Episode Transcript
[00:00:03] Hey, everyone. This week I want to give you some updates on the AI landscape. Two big announcements coming next week, at least two. One from LinkedIn, one from SuccessFactors. I won't give you the details, but you'll see them on Monday and Tuesday.
[00:00:18] And generally speaking, based on the feedback I got in Europe last week and the conversations that I talked about in the article about digital twins, these AI tools are turning into digital employees. And the question that a lot of companies are asking is, how do I reengineer, redesign, reorganize my company around these highly productive, highly intelligent digital agents? By the way, Microsoft also announced a bunch of digital agents. And there's a big PR war between Salesforce and Microsoft because Marc Benioff thinks he's going to own digital agents, which I seriously doubt. So I had some conversations with a bunch of L and D people this week about this, because we're trying to figure out how to apply skills to work. And believe it or not, this is all connected. And here's how it's connected. Take a job salesperson, hr, business partner, recruiter. You name the job. They're all analogous. And there are business processes and tools we use to do our jobs. There's a certain way we write a job description. There's a certain way we open an opportunity in Salesforce. There's a certain way we're supposed to talk to a client or a prospect. There's a certain way we're supposed to talk to a job candidate. If the client decides to buy something, we've got to go through contracts, reviews, et cetera. Same thing in recruiting. These are step by step things that the humans do at work. And if you're building a construction site, these may include licensing the subcontractors, defining the scope of work, sending things out for bid, creating contracts, talking to the city about the zoning laws and other items that the city has. And if you're a software company, they would involve the steps of building the software that the customer wants if you're a contract software company. And in all of these situations, what we did in the past is we tried to organize this work around job titles. So we said sdr, sales development rep, recruiter, software engineer, software test engineer, construction project manager, construction site manager. And those job titles were what I call mashed potatoes. They had a bunch of stuff mixed into them, and they. But. But they were broad enough, yet specific enough that we could define the work that needed to be done in the context of a bunch of human activities, which we called a job.
[00:02:45] And that thing that job was scoped and defined usually by a hiring manager, not by an organizational designer. And so if you were the manager of the sales team, you would decide what the SDRs did versus the account reps. And maybe if your company was really rigorous and engineering oriented, you would industrialize that and standardize it. But honestly that's not very common. I find. Unfortunately, most companies aren't very good at standardizing things. They like to delegate organizational responsibilities to VPs and managers and they expect the VPs and managers to figure it out on their own. In fact, I remember talking to Nicole Lamoreaux at IBM about this and she said that there was a tendency at IBM to believe in the successful leaders and assume that because they had hit their numbers in the past, they, they knew how to do things like Org design, which of course they didn't, but maybe they just knew how to manage people well. So anyway, this is kind of where we are in all of our companies. And along come these AI agents and they're really good at a lot of things. They're good at information retrieval, analysis. They can take multi step processes and gang them together and do them very fast. They can do scheduling. You can interact with them as if they're humans to make them smarter. You can load them with information so that they understand your brand and your culture. And to some degree you can decide what you want them to do because they're programmable. So they're not like human beings, which aren't really that programmable. Human beings can be trained, but you know, we're all kind of born with certain DNA and we kind of are good at what we're good at. So now we have the opportunity to redesign all these work things and use AI in the middle. And there's kind of two ways to do this.
[00:04:33] One is you buy the AI, you experiment with it, and you figure out one team at a time or one function at a time, what it's good at. And you try to find ways to reduce routine work and speed things up. And that's not a highly engineered process, but it works. This is what's going on in most companies today. People are experimenting with AI and they're learning where it's useful. And then product companies like LinkedIn, who is about to announce some stuff on this area, and Paradox and others productize this by looking at user journeys and studying where the glitches are and tweaking the AI to be good at certain things. In the case of Paradox, for example, which is a very, very sophisticated recruiting system, they know exactly what kinds of questions candidates ask. They know what kinds of questions candidates ask to apply to the job. They know how to do interview scheduling, they know how to do assessment. All of the stuff that recruiters do. And there's at least 40 steps in recruiting. They've engineered AI interfaces to make this all faster and easier and less human centered, or rather less burdensome, I guess is the word on recruiters. But in a lot of the new things where we haven't automated like sales, we don't really know what all these AI things could do. So companies like Salesforce and LinkedIn and others are experimenting with agents that they believe from a product management standpoint are going to automate your work. So we're going to experiment and we're going to see these agents appear in the product space and we're going to use them. And we're going to probably believe the vendors when they tell us they save 70% of your time. Okay, that's all well and good and that's going to happen, but there's another way to do it which is a little more conceptual but might be better. And I'm offering you this as an alternative. I'm not telling you which one is preferable. And that is you sit down with the work team and you decompose the work into tasks, critical tasks. And you know, that takes a little bit of time, but it's not very hard. If I sat down with the sales team in our company and I said, let's just walk through what happens from beginning to end in a sales cycle, it would be very easy for them to explain. Well, first we have to do this and then we have to do this and then this happens and we have to call this person to do that. And you could put it on a whiteboard and I would be willing to bet in one day we could write down every possible iteration of, of what goes on and we could look at it visually and we could say, you know what, there's four massive bottlenecks here. Some of these are very human centered, some of these are not. You know, HubSpot's not capable of doing this. Monday.com is not capable of doing that. And we could design agents to fill those gaps.
