The minute those platforms or the ability to bring insights out requires an explanation or requires a user manual.

I mean, the philosophy I've always thought through is, and this is something a mentor of mine told me when I was designing, you know, UI and workflows for at and t is like when consumers are able to buy a car in Tesla in three or four clicks on a website, you're telling me you can't get them to buy a phone.

But when you talk to these large organizations, but this is how we do work. I mean, I don't think you understand. Karthik Healthcare is so complicated. I'm like, is it though? You just have to get people to log into a digital ecosystem, show them their profile and their membership details. Get them their labs, get [00:07:00] them access to their medication, get them to nurse when they need it.

That's it. Why over complicate it. And I think that's what we are trying to fundamentally change. How do we take the liability out of these so-called best of breed systems? The other thing is, and specifically for HR Tech, when you have multiple systems, give you multiple insights, because they're all sitting on different data sets, right?

So when your a TS says one thing and your CRM says another, your sourcing tools course completely differently. Each one of them has a chat bot because that's the cool thing to do these days. Then analytics is sitting in Google or somewhere else. You only have noise. You don't have intelligence. By the time you make sense of it, either you're making a bad decision or you're justifying your bad decision because you're running out of time.

And that's not a human failure. That's a structural failure. 'cause I think humans are inherently smart. What is your bigger view on solution here? Do we need to glow things up a little bit and kind of restart? Or can all of this be salvaged? I mean, clearly on the HR tech side, your answer was Go start Toro and.

[00:08:00] Bring a fully capable platform, but it's far from the only functional area suffering from this that's in itself a a one hour podcast discussion, I guess, and I by no means qualify as a futurist, but what I will tell you is having done a few different transformations, I think it's an answer that's very personal to each organization.

You know, mid-size companies, startups are companies that consider themselves fairly lean and small. I think for them to leverage everything that AI has to offer, be it from a technical. Or technological prowess all the way through marketing or content creation prowess. To really transform the way they do workflows, I think is a lot easier, and it's something that is, they just, they just have to do it.

There's no other way. I think it gets a lot more complicated when it comes to large organizations because their concept of success is vastly different because they have to think through what their boat is gonna say, what the dividend is gonna look like. Alongside what they need to do to kind of stay competitive over the next 20, [00:09:00] 30 years.

Right. But it's one of those where misery loves company. Unless one of their core competition transforms overnight and does business a different way, they have no reason to kind of pivot drastically and kind of. Change the way they've always been doing business. So that, I guess mediocrity, if I can call it, that also puts a level of inertia in terms of what transformation looks like for a larger organization.

But there are a bunch of different organizations that are really leaning into experimentation, doing a lot of pilots. But the one thing that's common between everything is that you start with the end customer or consumer in mind and work backwards and figure out which technology, solution or stack is the fastest way to get you there.

It's gonna be really, really hard for a company, especially a startup, to do each of those eight things nearly as well. Mm-hmm. To look at tel, how did you and the team look at building something that is great at each of these many things at the same time, you don't have this just like horrendous UI of over complication and hard to actually find [00:10:00] these things or use them or all of that.

It's like that's a balance that takes a lot of work to get right. Potentially. And I think Taleo had the luxury of starting from a blank slate. We could stand on the shoulders of people who have kind of gone ahead of us and kind of figure out what we can do right. And what we should in essence, mimic or learn from in certain cases.

Right. Even for us, because we use a combination of both public and private models. I think creation of a, what we call a semantic layer was very important, and this semantic layer is what gives us the clarity from a business logic standpoint to figure out what the context is and why, and what is relevant for every decisions you're making, which then allows our agentic ais to make a decision of what is the next best action for.

Either the system in an autonomous scenario take or provide a recommendation to a user if they're kind of doing it manually. And I think the same thing could be used for enterprises that are not starting from scratch. And a lot of the companies that specialize in [00:11:00] agentic AI right now are going down that path.

