Episode Transcript
[00:00:05] Speaker A: Good morning everyone. Happy Thanksgiving. For those of you in the United States, this last week I've had a massive number of discussions with companies and vendors about the AI impact on job design. So in the middle of writing our predictions for next year, which is almost written, I want to give you some.
[00:00:24] Speaker B: Things to think about which are going.
[00:00:26] Speaker A: To become very significant initiatives for 2025. Now the way we've developed jobs in companies is they're like little containers and they have titles and they have levels and they have role descriptions, task descriptions, skills required pay bands and all sorts of attributes. And these little containers are used for hiring, for promotion, for organization design and many other things. And we embed them into our HR capital systems. So they fixed because once we write.
[00:00:59] Speaker B: Them and they're in the system, we.
[00:01:00] Speaker A: Can'T create 25 versions of the same thing.
And the reason we did that goes back to the industrial age when we had highly scalable companies where we could basically hire a lot of people into the same job and replace the human in the job.
[00:01:16] Speaker B: And the job was the static part.
[00:01:18] Speaker A: And the human was the replaceable part. So it's a little bit like a factory really. Now, as you know, we've talked about this a lot, that doesn't keep up.
[00:01:27] Speaker B: With all of the changes that go on in companies. New technologies, new tools, new organization models.
[00:01:31] Speaker A: New go to market models, new supply chain models, pandemics, people working at home, part time work, gig work. I mean all these things that have happened have punctured the job architecture idea. So companies still have these and we need them because you know, you need to be certified to be a certain type of nurse, or a certain type of truck driver, or a certain type type of repair technician. And we don't want to just wing it all the time and let anybody do whatever they want without knowing who's responsible for what and who has what skills and capabilities and who's trained in what role. And we don't want people stepping all.
[00:02:05] Speaker B: Over each other's jobs.
[00:02:07] Speaker A: So I'm not saying this is going away, but if you look at what's going on in AI, we're going to slice across this in a very disruptive way.
[00:02:16] Speaker B: And what we're essentially finding, and there.
[00:02:19] Speaker A: Are vendors working on this right now, is that AI will interrupt and force you to re engineer job architectures in four distinct ways. The first is today out of the box tools like The Copilot, Galileo, ChatGPT. Whatever you use will take the current job you're doing, whatever it is and make it more efficient. You'll be able to build emails faster, send letters faster, write reports faster, analyze data faster, use the copilot to listen to meetings, summarize meetings, skip meetings. I mean, those things are going to happen. They're already happening.
And you know, Microsoft's doing these studies and finding out that people are saving, you know, 30% of their time every week. Our Galileo customers are telling us they're saving a day, a week of work, which is 20% of their time. I think it's more than that because they're doing more work. So that's happening. And we call that the efficiency super worker.
The person who's doing the same job.
[00:03:18] Speaker B: Can do it a little bit faster.
[00:03:19] Speaker A: Maybe a lot faster using these tools. And we're not tweaking the tools that much for these use cases, although that's coming. The second implementation of this is what we call the empowered super worker. So let's suppose the AI tool could.
[00:03:34] Speaker B: Do the work you're doing faster than.
[00:03:37] Speaker A: You could do it.
[00:03:38] Speaker B: So not only will it help you.
[00:03:39] Speaker A: But it will take it over for you. This is the software engineer that can prompt the agent to write code 70%.
[00:03:48] Speaker B: Faster than they could do it themselves.
[00:03:50] Speaker A: So now the engineer is not writing the code, the machine is writing the code and the engineer is prompting it and testing it and improving it. That's going to happen to all of us too. I'm waiting for Microsoft to announce a copilot in office that will read my emails and tell me which ones are the most important and stop me from spending all day watching them come filtering in and figure out which ones to read. I mean, there's going to be all sorts of tools like that that replace things that we do automatically and we will have to modify and edit and tweak them, but that's going to create.
[00:04:24] Speaker B: An even higher level of productivity. So those are two use cases for everything we're doing today.
[00:04:30] Speaker A: But that's not really what's going to happen. What's really going to happen is we're not going to do it this way anymore because these industrial designed job architectures.
[00:04:39] Speaker B: Just don't even make sense.
[00:04:41] Speaker A: So let me go to two more use cases.
[00:04:43] Speaker B: The third is what we call the highly productive super worker.
[00:04:47] Speaker A: So when you go to McDonald's today, and I don't go very often, but every now and then I do, there.
[00:04:52] Speaker B: Is no counter to order anymore.
[00:04:54] Speaker A: There's a kiosk or an app, and the same thing's true at Starbucks and Pete's and lots and lots of restaurants too, where you get through some AI like experience. You pick what you want to order. Maybe the AI gives you recommendations because.
