Podcast: Phenom’s Mahe Bayireddi on Data, AI and Advanced Technology

Phenom Explain AI


Transcript

Mark:

Welcome to PeopleTech, the podcast of the HCM Technology Report. I’m Mark Feffer. My guest today is Mahe Bayireddi, the CEO of Phenom. They use AI in their talent acquisition platform of course, but they also talk a lot about data. Mahe says that’s because he’s seen firsthand the impact data can have on business outcomes, especially when it’s used with the right technology. So we’re going to talk about that and also about the growing impact of automation and where this is all leading on this edition of PeopleTech.

Hi, Mahe. Welcome. So Phenom, like pretty much everybody in the industry right now, is talking a lot about AI. And I’m wondering what’s distinctive about your approach compared to others in the space?

Mahe:

The thing is, we started the company with data as the center even in 2011 when we are starting. And there is a reason for it. My previous business is in a data company, in patient recruiting. And we have always thought about data on the AI will make a humongous difference for what we are really doing, and HR is lagging. So that’s what is the infrastructure, what we are building for years. And the most important thing is there are two types of AI which we constantly think about. The AI which is discriminative AI, which is how do you segment the data? And then what we are really looking at is a generative AI, which is what you can really create on top of it.

But the last couple of years, everybody is thinking about how human think and how human act. But right now, how humans actually become rational and how humans can rationally think about what they’re really doing, at the forefront’s what we are really looking at right now. Because of the foundation, what we have built in the last 10 years of how datasets has to work, how data engine has to work, we have the enterprise context really, really well-defined, so that the overall infrastructure will work in a much more accurate format comparatively. But nobody’s 100% on the track, but everybody can improve. But we have a very good foundation for what the data is. The second thing is how can we make sure it’s a responsible AI? Those two areas we are very, very strong.

Mark:

There’s so much buzz about AI now. It seems like every place you turn around, someone’s got a new AI-based product. And I’m wondering what is behind all of this talk? Is it users expressing their needs? Or is it solutions providers hyping their products? Is it a mix? Or what’s the dynamic of all that talk?

Mahe:

So if you really look at right now, the most buzz is towards generative AI. So one of the most important thing in generative AI is it’s a non-algorithmic compute. So what that means is there are no algorithms you need to build. They’re built by the foundation infrastructure, like OpenAI or Cohere or Midjourney. All you’re really doing is you’re calling through an API, which is in the cloud infrastructure to really consume it. What is the most complex thing in this is the infrastructure, what you need, how you really built a data engine, how you built intelligent applications, how you built knowledge graphs. And how do you make the context really friendly is the most critical element, which is not hyped. It’s actually the foundation in the last 10 years getting matured over a period of time for many, many companies, like Palantir is one example. Or if you really look at companies like Scale is another example. There are a bunch of companies which actually really built the foundation. In HR, we build a foundation which is extensively connected on that particular front.

The hype is about generative AI, which is about it’s easier to really utilize it, and everybody is talking about it. But if you really don’t do your predictive AI and discriminative AI really well, your generative AI working is complex, is how we really look at the world of HR.

Mark:

Now, you launched in April X+, and you also added a number of AI features to your platform. What’s the reaction been to those? And has it been along the lines you anticipated?

Mahe:

Yeah, so when we really look at the overall infrastructure of whatever the AI we want to deploy, there are multiple angles we really look at this. One is how do we deliver an experience which is relevant to the people? That’s the most important thread. And when we say people, there are two sides of the equation. Side one is talent, which is either a candidate or an employee. And the other equation is in an enterprise. It’s about your team, your managers, your HR, your recruiters, your recruiting coordinators or HRIS people, in which particular framework we can make the maximum impact. And we look at the most important thread when we are using AI as AI should be invisible in the experience. The minute it is visible, it’s not AI. So the way we really think about it is which experience we are making, what kind of difference and how that can really impact the end user in effective format.

We actually look at experiences in three different formats. Format. One is can we empower people who are using, especially the talent, the candidates and the employees? Then we look at productive experiences. How can we make recruiters more productive, managers more productive, talent managers more productive or HR more productive? Then we really look at operational experiences. How can HRIS be more effortless and engaging and talent ops can be much more easier to use the infrastructure what we build? In those use cases, where is the AI really going to come in and where it can help?

There are two ways to really look at. One is where can we personalize, and what pieces we can automate? And that’s the framework what we constantly reference. It’s about the personas. It’s about the experiences we want to impact. And within the experiences, where can we deploy the personalization using intelligence? And where can we bring automation so that we can bring productivity is how we look at the overall framework.

Mark:

It seems like a lot of this AI based or otherwise… But the use of technology, like Phenom’s, it seems like it’s based on certain data or a certain task, and you’ve got two sets of people looking at it from opposite directions, the employer or talent professionals and the candidates themselves. When you’re building an application, does that work to your advantage, where you can leverage different capabilities for each audience? Or are you having to almost build two paths?