So rather than transforming broad systems and starting years of work, again, what they're saying is, well, give us the data. We'll slap in agentic AI or an avatar on top of it, and now that becomes the main medium that you do work with. So there are multiple ways to skin the cat, I guess. But at least for rum, it was the use of doing.

Clean semantic layer that allows us to kind of thread the entire platform and ecosystem to learning from and avoiding mistakes from people who have kind of done this in the past. The third thing is, I think Tiller doesn't promise to do everything that every other system does. What we have taken an approach on is to say, let's focus on the 20% that drives the 80% value for agencies, and let's do that in the most simplistic and the most expedient way possible.

For an example, within an A TS, if the top three things people use in a TS four is to track their jobs, track their candidates, [00:12:00] and be able to communicate with them well, we have that. Do we have 10,000 templates sitting out there where they can send to candidates or vendors? No, but if we have an AI that can write and email to your specifications when you need.

So there are ways to kind of get the same level of functionality, but doing it in a more simplistic manner and in a way that people understand and use in other places, be it Amazon or be it in an Apple store, or be it in the tesla.com website. You kinda described maybe the a s layer there. What does that look like if you're using one of the other ones out there?

Like what does a recruiter's life look like when they're going through that versus in something like ERO, where it's kinda all just like magically together. So I think if you'll take any of the larger, more established or incumbent, athe is out there. I think you'll see a plethora of choices, at least a dozen tabs, some semblance of ai, but largely to improve writing or give you some high level suggestions or simplify or summarize this.

It [00:13:00] isn't anything deep or insightful. What we have taken an approach within our web application is that any of those. Core workflows either needs to be automated, so saving you time, or if you're doing it manually, needs to be completed within three steps or less. Most, if not all of our tasks that you do within the mobile app will get done within 45 seconds because we honestly believe if we can give back one gift.

To anybody using ro. It's a power of time. Again, we are nowhere in the world of replacing humans as much as allowing them to feel more empowered and actually feeling human at log rocket. We take a very similar look to what you're describing there. Obviously very different platform sets. But the core thing people were trying to do here for us is you just want to understand how people are using your digital product and where they're running into problems.

What's working? You know, I made a change. Is that bad or good? You just want answers. And a lot of competition has been maybe building out, like you said, the ubiquitous chat bot, right? So good. You have a [00:14:00] chat that can help you build charts. It doesn't save you that much times. You have to fiddle with the chart, make it right, like you just have to know what you're looking for.

The idea here has been can we connect all these disparate points and bring in the data? We just ship you the answer. Like your question is always going to be, what do I need to think about? What's important, what's salient? What do I need to do to improve the experience? If we can just ship that to you magically, or have it show up what seems like magically, you're gonna be very happy with that.

Mm-hmm. And as you brought up, that used to be a thing that you had to go look at multiple different systems. Ask the question, try and use some searches. Build a query. Build a chart. But we can just send you the summary and the thesis of like, here's what's important. Is why it is. Here's the evidence, we wanna go check it out.

To your point, in HR Tech, they don't wanna go through 12 different tabs to look at analytics and who do I need to send email to and which I send. Where are my templates? Is just, I wanna do the right thing right now. I couldn't agree more. And having seen some of the log rocket capabilities, I completely agree.

I think that's really what companies like Log Rocket should be doing and investing their time in. Aside from the monetization side of it, that's the right thing to do to go help people. [00:15:00] Right? So I completely agree and kudos to the team for working on it. But I think broadly, if I had the proverbial microphone, I would say I really wanted leaders to stop asking what can AI replace?

And as leaders, we should start asking. What human potential are we wasting today? And I, you know, where can we feel like we may need more ai, at least in recruiting to make people feel more human? Right? And that sounds counterintuitive. I get it. But I honestly feel like I alluded before when somebody says, man, I applied to this job and that recruiter, God damn it, they won't even get back to me.

