Latest / Human Capital Business / The HR Leaders Guide To Multiagent HR Architecture. With Josh Bersin
Transcript
- 0:04Speaker 1: Josh Burson, welcome to the conversation today.
- 0:07Speaker 2: Thank you, Jonathan. It's great to be here.
- 0:10Speaker 1: It is great to be with you again. You're joining
- 0:12Speaker 1: us from San Francisco. I'm south of Salt Lake City
- 0:14Speaker 1: in Utah. Today, we are going to be talking about
- 0:18Speaker 1: the latest report out from the Josh Burson Company, the
- 0:23Speaker 1: HR Leader's Guide to Multi-Agent HR Architecture, which is part
- 0:27Speaker 1: of the HR 2030 Initiative, which outlines how HR leaders
- 0:32Speaker 1: must govern and specialized AI agents, and functional super agents.
- 0:37Speaker 1: I'm super excited to have the opportunity to sit down
- 0:39Speaker 1: with you and to discuss this. We were scheduled previously,
- 0:43Speaker 1: had some travel mishaps and whatnot. We had to reschedule. Anyways,
- 0:47Speaker 1: thank you for your patience, and it's a pleasure to
- 0:50Speaker 1: sit down. As we get started, I wanted to share
- 0:53Speaker 1: Josh's bio with everyone, though I think everyone knows who
- 0:57Speaker 1: you are, Josh. But anyways, Josh is... well-known analyst, author, educator,
- 1:03Speaker 1: and thought leader focusing on the global talent market and
- 1:07Speaker 1: the changes and trends impacting business workforces around the world.
- 1:12Speaker 1: He studies the world of work, HR, and leadership practices
- 1:16Speaker 1: in the broad talent technology market. He is often cited
- 1:20Speaker 1: as one of the leading HR and workplace industry analysts
- 1:23Speaker 1: in the world and And I think this may be,
- 1:27Speaker 1: I don't know, the third or fourth time we've had
- 1:29Speaker 1: an opportunity to chat. Always a pleasure. Before we dive
- 1:33Speaker 1: on in, anything you would like to share with the
- 1:35Speaker 1: audience by way of your own background, personal context, or
- 1:38Speaker 1: about the company? And then we'll get going.
- 1:40Speaker 2: With the report. I mean, at the moment, it's all
- 1:43Speaker 2: about AI. Yeah, yeah.
- 1:46Speaker 3: It's all everybody wants to talk about, so we're spending
- 1:48Speaker 3: most of our time on that. But of course, We
- 1:51Speaker 3: don't have to talk about it today, but there's lots
- 1:52Speaker 3: of other questions about pay and job titles and roles
- 1:57Speaker 3: and structure of companies and so forth that's being affected too.
- 2:00Speaker 2: But let's talk about AI first.
- 2:03Speaker 1: Yeah, for sure. It's ubiquitous. It's really at the forefront
- 2:08Speaker 1: of everyone's attention, which is understandable, sometimes frustrating. It is
- 2:16Speaker 1: what it is, right? Well, first, tell us a little
- 2:19Speaker 1: bit about the report generally? Maybe even start with HR
- 2:26Speaker 1: 2030 generally, the initiative, and why this report?
- 2:32Speaker 3: Well, we started HR 2030 about a year and a
- 2:35Speaker 3: half ago because there has been this proliferation.
- 2:38Speaker 2: Of ideas of what AI could do to.
- 2:43Speaker 3: Different parts of HR from recruiting, interviewing, assessing, training, employee experience, analysis,
- 2:51Speaker 3: pay rewards, et cetera. And there's going to be hundreds
- 2:55Speaker 3: of use cases of AI, of course, but we wanted
- 2:58Speaker 3: to kind of get the big picture. So we sat
- 3:01Speaker 3: down and said, let's imagine we could fast forward ahead four.
- 3:05Speaker 2: And a half years.
- 3:07Speaker 3: What do we want the world to look like ideally?
