Predictive skills architecture is reshaping talent intelligence with Mickey Raie

[00:00:00] Vijay Swaminathan: Good morning, good evening, or good afternoon to our global audience. I'm super thrilled to be here, my name is Vijay Swami, I'm CEO of Draup. We conceptualize this Draup dialogues primarily to bring all the great best practices in workforce planning, skills architecture, and other related topics. And we get to work with global customers and sometimes we see wow.

This is a great approach they're doing is honestly a moment of learning. And when that happens, we wanna bring it to broader audience. And for me, one such, inspirational leader is Mickey Mohit Raie, from Accenture. We call him Mickey. So Mickey, glad have you here. Super thrilled to be here. Just by way of background, Mickey

brings in huge analytical and leadership experience. He has put that to good use in predictive skills architecture, a topic that pretty much every other organization is attempting to solve for. I thought we will take the next 20, 25 minutes and have some conversation.

Mickey, maybe we'll start with your introduction and then jump right into a few questions.

[00:01:33] Mickey Mohit Raie: Absolutely, well thank you for the kind words first of all. My name is Mickey Mohit Raie, work for Accenture and in here I lead the skills, analytics and insights for us.

It's an absolute pleasure to be talking to you today, Vijay.

[00:01:48] Vijay Swaminathan: Thank you.

So Mickey, one question I ask is, I've been asking this across many forums and questions, right? So the skills architecture is a problem that we've been attempting to solve for in the last say, post cloud disruption now AI era and so on. But very few organizations have actually gone into.

Really implementing predictive skills architecture, right? So make that predictive, make that utility driven. And when that happens, you not only become reporting sort of COE, but you actually start recommending or dictating, what are the skills that the organization really need to acquire and become a very strategic partner to the organization.

I know many companies are attempting to solve that, and you have done that brilliantly. What has your experience been in moving from skills architecture to

predictive skills architecture? Maybe we'll start from there and then, have some interesting follow ups on that.

[00:03:04] Mickey Mohit Raie: Absolutely. In fact, before we dive into that, let's, you know, quickly demystify what predictive skills architecture would actually mean for a lot of the people here listening to this.

And let's start with the skills architecture itself, right? When you think of skills architecture at its core, it's how an organization identifies, organizes, manages the skills that its workforce needs, both now and in the future as well. Think of it, and we like to call it the common thread that connects all HR processes like hiring, learning, staffing, et cetera.

A predictive skills architecture takes this a step further. It helps organizations look ahead, anticipating future skill needs, proactively building or hiring for them today. And with this, with this predictive skills architecture, companies like ourselves, we leverage data that we already have on our people. Which could be from their project assignments, their learning completions, certifications, to get a view of what is the current

landscape of our skills. And then use AI, use intelligence to identify what that future need would be and to map it together, bring it all together, and help solve- help the organization solve many challenges that exist today, but also can be anticipated for the future.

[00:04:42] Vijay Swaminathan: Got it, that is really well said, and thank you for demystifying that a bit

because sometimes these terms can have different meanings in different organizations. Is there any way you can give a practical case study on these are all the specific skill gaps that we were able to point out that the organization found it very valuable. Any- anything on that lens, Mickey?

[00:05:11] Mickey Mohit Raie: Yeah, so an interesting case study I think could be when we think of how, let's say staffing, happens, used to happen in our organization and that basically let's say was it- we're a skill-based organization, so staffing based on skills. But what with the advent of thinking more predictive in nature, we're- we've also brought in information which, which brings in the proximity related aspects to a skill.

So when now we staff it is not just based on the skills that a person has, but also on proximate skills that a person would have. So thereby, our let's say the talent base on, based on which we're gonna be doing the staffing has become

manyfold. So if Mickey had 10 skills, now Mickey has 10 plus maybe 50 more skills which are approximate.

So that's how we trying to solve that problem in the staffing scenario.

[00:06:19] Vijay Swaminathan: Wow, that's pretty good. You- this conversation went really deep, real quick. So the proximity skills, or the non-obvious skills is a problem that we've been trying to address from the industrial age, right?