Well, recruiters aren't ghosting candidates because they don't care or because they, you know, it's one of those, um. Now the movie Jerk, I dunno if you've seen it, but like, you know, the guy just flips through a yellow pages and he puts on Navin Johnson and he's like, this guy seems like a jerk. No, it's not that at all.

They're coasting people because they're drowning. And when you're spending 60% of your day in like just admin screening and everybody looks the same, everybody's doing the same. AI resumes quick, whatever [00:16:00] it may be, your emotional bandwidth. As a recruiter disappears. Yeah. And when that goes away, you stop acting logically.

It's one of those, when you try to be a monster, you become one. It's one of those things, right? We talked about this at dinner like a couple months back. Humans became the integration layer. Yeah. And so when we built all these disparate things where you had all this kind of lack of efficiency, let's call it, where things didn't kind of go together.

The glue became just human effort. And that was great. Like in 2021 when we were all raising enormous rounds with no dilution and money was basically free. Sure. Like go hire a bunch of people and and just throw people at the problem. Yeah. But that's never gonna work long term. It's, it's not good. Like, it's not a good outcome, I would argue.

And so like, it is kind of, I think, you know, a little bit of self contradiction here is interesting where like more ai. Can potentially help make us all a little bit more human in our jobs. But I think just onto that point, I do also believe in this notion of decision budget, right? As humans, we are only equipped to make three or five really good decisions.[00:17:00] 

One of those decisions should not be, how do I copy paste from Excel to this CSV file? It just cannot be that. Right? And quite honestly, just to put a, a bow onto like, I honestly think. RO is considered a success in its generation, so to speak. It won't be because we automated hiring. I mean, great, we'll take that as a means to an end, but that's not what success looks like for us.

It is if we are able to give back people their time, their ability to kind of make the right judgment, and quite honestly, the dignity back. If we can do all of that, I think aside from what money we end up making or don't end up making, that would be the true measure of success for us. And I say this because I feel, at least in our conversations, log Rocket is the same way.

You're really trying to help people do their work better and amplify their effort in the, in the right perspective. I think about, we have a, we recruited here who are working with now, and I have talked to so many candidates because we're hiring a bunch of roles right now. So I've been spending my time on recruiting a lot.

And every single person I've talked to, I'm like, Hey, you seem like you're [00:18:00] doing excellent. Your current role, your company's blowing up. Like, why are you talking with us? Their answers repeatedly. This person was so amazing in her outreach and how she talked with me about this and how she explained everything.

I was compelled to continue with the process. I'm really interested, and if we could get. All the BS outta her way. That would be so incredibly powerful. And that's true on the product side too. Like you have great product managers who are just so disillusioned with the manual repetition of a BS they have to do.

If you could imagine like all these people, if you could just unlock the humanness of what they do. Automate the crappy parts. That's gonna be fantastic. And that seems to be like the thesis of what you're looking at here. You just applied it to HR tech. One of the things that I learned, I think it was actually my wife who told me this, which is actually pretty poignant on her part.

When you're constructing teams, you always need a mix of people who are thinkers, doors and talkers. Right. And I don't mean talkers as in like gossip mongers, but people who are building trust and communicating updates, so on and so forth, right? And if you take that same viewpoint to a singular human, we should be spending our time [00:19:00] thinking.

Doing and speaking or talking, right? And for each role, in this case, for this person who's doing outreach, maybe they need to spend more time talking or thinking versus actually doing, which is what a thing you're alluding to. But I think having AI provide that flexibility in terms of what you need to do, at what point in time of your day or your journey within a week.

I think it's extremely important because going back to the word amplification, that way automation isn't replacing anything. It is just helping you amplify your day and amplify the ROI and amplify the outcomes that everybody is trying to go towards. Right. And that just, I guess in my definition is what fulfillment looks like.