- 3:11Speaker 3: And so we ended up with an architecture of 160 agents,
- 3:16Speaker 3: which sounds ridiculous, but it makes sense when you think
- 3:19Speaker 3: about it. interoperating with each other in a certain sort
- 3:23Speaker 3: of networking effect with eight super agents that manage and
- 3:28Speaker 3: coordinate the 160 other agents. And we introduced this. And
- 3:34Speaker 3: the way we came up with that is we had
- 3:38Speaker 3: spent the last 10 or 15 years studying HR and
- 3:40Speaker 3: the 95 functional domains of HR. So we kind of
- 3:44Speaker 3: knew a lot about the complexity behind this. And we
- 3:47Speaker 3: had already done a whole bunch of research on what
- 3:49Speaker 3: we call systemic HR, which is HR operating as an
- 3:52Speaker 3: integrated system as opposed to a whole bunch of standalone services. And,
- 3:58Speaker 3: you know, I didn't really know what people were going
- 4:01Speaker 3: to think of it. It was almost an intellectual exercise
- 4:03Speaker 3: as much as anything else. And we built a whole model,
- 4:05Speaker 3: a visual model of it and everything. And we started
- 4:08Speaker 3: showing it to people and they said, oh, my God.
- 4:10Speaker 3: please come help us because we're getting flooded with vendors
- 4:14Speaker 3: and tools and technologies and bots and stuff. So it's
- 4:18Speaker 3: become a body of research and a set of tools
- 4:22Speaker 3: to help companies decide where they are and where they
- 4:25Speaker 3: want to go in their roadmaps. Because this is a
- 4:29Speaker 3: lot of change happening out there. And then the multi-agent
- 4:33Speaker 3: architecture piece is one chapter. There's a bunch of elements
- 4:38Speaker 3: to this, but to deal with the issue of this
- 4:41Speaker 3: is a system of agents. This isn't an agent. And
- 4:45Speaker 3: you could have a debate. I just had a meeting
- 4:47Speaker 3: last night with a computer science guy. You know, is
- 4:51Speaker 3: a big agent better or a small agent? And that's
- 4:54Speaker 3: part of what this is about. And if you do
- 4:56Speaker 3: have a bunch of small agents, how do you manage
- 4:58Speaker 3: them all?
- 4:59Speaker 1: Right. Yeah. So 160, you know, specialized agents, the eight
- 5:05Speaker 1: functional super agents, You know, we are in the world
- 5:09Speaker 1: of agentic AI. Long gone are the days of just,
- 5:15Speaker 1: you know, chat GPT and large language models and bots, right? Like,
- 5:21Speaker 1: that is so far behind us now. And like you said,
- 5:26Speaker 1: there were so many tools, so many vendors. Um, I
- 5:29Speaker 1: feel like I'm on calls constantly talking with folks about
- 5:33Speaker 1: the new and greatest coolest thing. And, and there really
- 5:36Speaker 1: are so many cool things, right? There, there's no doubt
- 5:39Speaker 1: about it. I'm sure you're constantly on calls too about
- 5:43Speaker 1: all of these things. And, uh, but, but yeah, getting beyond,
- 5:48Speaker 1: getting beyond just the cool little tool, uh, or, or
- 5:52Speaker 1: the new, the new fad or the new thing that
- 5:54Speaker 1: comes out and every week there's something new, uh, It's
- 5:58Speaker 1: all about the agents and it's all about how to
- 6:02Speaker 1: organize and manage the agents, which I just find so fascinating.
- 6:07Speaker 1: And anytime I get a chance to talk with somebody
- 6:10Speaker 1: who really has figured out how to connect and integrate
- 6:17Speaker 1: the agents, that's really what it's all about. Because so
- 6:21Speaker 1: many people have agents and so many people are talking
- 6:25Speaker 1: about agentic AI and I've seen very few who actually
- 6:31Speaker 1: have it systematized, who really have it integrated, who really
- 6:35Speaker 1: have an approach that seems to be working in any
- 6:41Speaker 1: systematic way. Just so many teams, so many organizations, it's
- 6:46Speaker 1: so haphazard, right?