[00:05:09] Speaker B: If it knows who you are, it.
[00:05:10] Speaker A: Remembers what you ate last time. It might even know the time of.
[00:05:13] Speaker B: Day and it might know what type.
[00:05:14] Speaker A: Of food you're interested in, this time.
[00:05:16] Speaker B: Of day, or what's on sale or.
[00:05:18] Speaker A: What'S fresh or whatever. And you do your thing, you order it and you walk in and you pick it up. Now the human worker is watching these.
[00:05:27] Speaker B: Transactions come in and they're doing the.
[00:05:28] Speaker A: Work to fulfill it. So now their job is completely different than it was before and the scale is going up by an order of magnitude. Now you could get five times as many people through the restaurant.
[00:05:39] Speaker B: Now the customer service is not quite the same.
[00:05:41] Speaker A: You can't ask the AI the same questions about what the specials are today and so forth.
[00:05:46] Speaker B: But that's going to happen too.
[00:05:47] Speaker A: So that's number three where we've used the power of the AI to redesign the flow of work itself. And now the humans are supporting a much more automated process.
[00:06:00] Speaker B: High, high, high levels of productivity. Lots and lots of examples of that in retail and customer service.
[00:06:05] Speaker A: I mean self service in HR is.
[00:06:07] Speaker B: Clearly this type of application.
[00:06:09] Speaker A: There's going to be self directed learning, there's going to be self directed, you know, coaching and all sorts of things like that.
[00:06:16] Speaker B: That's level three.
[00:06:17] Speaker A: Level four is where we have specially.
[00:06:19] Speaker B: Designed AI agents that are highly integrated systems. Think about a self driving car.
[00:06:24] Speaker A: Think about an automated recruiter that can do all the steps of recruiting. Think about an automated course developer that can build a course and interview prosp.
[00:06:33] Speaker B: Interview subject matter experts for you to build the course.
[00:06:37] Speaker A: Think about a jet plane, a fighter plane, a drone, a machine that can watch the nature of the incoming parts.
[00:06:44] Speaker B: To see if the parts or out.
[00:06:46] Speaker A: Of scale and then change its tuning.
[00:06:48] Speaker B: Of how it's doing.
[00:06:49] Speaker A: Very machining operations based on the parts coming in, based on visual recognition. I mean these. Now we're at a point where the machine is actually making decisions on its own using vast amounts of data that.
[00:07:01] Speaker B: Humans couldn't possibly accommodate in at once.
[00:07:04] Speaker A: And we're monitoring the machine so we're.
[00:07:06] Speaker B: The manager or the coach or the.
[00:07:09] Speaker A: Monitor or the supporter of the machine. And that's level four, you know, and that's where we're going very, very quickly.
[00:07:15] Speaker B: I'm in a self driving car.
[00:07:17] Speaker A: There are people behind the scenes in.
[00:07:19] Speaker B: A self driving car, but there won't be forever.
[00:07:21] Speaker A: There's certainly not people behind the scenes in a lot of the Self running factories we have and this is going to hit us in the white collar world a lot. And as I talk about in the predictions report, you're going to see that.
[00:07:33] Speaker B: The reason this is so powerful in.
[00:07:35] Speaker A: White collar work is that the complex.
[00:07:38] Speaker B: Tapestry of data and information that we.
[00:07:40] Speaker A: Have to use to make decisions in.
[00:07:42] Speaker B: White collar is very slow for humans to adopt.
[00:07:45] Speaker A: Okay, let me give you two examples.
[00:07:47] Speaker B: So sales. I have to make a lot of.
[00:07:48] Speaker A: Calls on clients and before I make a call I want to know who this person is, I want to know who's talked to them, I want to know what's going on in their company, I want to know what's going on in their industry, I want to know if we've they bought anything from us before. I want to know something about can.
[00:08:03] Speaker B: I listen to their speeches online, can.
[00:08:04] Speaker A: I look at their LinkedIn profile before I get on the phone with them.
[00:08:07] Speaker B: And ask them a bunch of silly questions? If I can get smarter about them, we can have a much more meaningful.
[00:08:12] Speaker A: Conversation that takes hours of work or.
[00:08:16] Speaker B: It'S certainly an hour or sometimes longer.
[00:08:18] Speaker A: Well, there's no reason AI couldn't do that for me. AI couldn't collect all that information, write me a script or a narrative or tell me in words, in language and sound, what is this person going through and what is their job and role and company like. And then we can have a much, much more meaningful conversation. And I don't have to do all that research. That's one example. Another example is an HR business partner. The HR business partner gets dragged into a meeting about performance management because somebody's underperforming. And we don't know whether to put this person on a PIP or ask them to leave or give them training.