Yeah, right. You are actually happy about what you're doing at any role you're part of, or any company you're part of. And that in itself should increase everything from ENPS course to job satisfaction for people. I do think this concept of fulfillment has a lot to do with the kind of work we're doing.

And so if we can help people focus on the stuff that really floats their boat there, they can do a lot more and be [00:20:00] happier doing it, which is a great combo. Correct. I think that's really what we have to focus on, right? Rather than fearmongering and really kind of getting people worked up about, and it's hard in given the macroeconomic conditions we are in, and there are in reality a lot of people looking for roles, people losing jobs.

I, I don't want to trivialize that. Yeah, but at the same time, fear-mongering and saying AI is, you know, the one evil that we have to all fight against is also not portraying technology in its most useful light. And that's what we ought to figure out. How can we use this to amplify fulfillment, amplify outcomes versus amplifying fear.

To bring a full circle. Right? Like you said, recruiters aren't ghosting people 'cause they're jerks. Well, you know, maybe a few of 'em are, but for the most part they're not to, you know, to be malicious. Yeah. They're just overburdened and overworked and don't have the time. And if you can take some of the things that aren't moving towards that, like making sure candidates have a great experience, you can set them up to have more human interaction, more humanists, which is.

There's one more [00:21:00] thing I did want to talk to you about real quick while I've got you here. I do always find it really interesting seeing people who have spent a ton of career at really, really large companies. Like at t is, I don't think it's argument to be made that it's anything but huge. Macy's, Optum, and then you didn't just go to a startup, you founded a company, so you went down to like the absolute smallest thing you can do at the beginning of it.

How has that gone? What's been beneficial? Like what stuff have you learned and, and brought with you that was really, really positive and maybe where did you have some hangups or get caught some things that were really different from that experience of, you know, the previous years? It's one of those where if you have heard, uh, Steve Jobs talk about how you look back at the past and like just these dots just magically connect and there's a story that always forms, although you don't understand it, I think the timing is a very important factor in terms of why and when the move happened.

Again, not sure if everything Sam Altman says is gonna come to fruition, but I do agree with him that I think the next meaningful organizations, [00:22:00] regardless of if they're dragons or unicorns, doesn't really matter. They're gonna be small in size and extremely nimble because I think the one aspect of AI that we haven't spoken about is it really empowers people to have superpowers, right?

Like you don't have to be a technologist to build an app and monetize your idea. You don't have to be a recruiter to go help people recruit and make money out of it. You can just use RO or any other tools that you wish to choose. So I think that part of technology I think is important. And if I felt like if I didn't jump in and build something that is meaningful from scratch, it'll just be a waste of my experience and the time I have left on this earth.

Right? One of the big things I wanna do for the larger social good is leave this place better than I found it. And this is, I guess, one. Or one conduit to go do that. But some of the lessons I think that have remained with me is you realize very quickly that titles that you carry don't travel with you.

When you leave an organization, the only thing that does is the accrued knowledge and the clarity that you gain. Having those [00:23:00] stumbles and failures across the number of years you worked in larger organization, that is the only thing you're left with, right? In a large enterprise, you really inherit your credibility.

You walk in your bu, your boss, your team. Empower with, you have like auto magical credibility as you go into any room or any conference, but at the same time, as a founder of a startup, you have to earn it every single day. People have very short memories. I still remember there was a, a conference that I had spoken without naming them multiple times, and I said, Hey, is there any way I could come and pitch my idea?

They're like, they didn't gimme the time of day, they didn't return my email, and this person knew me. Right. Is there something you have to earn that respect and earn that credibility in our daily basis? But honestly, that has forced me to really find and own my authentic self and voice, right? Mm-hmm. Not as an executive, but truly as a builder who has very, very deep conviction on what.

Where he invests time and why he does it, it has been uncomfortable. Um, someone told me that 90% of days as a startup [00:24:00] founder or working in startups are very dark and bleak, and that is absolutely true. There are some days that are extremely awesome and they get better and better as you grow and you start to scale.