- 6:49Speaker 3: Well, in fact, you know, the, the, the, the misnomer
- 6:52Speaker 3: or misunderstanding is that if I build a little tool,
- 6:55Speaker 3: a little AI agent that does something, that's a good thing. Well,
- 6:59Speaker 3: it may be, but it may also not be. And
- 7:03Speaker 3: I actually think there's a downside to doing that because
- 7:07Speaker 3: if you have your own little personal agent that does
- 7:09Speaker 3: something and it's not correct.
- 7:12Speaker 1: Right.
- 7:14Speaker 2: What other, first of all, what are you going to.
- 7:16Speaker 3: Do to fix it in a big company? And who
- 7:19Speaker 3: else is using it or relying upon it for what?
- 7:22Speaker 3: And are you proliferating the mistakes you're making into other
- 7:25Speaker 3: people's agents? There's just, you know, these are not simple
- 7:29Speaker 3: little tools. They look simple when you play around with
- 7:32Speaker 3: chat GPT, but they learn over time and they get
- 7:35Speaker 3: smarter and smarter and smarter. So I'm as an architectural thinker,
- 7:40Speaker 3: really trying to tell people to just think about the
- 7:43Speaker 3: big picture before you dive into this cool little.
- 7:45Speaker 2: Thing that you want to work with.
- 7:48Speaker 3: Um, in terms of scale, ultimate connectivity, credibility, trust, protection
- 7:55Speaker 3: of assets, protection of data, protection of security.
- 8:00Speaker 2: And I think that's doable.
- 8:01Speaker 3: But the funny thing about it, Jonathan, is since we
- 8:04Speaker 3: started this, I've been reading up on all these people
- 8:06Speaker 3: that have done lots of work on this, Palantir and
- 8:09Speaker 3: IBM and others. There's not that much written about how
- 8:11Speaker 3: to manage a complex multi-agent system. It's not because the actual...
- 8:17Speaker 3: activity and behavior of an agent is not well known.
- 8:21Speaker 3: I mean, the guy I was talking to last night
- 8:23Speaker 3: does AI safety research. And he said, you know, there's a,
- 8:26Speaker 3: there's like five or six different tributaries of research on
- 8:29Speaker 3: understanding why an agent does what it does. And so, um,
- 8:35Speaker 3: so I think what we've come up with is enough information.
- 8:39Speaker 3: And I talked to SAP and great detail about this
- 8:41Speaker 3: at their conference. I've talked to work day about it. Um,
- 8:45Speaker 3: I think there's some really important lessons here. And the
- 8:49Speaker 3: other reason that I think we have to think of
- 8:52Speaker 3: this systemically is if you think of HR as a
- 8:55Speaker 3: bunch of standalone services, you're going to get your budget cut.
- 8:59Speaker 3: People are going to say, well, why are we spending
- 9:01Speaker 3: so much money? Cut the headcount in half and let's
- 9:04Speaker 3: just do less of this and turn it more over
- 9:06Speaker 3: to agents. If you think about it as an enablement
- 9:10Speaker 3: function for growth, and that's what I think HR ultimately is,
- 9:15Speaker 3: is a dynamic enablement for growth through hiring, through training,
- 9:19Speaker 3: through redeployment, through skilling, whatever it may be, through pay
- 9:24Speaker 3: or job redesign or scheduling or whatever benefits. All of
- 9:28Speaker 3: a sudden, every decision has an impact on the bottom
- 9:32Speaker 3: line and the top line of the company, not just
- 9:34Speaker 3: making somebody happy. Then you want these things to work
- 9:37Speaker 3: well so you can learn. Because what we're going to
- 9:40Speaker 3: do is we're going to learn lots and lots of
- 9:41Speaker 3: things about our companies that we didn't know before because
- 9:44Speaker 3: AI is so smart.
- 9:46Speaker 1: Yeah, I love that. The enablement function for growth. I
- 9:50Speaker 1: love that framing. And you just mentioned it a minute ago.