So if you look at all the way from assembly line that Ford invented. Ford said, you know, you do a lot more than the basic operations, so I'm gonna transform the operations right? But very- when it comes to actual implementation, how do you make sure business actually uses that data? And I know you are a big proponent of integrating different systems and exchanging data.

Could you talk about that? What are the best practices and maybe even KPIs we should consider in this journey?

Sure, so for an

[00:07:21] Mickey Mohit Raie: organization that's just starting out, the foundation of something like this, perhaps is a very robust market, relevant skills taxonomy. That I would say is the bedrock.

It should reflect both external market trends, and the company's very unique business strategy. If you get that right, everything else, builds around it effectively. The next layer could be you bring in AI based skill influences. So rather than, let's say in, the case of trying to identify people's skills.

Just waiting to rely on self-reported skills, which might often be incomplete. Organizations can start using machine learning to infer skills from data that already exists. That there is no dear of data that organizations would have on their people, like job histories, the learnings that they've taken, the certifications, the project work they've done.

Ans especially for very, very large organizations, I think it's really important that they get onto this process of using AI to look at people skills. The next step probably is then to look at slightly in the future, look at demand and try and forecast that either through intelligence or through basic models, which can help align future skilled needs with the business direction.

And this is where in fact predictive analytics plays a very key role. Spoting with skills are emerging, which are becoming less relevant, can high- help guide strategic workforce decisions. And finally all of this, it needs to come together in the form of either workforce planning tools, so that it can be embedded into each of the HR processes.

It cannot work in isolation. It needs to be lockstep with all of the processes that exist. You also mentioned KPIs. I think the way to look at this is, after you've put this in place, do you see certain things, right? Do you observe, let's say, reduce skill gaps in some critical roles that you've always hired in the past?

Mapping that the past to what you see now, higher internal mobility, that definitely is one of the things that you could look at. Increase adoption of learning, targeted learning. Alignment to in demand skills and so on. So, if you observe the changes you have made in light of some of these things, it can tell you whether the framework is working and also adding real business value.

[00:10:20] Vijay Swaminathan: Got it, got it. So it is internal mobility, resource utility, resource utilization, and things like that. If I summarize. And you also, probably only very few skills architecture professional who use the word business strategy at the beginning of your narration. And I feel that is absolutely critical and skills architecture folks normally

do not, get that part right. I've done a lot of experiments saying that, do you know your company is making money? And many of them will not be able to give a very concrete answer to that simple question. But I think in your case, you also mentioned that the staffing aspect of that means that every resource that walks into your enterprise is a sort of a treasure. You bring in not just for the immediate need, but also for the future need.

And you also said that the AI is helping you, map the potential. So has anyone called you at the middle of the night and say, thank you, Mickey. I'm now able to staff the project that I would not have been able to before. Any such Archimedes moment for you that you experienced in the last couple of years or so?

[00:11:56] Mickey Mohit Raie: Well definitely there have been a few, there have been a few but I'll tell you this has not been a short journey, this has not been an easy journey. We say business relevance because we've been doing it for a long time now and over many years we've realized how if you don't start there, it's just not gonna work.

The simple fact that we all talk about skills taxonomy, right? You could build it in isolation, it might work, but in order for it to definitely work you need to co-create it with the business. And a professional services organization like ourselves, we are very vast, we cater to all businesses. So in fact, slowly but surely, we worked with everybody.

All businesses across geographies, across industries, across function areas to co-create what we would then call the a robust skills taxonomy. To answer, yes, there have been quite a few wins, definitely there have been a lot of learnings along the way well, but we've definitly recieved good feedback from a lot of folks from business saying we've been able to staff more broadly now, we've been able to broaden that. We've been able to reduce our bench strength. In fact, that was one of the examples that we had in- across Europe. So yeah, we've been, we've had a few, and a lot of learnings on the way.