[00:08:53] Speaker B: Or give them more coaching or talk.
[00:08:56] Speaker A: To their manager or whatever it may be. The HR business partner goes digging around looking for all sorts of background, does a bunch of surveys, interviews a bunch of people, spends a month analyzing what to do and comes back with some recommendations.
[00:09:07] Speaker B: All that information could be available in.
[00:09:08] Speaker A: The current systems and could be generated by AI. I mean there's a whole bunch of these examples where the complexity of the information is voluminous.
[00:09:19] Speaker B: It takes time for us to get.
[00:09:21] Speaker A: It and we may or may not have the wherewithal to make sense of.
[00:09:25] Speaker B: All of it, but the AI can do that.
[00:09:27] Speaker A: Rippling's new performance management system does this for example.
[00:09:31] Speaker B: It looks at all of the engineers.
[00:09:33] Speaker A: All of the call center agents, all of the salespeople in a company and they're doing this in a few clients now. And it looks at their activities, which.
[00:09:41] Speaker B: Are recorded, because most of us have.
[00:09:43] Speaker A: A lot of recorded meetings and a lot of recorded activities.
[00:09:46] Speaker B: And it analyzes for the high performers.
[00:09:48] Speaker A: What are they doing versus what are the low performers doing, and gives the managers tips and coaches on how to evaluate and support the lower performing people in the organization. Again, managers can do that if they want, but it takes a lot of time and energy and most of them are not even interested in spending the time on it because they have other things to do or they're not even sure how to do it. So these level four, level three and level four changes are not the same job. You're not going to have the same job title, the same job description, the same job level. In fact, I think in a lot of the super worker jobs we're going to see at level three and level four, we're going to have deeper skills needed, more extensive understanding of the organization, better managerial capabilities, higher pay.
[00:10:36] Speaker B: I think in all four of these.
[00:10:38] Speaker A: Scenarios, there will be routine work that will go away, as happens in hospitals and lots of other examples, but the people that remain are going to make more money.
[00:10:47] Speaker B: And so the 2025 world of AI.
[00:10:50] Speaker A: Isn'T just about finding all sorts of great corpuses of data and managing it and using it and learning how to, you know, adopt and understand these tools. We're going to have to rethink how much money people make, what their skill level is, who's best suited for managing AI versus implementing AI versus curating AI versus monitoring AI. And there's no perfect answer to these things yet. I mean, I've been in some automated factories and I go to Whole Foods where there's all sorts of automated stuff. And you know, the workers there are just doing their jobs. They're just assuming that these systems are there. So as magical as they may seem.
[00:11:28] Speaker B: In the design phase, in the early.
[00:11:30] Speaker A: Days of these conversations, they become rather.
[00:11:33] Speaker B: Mundane after a while and we learn.
[00:11:34] Speaker A: How to use them and they get smarter and smarter and behind the scenes there's somebody optimizing them and then the human rules evolve. Now, one of the big issues that we're going to talk a lot about next year is the unemployment rate, layoffs, reskilling career pathways, and what's going on in the economy and in the United States. This last couple of weeks. We had a big election, of course, and the Democrats got thrown out for not understanding the role of income inequality correctly. This is the analysis that most people have. And the Republicans basically came and said we're going to take care of your earnings, we're going to take care of your lives, we're going to take care of your jobs, we're going to reduce immigration, we're going to add tariffs and we're going to protect your companies so your companies can grow and you can make more money. And the theory behind that is that we as employers, which basically the economy is businesses, the economy is not the government. The government uses the economy, but I mean the economy is us, that we will improve the standards of living of our employees through these automation projects. Well, let me sort of say publicly that that is a question to be answered. And you know, my philosophy on this and my advice and the experience I've had as an analyst is that if you think the role of automation is.
[00:12:53] Speaker B: To eliminate people and to run your.
[00:12:55] Speaker A: Company with no people, you probably aren't.
[00:12:57] Speaker B: Thinking about this holistically.
[00:12:59] Speaker A: There are companies that have very few people. Craigslist was one of those companies, by the way. Craigslist had almost no staff. Their revenue to employee ratio was just spectacularly high. But they don't do much.
[00:13:10] Speaker B: It's a fairly simple system. What we're going to do, what we're.
[00:13:13] Speaker A: Really going to find is that the human being, the human body, the human soul is a value creation animal. And as these tools come in and.
[00:13:22] Speaker B: We learn to use them and they.
[00:13:23] Speaker A: Replace more and more routine things, we are going to add value on top of them.