- 9:54Speaker 1: I want to double click on it. You know, the,
- 9:56Speaker 1: the failure that comes from unconnected agents, uh, right. Which,
- 10:02Speaker 1: which is happening all around us. So everyone, you know, we,
- 10:06Speaker 1: we feel so special because we, we create our own
- 10:09Speaker 1: little agent and then, and then, you know, which may
- 10:12Speaker 1: work really well for our own little workflow in our
- 10:15Speaker 1: own little world and our own little silo. Um, but,
- 10:19Speaker 1: it can cause problems.
- 10:22Speaker 3: Well, just yesterday, I don't know if you saw yesterday,
- 10:24Speaker 3: Goldman Sachs just shut off Anthropic for all their non-ITF professionals.
- 10:29Speaker 2: They don't want people hacking around with.
- 10:30Speaker 3: It anymore because it was expensive and it was actually
- 10:33Speaker 3: not working out very well. I think one simple way
- 10:37Speaker 3: to think of it is if you gave everybody a
- 10:38Speaker 3: copy of Excel and said, build a spreadsheet that analyzes
- 10:44Speaker 3: whatever it is you're doing and just share it with
- 10:46Speaker 3: everybody else in the company. Can you imagine how many
- 10:49Speaker 3: errors would be suddenly just proliferated.
- 10:51Speaker 2: Throughout the company, right? Because everybody's spreadsheet's.
- 10:55Speaker 3: Got glitches in it, and it works for me, but
- 10:57Speaker 3: it wouldn't work for you because you don't know how
- 10:58Speaker 3: I did it or why I did it. It's a
- 11:01Speaker 3: little bit like that. So I'm not saying personal agents
- 11:05Speaker 3: are going away. We're all going to have all sorts
- 11:07Speaker 3: of cool personal agents, but let's just do it in
- 11:09Speaker 3: a managed process. Another analogy that I think is useful
- 11:13Speaker 3: is the more I talk to companies about this, the
- 11:17Speaker 3: more I do believe that the agents are like people.
- 11:21Speaker 2: So if you think about a.
- 11:25Speaker 3: New employee in a company, the employee has a job
- 11:29Speaker 3: title and a role and a set of goals and responsibilities,
- 11:33Speaker 3: and you don't give the employee access to all of
- 11:36Speaker 3: the information in the company, all of the secret projects,
- 11:42Speaker 3: all the financial information, because they're going to make a mess.
- 11:47Speaker 3: That's not their job. So we think of the agents
- 11:50Speaker 3: as individuals with very clearly defined specific responsibilities that we
- 11:55Speaker 3: want them to be really, really, really good at this.
- 11:59Speaker 2: All of a sudden, that.
- 12:00Speaker 3: Point of expertise is really useful to other agents who
- 12:04Speaker 3: need it and don't want to replicate it.
- 12:07Speaker 2: So I think the.
- 12:08Speaker 3: Division of responsibility or whatever the right phrase, there is
- 12:11Speaker 3: a word for that. I don't think that's exactly the phrase.
- 12:14Speaker 3: clear delineation of who's responsible for what really makes AI
- 12:19Speaker 3: safer and more scalable and more powerful. And there's different
- 12:24Speaker 3: ways of mixing things up, as the paper discusses. But,
- 12:29Speaker 3: you know, I think a lot of these rogue agents
- 12:31Speaker 3: that went nuts and, you know, hacked into other websites
- 12:35Speaker 3: were probably experiments where the agent was deliberately given a
- 12:39Speaker 3: whole bunch of uh, opportunities to make mistakes and they
- 12:42Speaker 3: were testing how far it would go.
- 12:44Speaker 1: Yeah.
- 12:45Speaker 3: And, you know, because these things are probabilistic, they're going
- 12:48Speaker 3: to keep trying to achieve their goal until they find
- 12:51Speaker 3: a way to achieve it. And if you give it
- 12:53Speaker 3: too much opportunity, it's going to go down to the
- 12:55Speaker 3: tail end of this probabilities curve and do something dumb.
- 13:00Speaker 2: Right.