But you know what? It was necessary, and it's important to kind of bring the rigor and the discipline that enterprises teach you, but find a way to use. The agility and nimbleness that I do think that speed really becomes a differentiator as you're working through this AI revolution as a startup or an enterprise for that matter.

So I think I've carried a lot of those learnings, be it on the architectural side, how to get data models right? We spoke about semantic layers. All of that is great. I by no means see RO or starting a startup as starting from the bottoms up. It really is an extension of what I've done in the enterprise world and really figuring out ways to move even faster.

Because in enterprises sometimes you are unable to move fast just because you have to get that buy-in. 'cause as you scale, the buy-in has to scale as well. Right. So it's been a fantastic journey. Um, looking forward to how 2026 unfolds. I [00:25:00] remember joking with peers of mine at a prior company, you know, it doesn't feel like I'm doing work.

It feels like I talk about work and it feels like I plan work, and then other people go do it. That's what I often like about here at Log Rocket. We're about a hundred people. I still write, I still create, you know, and it's nice to have that mix of managing fairly large team and contributing, but when you start a company at the beginning, there's no one else you're thinking and you're doing.

How has that experience been? Has it been rewarding and fulfilling? Has it been great to get back to it? Or is it tough sometimes to turn off the, how do you coordinate a hundred people or whatever and get back to, I just need to do this two things today and they need to be done 'cause no one else is gonna do it.

I think there's definitely unlearning a lot of the things that. You know, you have accrued over the years working in, in an enterprise organization. I think the one thing that's really helped me is as you climb it starts to get really lonely. There isn't anybody who's telling you, well, Karthik, these are the five things you gotta go do today.

The organization tells you, well, this is where we directionally want to be at the end of next year in a couple of years. You go figure it out. [00:26:00] So I think there is definitely a ference of skills, but yes, there is absolutely relearning of skills that you have to do. But you know, that's what a revolution really forces upon you.

Right? And it's very hard to talk about AI as an executive unless you've been in the trenches and really figured out how vibe coding works and how prompt engineering works, and what are the challenges with bias. All of those things. And so regardless of how 2026 unfolds, I feel I am a better student. I ignore close to being enlightened, but better student than I was in 2024.

And all things ai, all things technology and the where the future of this ecosystem goes through. And the other part about talent and the reason that drew me is there may be retail companies and there may be telco companies. But everybody's a talent company. Everybody needs talent. Log rocket does, apple does, Walmart does.

If you have two people in the organization, if you have one person in the organization, you need talent and technology is a team sport, regardless of what people say. You can use ai, but you still need people to kind of code it. Do prompt engineering use it, so it's [00:27:00] still a team sport. So being able to understand how talent works and how to really transform talent in this age of massive shifts in change and disruption, there's no better way than to kind of be with a company like RO and the HR tech space.

Makes sense. Cool. Well, appreciate you coming. I mean, this is super interesting. Gonna go through the idea of like, where does some bundling come from? The chaos tax that lies on top of it, but also just like how the solution is gonna make us all, uh, a little bit more human, hopefully. So yeah, thanks for coming on.

Appreciate you joining us. If people wanna ask more questions or, or reach out to you, like you said, it can be a little lonely at the top. So like people just want to connect. Is LinkedIn a good spot or is there a, a better way? LinkedIn is a great spot. I'm on all social channels, although I've kind of neglected my followership a little bit in the last couple of years, but I'll definitely happily send my email and my social handles in case any of your audiences are interested.

Awesome. Well, great to have you on. Anyone who wants to chat humanizing roles, Kartik is a, a great resource. But yeah, I hope to see you again in San Francisco. We'll be out there in a couple months, so hopefully See ya. [00:28:00] See ya soon. Thanks man, for coming on. Have a good rest of your day. Of course. Thanks for having me and have a great start of the year again.