Mahe:

So the thing is, in the most important infrastructure of data-driven applications, each particular side is a feedback loop to the other and which are not naturally understood in HR for a very long time. What does that mean? A candidate getting hired in a job, and a recruiter’s activity and a manager’s activity, is a direct correlation of how do you improve the experience of the candidate? So which particular piece of the recruiters actually really like a particular candidate in a particular division or how a manager really hires, based on that feedback loop, you can improve the candidate experience.

Let’s go further more. How an employee actually grows within a company, and retains within a company, has a direct correlation to what candidate experience you can really get and which person is a perfect fit. So you can use all these data points in the FIT score based on how the candidate is engaged, how the recruiter is actually really moving a person from one spot to another and how the managers and the interviewers are actually considering people and then how an employee is growing within a company. All the data can really feed into the candidate experience.

For the last 30 years, that’s not how we actually did. But after we really came into existence, we basically said, “The candidate’s experience can be improved, not just by watching what the candidate is doing, but what is happening in the backend.” And we call this as human-centered system thinking. And that’s what we primarily focus on, how to think about people, how to think about these personas, where can we really deploy AI, where can we deploy intelligence and where can we deploy automation is what is the most important infrastructure what we’ve built.

Mark:

One of the things that a lot of solutions providers are saying is that by using generative AI, or sometimes just more broadly AI, the users are freed up to spend more time on “strategic tasks.” They get rid of a lot of the mundane chores. Honestly, that always seems like kind of a squishy thing to me. And I’m wondering, do you see real potential for that? I mean, it seems like it’s an easy thing to say, but is this technology-driven revision to the workflow really going to make that much of a difference that people will notice?

Mahe:

So the thing is right now in the country we have… Especially in US and the western world, we have a most complex problem going on, which is the demographics problem. We have unemployment at 3.5% in US, 4.8% in Europe. And that is not going to come down, because we have demographically and birth rates are really, really low. So the future of the work is going to be two levels, robotics and copilots. That will really do a bulk of demographics problems. What we have can be really resolved. And how do you do it? So we look at the overall work in five different zones, zone one, two, three, four, and five. And these zones are based on the preparation, what you have to really build your working career, so how much preparation you need to do the work.

So zone one and two is more like a frontline worker. Zone four and five is more like a knowledge worker. In frontline worker use cases, robotics are the most important players, which are really making a humongous difference. I’ll give you a couple of examples. A company like DHL or Kuehne + Nagel, who are in logistics businesses, which are some of our customers, they basically use picker robots in their warehouses. What’s a picker robot? It actually picks whatever the infrastructure, like boxes or whatever it is, and really places it in another spots.

Now, how is the industry changing? A pick a robot, one robot, can do four people’s work, four human beings work. Now, a manager used to have 15 people reporting to them. Right now, they have three robots reporting as resources and three human beings. So now, you have six resources which can handle 15 human beings work. And that is how all the fulfillment centers are really running. Whether it is Amazon, it’s UPS, it doesn’t matter, because the markets are so tight in terms of, what do you call as, jobs to people. And the workforce is not available at all. So robotics are changing demographic issues in a very, very different format, but not every job. There are only certain jobs it actually fits. Now the manager’s job actually have changed dramatically. Now, that is a frontline worker use case.

Now you switch back to knowledge worker. Take a developer or a programmer who is actually really writing code. Today, using copilot, you can optimize 20 to 30% of your coding in a much more effective format than where you used to be. So what does that means is your developer productivity is different, so the managers has to really look at productivity from a different angle. So now each particular zone, things are really working differently, so the skillsets right now, what people need are entirely differently than what they used to. And these kinds of changes are drastic and different.

But will this really replace human beings completely? No, you still need a lot of humans. Right now, we have a problems with availability of human beings, so we are using robotics and copilots in extensive format. And because of that, the way you hire, the way you retain, the way you grow people will also will change based on the needs of the market. And that is what is really shaping up, right now, the whole HR industry.

Mark:

When people talk about this, it sometimes seems that they anticipate the changes coming very quickly. In a year or two, the whole approach to work will be different. Do you think it’s really happening that quickly? Or do you think what you just described is more of an evolution that’s going to take decades, or a decade or something like that?

Mahe:

It is happening in pockets, in a very fast format. In pockets, it’s very slow. So what does that mean? If you really look at digital artists, or anybody who’s doing multimedia side, things are really happening rapidly. If you’re looking at programming side, things are happening rapidly. In financial analyst side, things are really at a different pace than we have ever seen. That is on knowledge worker side.

On the frontline worker side, there are pockets where things are really rapid, but nothing is actually changing on a productivity tone of every job family at the right spot. What that means is there is a reason why productivity cannot be increased in one single shot across every industry. The reason is the talent operational metrics are not really well calculated for most of the jobs. So what does that mean? The data of people data, the data which is with respect to financial data, customer data and worker activity data has to be merged together. And that merger is not happening at the pace where we are really thinking, because otherwise, the data is very complicated to pull out.