[00:13:26] Vijay Swaminathan: Got it, one of my curiosity, I've been wanting to ask you this question.

In- prior to AI, we viewed resources as, human resources, as a bit of rigid, in the box, you know, here is a DevOps person, here is a UI person, here is a backend like that. And in AI we are beginning to see, hey, you know maybe some software engineer doing big data work can get into, AI/ML and so on.

But the problem I see is also is sometimes AI over pronouncing. Those links and then suddenly everybody can do everything. And then the interpreted skills is or the proximal skills pretty much means you can pick any resource from, in your example, the bench and then they should work because there is this first degree, second degree, third degree link.

How do you- that is also a challenge for businesses to prioritize or rank order the candidates. How do you, how have we handled that or have you handled that? What is your thought on that?

[00:14:45] Mickey Mohit Raie: Sure, so we've tried to do that. We've tried to do it for our business, for our staffing business, but also for our learning teams and the concept of proximate sort of works really well in both scenarios. Let me explain that.

Let's say, we have skill, right? What we do is we're able to look at people across, let's say, a geography across wherever they need to be physically and say who are the best candidates based on what they have and some of the related skills. Everybody in the mind- on the bench, let's say in your example, might have one or two skills, but what we're doing is slightly more holistic

where we're looking at their entire skills profile and through algorithms we're able to say who is the closest map.

So there could be 50 candidates but there will be a number 1 candidate because the person either has that same exact scale at a lower proficiency or they have many others proximate skills at a very high proficiency. And then you could keep playing that with lesser scales at a higher proficiency or even lesser scale with low proficiency and so on. So we're able to almost rank people to say this candidate is the best candidate. This candidate is the second best candidate and so on, based on that.

[00:16:27] Vijay Swaminathan: That's beautiful, that's beautiful Mickey. I think that part is- I'm beginning to see that human resources have a little broader view typically and they have a expansive view on their people.

But some sort of algorithm to prioritize and say, yeah, here's the 50 people, but here is the top five or top three. I think that's fantastic. Is that an in proprietary- I know you build a lot of algorithms, but is there a proprietary algorithm you build? Is there any, without giving the secret sauce, anything you can tell us on how you went about doing that?

[00:17:11] Mickey Mohit Raie: Yep. So we've actually built it in-house. The whole concept of proximity and then that proximity feeding into tools that work off of a proximity but give us those rankings. I think some of the ways, without going into too many details, is we have a lot of proprietary algorithms like graph databases.

We have one which works within Accenture and we also take it to our clients. We have other things like we use affinity analysis, we look at NLP to start linking skills. And we use a lot of data because we- let's literally raise millions of demands every year. So we have- we look at the past

five years, we can even go to past 10 years to see what skills existed on our demands. What skills existed together on our demands. So you know a combination of four or five of these, let's call them algorithms and or approaches, come into helping us identify what could be those closest matches and helping us rank them, in order.

[00:18:31] Vijay Swaminathan: Got it, got it. No very, very helpful. Changing gears a little bit, I was listening to a podcast done by John Roese, CTO of Dell. One of the thing he said is, hey, enterprise is basically two things. The very foundation of an enterprise is just two things. You have some data about your customers, your people, that nobody else has.

And number two, you have people who can do the job that you have chosen to do in a way that no one else can do. These two are the foundation why an enterprise actually exists. So if you look at that as a framework, you have lot of data about, applying it to your scenario, right?

You have a lot of data about people. But in order to- I almost consider that you are using that to directly influence, hey, how do we create a workforce that can be the best of the best to get things done for our customers, for us. In that journey, we seem to have introduced lot of variations.

There is strategic workforce planning, there's people analytics, there is skills, predictive skills architecture team. Then another interesting thing, which is- which I see is, in certain high tech organizations or in engineering organizations, the business leader themselves are doing versions of this. Because they don't want HR to come and do this, or a central team come and do this. Because they think that they won't get it right? They won't get the skills.