[00:07:16] And even better, if we really did this form of task oriented look, we could then say what skills are needed in each of these tasks. And this is a massive break, breakthrough in the way we think about skills. Here's my problem with the skills work that's going on. We keep associating skills with people because we are humans. And so we naturally think that if we can associate skills with people, then we can use them for development and career and job placement and better recruiting and sourcing, et cetera. And that's all fine, but we haven't made the associated match of skills with work tasks. So if we say we need a person that knows how to do object oriented design, okay, I mean, that's a conceptual skill and maybe we can identify that. But what about the work? How much of the work requires object oriented design? Which part of the work, how long is that project in object oriented design? Do we need a full time person or a part time person to do that? So what's going to happen, I believe in more sophisticated companies is we're going to sit down and we're going to just look at these workflows and we're going to figure out not only where the bottlenecks are, but what are the skills needed in each of these steps. Because once we know that, now we know what human skills to develop, who to recruit, what to outsource, what to automate. And maybe something requires a highly technical skill, we should redesign the way it works so that it isn't so hard to do. And I had a really good conversation with a woman who's going to, I'm going to interview for the podcast later who works for EPAM Systems, a very, very sophisticated software engineering company. And she said the thing that we don't realize about skills is that subject matter experts, by the nature of their experience, have a lot of skills they don't even know about because they just use them and they just work. Whereas when you're sitting there recruiting people looking for skills, you're looking for these individual fine grained skills. But they really fall into clusters or groups, which the way humans work, we tend to cluster together. Like for example, my skills in writing and data analysis are not complemented by skills in acting. I'm not a good actor. I don't even tell funny jokes. It's just not the way I am, it's not the way I'm wired. So if there's a job that requires joke telling and data analysis, you're going to have a hard time finding a person that does that. You might, there are, I'm sure, engineers that are really funny people that are good at joke telling. It's just not me. In fact, one of my best friends is actually that type of person. But we have to sort of re engineer work from the bottom up. Now, I'm not saying you're going to do this for every job and every workflow in the company. But I'm telling you, if you don't do this, you're going to get incremental improvements. If you do do it, you can get spectacular improvements. That is how companies like Amazon have blown past their competition in things like product delivery and E commerce. They study granular activities on their websites, in their buying process, in their pricing, in their channel relationships, so that they can go in and decompose these things and automate them. And I think with the new AI stuff that we now have access to, you're all going to have to think about this more. Now, the final comment on this whole topic. We're really doing a lot of work on L and D and honestly I'm kind of disappointed with what I'm finding. We've interviewed 30 or 40 chief learning officers and I'm going to talk about this at the LinkedIn conference this week. I'm really kind of saddened by the fact that for the last decade I think the chief Learning officer in a lot of companies has been somewhat left aside and maybe even demoted because all of the exciting projects in hr skills analysis, AI flattening of the organization Org design has not involved these people. They've been left with their LXPs to try to figure out how to build great training. And I think when we get into this new world of AI where we're really trying to apply it surgically to strategic problems, we're going to want to engage L and D in this process. And my best example of this is Mary Glavaca at Rolls Royce who has decided not to sit around and build a bunch of courses, but to work directly with engineering at Rolls Royce to look at exactly what happens when new engineers join the company and build an AI based onboarding program that would be absolutely transformational for that company. So for those of you that are in L and D, this is your calling. This is your chance in 2025 to join and rise up into this challenge and exercise some leadership and help build a better solution to job design Org design and skills and AI. Now, I'm not saying you can do it alone. You're going to have to work with the Org designer Org development group in the company too. But these are projects that should not be left alone to the business to do without your help. Because all of your experience in training and micro learning and user interface design and understanding the role of content and content generation is critical in this re engineering that's going on now. I tried to put together a simple video on this, which I'll link to in the podcast. We're going to talk a lot more about it. EPAM will be at our conference in May. I'm going to talk to some other companies. If you are doing some of this job task re engineering and you would like to share what you're doing, let us know. I want to talk to you. I want to bring you to the conference because I think this is going to be a big, big new thing in the year ahead. Thanks for sa.