Because of that productivity is not really increasing on every section of workforce, but there are certain job families it is much more rapid. In certain families, it’s not happening as fast as it should be or it can be. And there are certain jobs which it cannot really change in a much more dramatically. And that’s also true. So it’s job family by job family. And task by task, it is actually orienting differently.

But this doesn’t mean people who are working will have less time to work and more free time. That’s not happening at all. You are getting newer job, or newer tasks, than what you used to. So your training becomes harder and harder, and your adoption is much more aggressive than ever. So in terms of talent, their time they spend to really do a particular task is not reducing. It’s actually increasing the number of hours you have to put in, because you have to learn so much than we’re used to.

Mark:

You launched, this year, a new multi-tiered support model called the Phenom Service Experience. Can you tell me about that and, first of all, why you took that course and also what the customer reaction has been?

Mahe:

Yep. So there is one fundamental thing which we also always think about, like what is the customer passion, what we can exhibit? In the current markets, there are so many things really changing. There are industries which are doing good. There are industries which are doing bad. And in the last three years, almost every company’s 30 to 50% of the HR teams have really left their jobs or moved on or were let go. Different things have happened. So what we started realizing was there is a service infrastructure, as a vendor, we have to provide to make sure in these kinds of times we can help our customers, because we understand their business in a very effective format. So we want to scale that so that we can really go to where their current situation is and help them with our tooling and our platform in a much more effective format.

So that’s the reason why we brought in Phenom Services as a business model, where we can really create a lot more impact. It’s not only that. When we really look at… Previously, the question you asked Mark is about Phenom X+. We brought Phenom X+ to bring productivity in the recruiter segment and manager segment, and also in terms of what we can do for talent managers, which is talent management and HR. In that, what we also observed is can we use X+ to help the service to be really delivered in a much more effective format, either in support or onboarding or basically really creating customer value? All of those things we are really observing, and that can make a difference for our customers at the end of the day. That’s what is the primary focus.

Mark:

So where are you going with all this? What’s on your roadmap in the immediate future? But in the longer term, what is Phenoms plan, and what are your long-term goals?

Mahe:

So what we believe in is the most fundamental thing is, in HR, the weakest link is experience. And the experience can really get better in multiple formats. Our thought process is dynamic personalized workflows is the future of HR, and you really do this dynamic personalized workflows using automation intelligence baked into the experience of the overall talent ecosystem. And if you can do it by each particular job zone, each particular vertical and each particular job family and look at different locations act differently by how the supply demand of talent actually exists. If you can deploy this infrastructure, it’s a humongous business to scale. Because the problems by each industry has really fabricated differently, and HR family has been really turned into something else.

So in this particular use case, how we use skills, both hard skills and soft skills in really recruiting people and retaining people and growing people, but making it super dynamic so that the workflows are adaptable for each use case. Let me give you an example. If you’re hiring a frontline use case, 90% you can automate. Only 10%, you need a human in the loop. If you want to retain in frontline, you can automate like 75%. In a knowledge worker, you can only automate certain jobs, like 30 to 50%, but you can make much more intelligent inputs with the people who are making decisions so that they can take action which is meaningful.

So based on that, we are really building these models for each particular company and each particular industry in a much more effective format, and we believe that’s where the future of HR is actually going.

Mark:

So what keeps you up at night? When you think about the market, you think about technology, you think about talent acquisition and business, what are you worrying about in terms of… Well, what are you worried you’re not thinking of? Or what do you think is going to turn around and hit you?

Mahe:

See, there are controllable inputs and then uncontrollable inputs. The controllable inputs is what we need to control, and that is where the culture has to be intact within a company so that we can really, at the end of the day, support our customers in a much more effective format and really generate value out of our platform for what they expect. And we always really look at how we, as a company, can really orient towards the culture of customer passion and really looking at building our purpose, which is about helping a billion people find the right work. And that’s the only thing which constantly really make me think about how can we shape the direction of our company and the industry and our customers so that we can really build the company with purpose?

Mark:

Mahe, thank you again for being here and talking with me. It was great to talk to you, and I appreciate it.

Mahe:

Thank you, Mark. It’s lovely to talk to you. You’re in Philly. Allow to me to … down the line.

Mark:

My guest today has been Mahe Bayireddi, the CEO of Phenom. And this has been PeopleTech, the podcast of the HCM Technology Report. We’re a publication of RecruitingDaily. We’re also a part of Evergreen Podcasts. To see all of their programs, visit www.evergreenpodcast.com. And to keep up with HR technology, visit the HCM technology report every day. We’re the most trusted source of news in the HR tech industry. Find us at www.hcmtechnologyreport.com. I’m Mark Feffer.

Image: Phenom

link