But in this complex journey, what is your recommendation to think about how do you integrate all these different pieces and think about implementation? So what is- how do I go about implementing predictive skills architecture with all these complexities because those are not gonna go away, right?

[00:20:57] Mickey Mohit Raie: Yeah, true. I think some of the things, and we've touched base a little bit on it, is to work with the business and do this. But let's say an organization looking to move or go on this journey. I think one of the key things to do is not to treat this as a project, as a one-time project.

The business might be doing it or the HR might be doing it, or they might be doing it together. But the key is never do it as a project, you know, relying on it as let's say creating a single tool or a project alone won't really solve the question. Without a business aligned skills framework and trust in the data, it won't succeed.

So what I would recommend in this case is working with the business to co-create the frameworks and integrate it into their strategic plan, treating it like a core business capability.

[00:22:01] Vijay Swaminathan: Okay.

[00:22:01] Mickey Mohit Raie: I think one of the other things which is quite important. And then something we've learned as well that there might be a lack of trust in the data sometimes.

And people may doubt the accuracy of what you're trying to predict or what you're trying to suggest with leading to resistance. So the idea always should be that you build trust through feedback loops. Run pilots to test predictions, compares results and don't over complicate it.

Just have a very simple sort of approach to it where you start small with some critical areas, some critical roles, and show impact. And again, Irrespective of whether a business is doing it or an HR team is doing it. Certain core principles need to be kept in mind, which we spoke about.

Do it like it's a way of life within the organization. Help people trust what you're doing. Keep it simple and also, very importantly, this shift is not just about tech. It's about embedding the change into the organizational culture. And if business is part of that journey, either doing it themselves or doing it through you, I think that's where the success lies.

[00:23:30] Vijay Swaminathan: Got it.

And, one interesting component that has happened perhaps after this gen AI advent, right? Like the employee aspirations are very high today, especially in a global business like you. People want to learn a lot and some ways they're also learning through so many other channels, in terms of re-skilling. So it becomes even more complicated,

re-skilling has become even more complicated because you are to match employee aspirations that are possibly a bit outside their proximal scales. Like today, there may not be enough AI skills that is proximal to my skills, but I wanna learn that. There is that type- how do you balance that? And and that's a huge opportunity for enterprises.

I think you have done some great things. I just want the listeners to really understand that.

[00:24:44] Mickey Mohit Raie: Absolutely. So there's two parts to it right? It might- this might sound strange. It has actually become much simpler now to bring together people's aspirations to where- what the organization needs. And we'll talk about that first, which is the- bringing proximity into the picture but just the fact that now we're able to see where, what skills people have today and

how does that, or how do those skills align to the skills that the organization is demanding for? Of course will be for people who have, let's say, robust skills profile or where their proximity exists as you were saying. But there could be the second scenario where there aren't enough, let's say, skills that are related.

For that, we've actually built very interesting frameworks in the organization, which we refer to as an aspirational specialization or think of it as where people can go in into the system and say, this is what I aspire for. Now, I might not have those skills.

[00:25:56] Vijay Swaminathan: Got it.

[00:25:56] Mickey Mohit Raie: Neither I have skills which are proximate or proximity don't- doesn't exist. But what I can do is I can put it in the system and what that does is, it helps connect people to the right network. So let's say if I am an HR professional, I say data and AI as an area. I become part of a data and AI community where I will start getting the right inputs. I will start getting the right learning because our learning teams are constantly looking at things in the market, things internally which they can bring.

Just by the fact that I'm able to raise my hand and say, this is what I wanna do. The whole ecosystem around, including learning, our professional networks, and even staffing sometimes where there might be certain projects that are open for shorter durations, all of that is information that starts flowing to you, thereby enabling you, irrespective of whether there is enough proximity for the new age skills or not.

[00:27:05] Vijay Swaminathan: Got it. I love how you put it. So it's current skills, proximal skills, and then aspirational skills. So that's and then aspirational skills. The start of the journey is almost always based on what you said, is through a peer group of sorts or mentorship of sorts. Okay,

got it. That's a deeply valuable utility Mickey, I think for people to consider that. Now

related to that, but also, changing gears a little bit in- yeah I attend one conference a week, present most of the time. I think one of the big thing is AI is not gonna destroy jobs, AI is gonna create more jobs. And then, when we probe a little bit, we say, okay, what is the job?