[00:13:27] Speaker B: We are learning creatures, that's what we do.
[00:13:30] Speaker A: That's really the way we're born from when we start as children.
[00:13:34] Speaker B: So what you really want to think.
[00:13:36] Speaker A: About in the super worker organization is further empowerment of the people around these objects. So if you're a bank and you're building wealth managers, which is a big role in banks, is how do we get more money into the bank through wealth management and, you know, whatever you call them, relationship bankers, the relationship banker is going to get smarter and smarter and better and better through these AI tools. We're not going to delegate all the work to a bot and hope that the bot does a better job than the human, because that can be replicated by another competitor. The bots that you have are not going to be that different than the bots that your competitors have.
[00:14:12] Speaker B: It's how you implement them, it's how.
[00:14:14] Speaker A: You train them, it's what you put on top of them and how you configure them and how you put the right information into them that's going to differentiate your company. And that's true for factories and all the other automation things we've done in the past. And so I would hope that as we move into this 2025 fascinating time.
[00:14:35] Speaker B: That we don't spend a lot of.
[00:14:36] Speaker A: Time talking about layoffs and salary reductions and pushing people down, but the opposite. We focus on pushing people up and.
[00:14:45] Speaker B: Finding ways to increase talent density and.
[00:14:49] Speaker A: Increase pay and increase standards of living and increase flexibility from these bots. Now, everybody doesn't think like me, and a lot of you are going to do the opposite. A lot of you are going to, you know, replace a lot of people with software and machines and let the IT department run the operation. I think that will turn out to be a mistake over time. Even if you look at Facebook or Meta and the problems they've had over the years with AI trying to clean up horrible social media posts and crime and all sorts of problems, they have had to hire tens of thousands or hundreds of thousands of humans in very, very difficult, stressful jobs to manage that gigantic mess of content that gets created on those systems. Amazon.com has more employees than almost anybody else on the planet. Walmart has more than a million employees.
[00:15:40] Speaker B: Great companies don't use automation to replace humans.
[00:15:44] Speaker A: They use autom to empower and leverage.
[00:15:48] Speaker B: Humans in a greater way. And so what I would hope will.
[00:15:51] Speaker A: Happen, and I can't speak to the Trump administration, but that over the next couple of years, we see pay levels go up for most roles.
[00:15:59] Speaker B: Now, the companies that can't automate or don't think about it, or don't care.
[00:16:03] Speaker A: About it, or they're just too tactical to deal with it, are going to fall behind. They're going to have more routine work. They're going to have lower paid workers. They're going to have to prove to their competitors that they can keep up with the AI machines in a human way that may work. I mean, that, that does happen. By the way, there are a lot of consulting firms. For example, if you look at Accenture, Deloitte, PwC, who I know really well, those three, certainly McKinsey, I mean, they don't have that much automation. You'd be surprised how unautomated they are. They have very smart people who work very hard and have very good practices for training and learning and collaborating and doing work together, they could potentially, theoretically be disrupted by a vendor, tcs. The folks at TCS and I have been talking about this. For example, there's a vendor that can go into ERP systems and do testing, automated, automated testing and change management using AI, theoretically, that could displace a lot of the services that companies like TCS.
[00:17:04] Speaker B: And Accenture do when you implement new.
[00:17:06] Speaker A: Systems, they know that, but they're willing to deal with it and they know it's happening and they'll use those tools and they'll have human beings to support them.
[00:17:14] Speaker B: None of this happens overnight. It happens over a period of years.
[00:17:18] Speaker A: And we'll probably look back 10 years from now, assuming we're all still here.
I hope we are, and we'll suddenly realize that there were some spectacular transformations that we couldn't have predicted that happened. So I'm not going to write the book in this podcast. You're going to read all about this in January when we publish the predictions. Stay tuned for some really cool stuff. One of the pieces we're going to put out before the end of the year is the 100 use cases for Galileo.
[00:17:45] Speaker B: We have developed a hundred comprehensive AI.
[00:17:49] Speaker A: Powered use cases for hr. And when you read through that, you're going to your jaw is going to drop and you're going to say, wow, let's do that. Let's do this. This is something that we're learning and going to continue to do over time to help you guys understand what the role is of AI in the year ahead. I hope you've all had a nice holiday week and more to come on this topic. Please go to the JBA to our Academy if you want to learn more about this. We've added a new course on skills based organization in there, just launched last week, and four courses on AI on different aspects of AI with lots of use cases and we'll continue to add more in there.
[00:18:25] Speaker B: And get your hands on Galileo.
[00:18:27] Speaker A: We're going to put more and more of our resources into Galileo over the time ahead. Talk to you again soon. Bye for now.
