What Do You Call the Talent Ecosystem in the Age of GenAI? The SBO.

A Timely Conversation with EPAM’s Chief Learning Scientist

Not long ago, Sandra Loughlin, PhD, EPAM’s Chief Learning Scientist, wrote: “I wish I could go back and redo this interview [about the talent ecosystem] from the lens of GenAI. So many of the ideas hold true—and we are leveraging the things I mentioned to support our own AI journey—but GenAI has added some new facets to the approach.” In the Q&A below, Loughlin’s wish becomes our command. Read on, as Loughlin speaks about the state of the contemporary ecosystem. It’s an education!

We’ve lately heard a lot about the idea of the skills-based organization… which sounds like the integrated talent ecosystem you’ve been talking about all these years. How are they related?

They’re the same thing—what we initially called the “integrated talent ecosystem” is what other people describe as the “skills-based organization.” We changed our terminology to match the language used in the broader field, but the concept is the same.

You said that GenAI has added some new facets to the approach. What did you mean?

I guess I should have said it has highlighted some things.

Please describe the glimmering facets of today’s talent ecosystem.

The foundation of our skills ontology at EPAM is an understanding of all the tasks that are performed for any role. For any job, we can break all the work down into its deliverables. We ask: “What skills are required to do each one of those tasks?” and “Who has the skills to do the work?” That's how we lay the foundation of skills-based everything: hiring, promotion, education.

What’s interesting is the degree to which we can use that task intelligence foundation to understand, predict, and plan for the impact of GenAI on various roles.

If you look at any task, each one is going to be automatable, or AI-enabled. When we lay that intelligence on top of the tasks, we get a precise sense of which jobs are going to be most impacted by AI.

This is essential when thinking about refactoring work. Let’s say you have five jobs. Those five jobs are made up of say, 100 tasks. In five years, 40% of those tasks will be automated, so now you have 60 tasks that are distributed between five different jobs, and all those jobs are incomplete. There's not enough work to be done to fill them. If you have a task intelligence foundation, you’re able to refactor those jobs in a data-driven way.

You lay out all the tasks on a table, so to speak. The tasks will naturally cluster: tasks one, five, 10, and 12 are going to be common. They're going to have common skills and so they naturally will cluster together. Now you can rebuild the job box on top of existing clusters of work. From here, you create a job architecture based on a collection of jobs.

Let's talk about data in this brave new GenAI world. What's new here?

At EPAM, all our systems of work and people are integrated. In most organizations, these are data silos, and you can’t do anything across them. You can’t even understand the data because you just can't get it together.

We put all our data into a single dynamic data layer. That data layer is where all the magic happens. That's where we apply our ontology of skills and tasks. It's where all our AI lives. It’s where our machine learning lives. That data layer is the magic that makes EPAM, as a business, skills-based. And, very importantly, it is where we are getting tremendous efficiencies from AI agents. AI agents can do the work of people because they have access to the data across all these different systems.

Looking at the current field, from a skills-based perspective, we see data problems. First, most organizations don’t have that data layer. That needs to be created. Second, even if they had that data layer, they’d need to clean up and coordinate all their data in there, which is a massive lift. Third, they must do governance with the skills data and task data. This whole skills thing is, in fact, a data problem.

Going forward, what do you think we should anticipate and look out for? Where do you think we will be two years from now?

The HR tech space is very dominated right now by Software as a Service (SaaS). Those SaaS products, the ones that are related to skills, have their own skills ontology. They don't talk to each other. This is a problem because you can’t make stuff happen because no one agrees on what “skill” is, how to define it, and you can't get all the different systems that have skills data to talk to each other and create value.

I suspect we're moving toward a Data as a Service play. Instead of thinking about the elements in little boxes, think about those boxes as data feeds that go into one place, and you build unique interaction or engagement layers on top of those. It’ll be like what Telescope does.

Also, this is connected to MACH. EPAM's Telescope infrastructure is MACH architected. It's a composable, headless approach. This is where I think, and a lot of people think, the HR tech market is going.

We must also consider also the upheaval that will occur as AI agents come into the mix. Most people have no idea of the impact that AI agents will have. They think about large language models right now, prompting them and getting them to clean up your email or whatever. That is not what makes them exciting. What makes LLMs exciting is their ability to do complete jobs. The work of multiple people can now be done by a single agent. That is going to change very significantly the landscape of work.

Organizations need to understand what tasks are left over for humans. They’ll need to refactor those and rebuild jobs for humans where the humans will leverage AI. This will require internal mobility sites, a need to hire differently, and much more data to make good decisions about individuals, understand them and their relationship to work.

Two years from now, I expect that the technology landscape will look very different. A lot of big players currently in the HR tech space will not be so big anymore. Look out for the little players who are data-focused, who do skills validation at scale—you’ll see a lot more of them and they’ll be much more prevalent. Watch out!

We’ve been talking about the integrated talent ecosystem for the past few years, but really it is the skills-based organization concept. Everybody else calls it this other thing, the SBO, and so now we’re calling it this other thing. There's a lot of stuff that we do that is unique to EPAM. I believe that about 80%, maybe even more of what we do, is how everyone's going to operate in the future. We're kind of like the day after tomorrow for organizations, and that is very exciting.