It's prompt engineer. Then what is the job? Human interaction analyst. What is the job? AI ethics, like beyond a few we truly are not able to come up with what is that massive set of new job roles that are coming up. And what has happened is just to fill up the slides, I see lot of tasks that people do as a job.

Like for example, adapt prompting or something like that. That's really not a job. Any data science machine learning software person should do that. So this

has become a big source of anxiety for many organizations. Hey, lot of there are a lot of automation potential through AI, but not much clarity on

what new jobs are actually truly evolving? What is your perspective on that? To some effect it is a problem that will get clearer in the next couple of years, but I was just curious what you think Mickey?

[00:29:20] Mickey Mohit Raie: Yeah, it definitely is something that a lot of the organizations are trying to solve for, including ourselves.

And, we've taken baby steps on that journey where we've tried to in fact, in order to help us do that, we work with organizations like yourselves. Talent intelligence organizations who help us continuously scan the external market. Look at job postings and industry movements, competitive benchmarks and evolving technologies.

This gives us insights into what skills are growing. And hence, based on that we are able to infer certain things on what roles could be important for the future. We also have some internal teams that work within the organization to identify the impact of AI on all of the jobs and the roles that exist.

And what that sort of does for us is we're able to say, for a job, 20% is what could be impacted by AI.

And what do you do with that then 20%? It could be either be you redeploy them or you look at what other skills you could sort of bring into the picture.

And sort of create that bandwidth. And it's not an easy sort of thing, looking at future. You could, like I said, work with talent intelligence organizations to understand that. Use a lot of- so we have a lot of systems which are connected which do sales for us. So look at that. We have our research teams that are constantly scanning the market out there to see what is selling most in the market working with organizations to review RFPs and so on and so forth.

So many, many different ways of doing this. I don't think there is a silver bullet yet. It's a journey that we- we're on and I think a lot of other organizations would be on as well.

[00:31:38] Vijay Swaminathan: Got it. No, it's a very fascinating problem. I'm sure we learn a lot more in the next couple of years. And one of the, when I speak to you throughout this conversation,

you emphasize a lot about systems, technology, models and it is not the typical HR talk. I know you have a different kind of a background. But it is an aspirational journey for many of the HR folks to understand technology, tech stack and many times we don't even know what is our current- what is our company's tech stack?

You have, in one of the conversations that I had before, you mentioned how important it's to map your own tech ecosystem. What are we using for data storage? What are we using for visualization? What is our central data like platform and things like that. Can you speak a little bit about that?

I think that's a big- a defining moment for our profession. I thought people will find that super valuable.

[00:32:55] Mickey Mohit Raie: Absolutely happy to share my journey. So I've been- I'm an HR professional, through and through, but I think over the past many years, I think me personally, I've been able to start looking at things slightly differently and think how one could drive value with what you have today in the world out there in terms of AI,

in terms the big data that we have on our people and so forth. So I wouldn't still call myself somebody who's very proficient in technology. I think what I like to say is that I'm somebody who can now ask the right questions to the right people so that we are able to journey together on bringing all of that. And yes,

because of the exposure, because of what I've learned on the way, I have become a little proficient in understanding the data visualization techniques for example. How data comes in to the organization and what's the best way to use it and so on. I think in summary,

trying to solve a problem is maybe the starting point, and then thinking of how an AI tool or data can help you solve that at scale. Because scale is very, very important for an organization like ourselves. So I guess that's probably would be it.

[00:34:33] Vijay Swaminathan: No, it's fascinating throughout this conversation I feel like what predictive skills architecture in a nutshell doing is it's unlocking the creativity and the problem solving of the organizations, right?

So that's really our mission because we say, it's not just current skills there is proximal skills, there is aspirational skills. You are- the amount of creativity that's available now through all these LLMs is just mind boggling. Anyone can

write like a Python code using ChatGPT, copy that in Google CoLab and produce some fantastic output.

But if you were to look at, since you mentioned about your journey in terms of- even though you've been a HR professional all through. We just picked up things here and there and then refined it, picked up all the technology, or at least the- may not be an expert at it as you, as per you say, but from a utility practitioner standpoint, the strengths, limitations, all that.

What should we not learn? Or what did you unlearn to get here? That could be a- that's my curiosity now.

[00:36:03] Mickey Mohit Raie: So I think, it's more a thought process that I've unlearned, which is that being responsible for something means having all the answers. I've, unlearned that, because the position and the role that I'm in, it's not easy for me to have all the answers.

[00:36:24] Vijay Swaminathan: Okay.

[00:36:25] Mickey Mohit Raie: And a lot of the work revolves around technology. And I might not be, let's say, a person who knows core technology or has a core AI background. So like I said before, I focus on asking the right questions and enabling others to find the answer. So the thought process of I should know it because I'm responsible is something I've unlearnt.

[00:36:50] Vijay Swaminathan: Got it, got it. Fantastic. Fantastic. Well said. And is there any key skill that you think, hey, I know this, it's gonna become very important. That's gonna be the- your Mickey's next generation skill. What are you aspiring to learn in the next couple of years, or even longer than that?

[00:37:20] Mickey Mohit Raie: In fact, one of the skills is probably not related to a technology or to something that I'm gonna learn but I definitely wanna start practicing it much more is learning agility.

[00:37:33] Vijay Swaminathan: Okay.

[00:37:34] Mickey Mohit Raie: I believe it's the most important skill, not just for an HR professional but any professional. The world is changing

very fast and we need to keep learning new things much quicker than we were required to in the past. So whatever I learned today, as a skill might become obsolete a year down the line. So just having that learning agility is what I wanna really really, practice, as a professional.

[00:38:03] Vijay Swaminathan: Fantastic, well said. Actually that's, I may have to do a separate podcast on learning agility itself because it's- it'll be interesting to see how you unpack that. We are coming up on time here but a few things I learned, Mickey are your ruthless focus on business strategy and what it is, how we can serve the business in terms of very specific outcomes.

That totally energized me. The second thing you said is really the trust, difficult to build, easy to break. The trust in data, right? How you co-build that. To your point, don't just take it on your shoulder and say I have to make sure this data is right. It's everyone's job too, for the data to be trustworthy.

I thought that was super fascinating. The most important thing that I also learned is this is not a- an overnight or one week one. And so you had to plan this a bit longer, be realistic, but have those winning stories on, in terms of KPIs and how you broke that in terms of- I think most of the leaders think in terms of current skills and proximal skills.

We have come that's that far, but you also ventured into the aspirational skills. Super thankful for you. Closing remarks, how do you stay updated? Any book, podcast, that you recommend the listeners get into love to hear that.

[00:40:08] Mickey Mohit Raie: I- there's no single definitive book yet that I- at least that I've read but I would definitely recommend reading articles published by various consulting firms, HR advisory firms.

[00:40:23] Vijay Swaminathan: Got it.

[00:40:24] Mickey Mohit Raie: Focused on things like skills based organizations. I think that, I would say is one of the starting points. There are a lot of things out there without naming the organizations.

[00:40:37] Vijay Swaminathan: Sure.

[00:40:38] Mickey Mohit Raie: But that I would say is a good starting point. And then from there you build on understanding really how skills play an important role and are gonna be the most important thing in the future, so no longer organizations will look for

you know roles or look for titles, they'll actually look for skills. So that I believe is the most important thing.

[00:41:05] Vijay Swaminathan: Fantastic, with that positive and encouraging note, thank you, Mickey you've done- doing fabulous work and more important,

I know how busy you are, so absolutely privileged to share 40, 50 minutes with you and, my sincere thanks for your contribution here.