- 13:00Speaker 3: Yeah.
- 13:01Speaker 1: Yeah, absolutely. Um, One of the things that's talked about
- 13:06Speaker 1: in the report is workload reduction. And of course, if
- 13:09Speaker 1: we have a bunch of productive agents, this is bound
- 13:15Speaker 1: to happen, right? So proper management of agentic families can
- 13:19Speaker 1: reduce HR workloads by up to 40% is one of
- 13:23Speaker 1: the findings in the paper. Talk to us a little
- 13:26Speaker 1: bit more about that, what you mean by these agentic families,
- 13:30Speaker 1: and what you talked a minute ago about, you know, reducing, um,
- 13:36Speaker 1: HR budgets. Uh, if, if, if we don't like works, uh,
- 13:42Speaker 1: systemically together, like different HR functions, we're just going to
- 13:45Speaker 1: find ourselves having our budgets reduced versus being the enablement
- 13:51Speaker 1: right for, for the business. How, how do you, um,
- 13:56Speaker 1: how do you look at this finding along with, uh,
- 14:00Speaker 1: kind of what we've been talking about up to this point?
- 14:02Speaker 3: Well, the first part of it is how you frame
- 14:05Speaker 3: the AI investment. If you go to your boss and say, hey,
- 14:08Speaker 3: as a result of spending all this money on AI,
- 14:10Speaker 3: I'm going to cut our costs by 72%, well, you
- 14:13Speaker 3: just basically put yourself in a hole. You just did
- 14:17Speaker 3: the wrong kind of business case. On the other hand,
- 14:21Speaker 3: if you go and say, as a result of this
- 14:23Speaker 3: AI project, we're going to be able to hire people
- 14:26Speaker 3: 75% faster with two-thirds higher quality of hire, Therefore, we
- 14:32Speaker 3: can grow our business at a higher rate of speed
- 14:35Speaker 3: and do more things than we wanted to do before.
- 14:37Speaker 3: That's a different way of framing the problem. So that's
- 14:39Speaker 3: the first thing. The second thing is if you look
- 14:45Speaker 3: at the behavior of the HR people in the company today,
- 14:49Speaker 3: we've got maybe an instructional performance consultant, an instructional designer,
- 14:55Speaker 3: a content developer, an LMS administrator. And then you go
- 15:00Speaker 3: to recruiting. We've got an HR business partner, a sourcer,
- 15:06Speaker 3: a screener, an interview scheduler. We have all these job
- 15:10Speaker 3: titles that are kind of narrow and kind of tactical
- 15:15Speaker 3: sounding because what they make up is a step in
- 15:19Speaker 3: a process that somebody years ago broke down into a
- 15:22Speaker 3: job family. They took a process and said, let's just
- 15:26Speaker 3: take it apart like an industrial process.
- 15:27Speaker 2: And put it into a job family.
- 15:29Speaker 3: Well, if you sort of think about the agents, if
- 15:33Speaker 3: you had a specialized agent that did each one of
- 15:36Speaker 3: those individual steps and it got really, really, really good
- 15:40Speaker 3: at that step, that job is going to go away.
- 15:44Speaker 3: But the person who used to do that job is
- 15:48Speaker 3: going to realize, well, now all of the intellectual knowledge
- 15:52Speaker 3: that I have about that domain, I can train this
- 15:55Speaker 3: thing and I can make it smarter over time. And
- 15:58Speaker 3: once I have it to help me, I can do
- 16:01Speaker 3: five times as more work or at a higher level
- 16:04Speaker 3: than I was before. So all of these individual agents
- 16:08Speaker 3: sound like job replacements, but what they end up doing
- 16:11Speaker 3: is upgrading or giving people opportunities to work at what
- 16:15Speaker 3: are called the top of their license to add more value.
- 16:18Speaker 3: So like most recruiters, when you talk to a senior recruiter,
- 16:22Speaker 3: They don't like doing sourcing and writing Boolean searches all
- 16:25Speaker 3: day to find people. They want to talk to people
- 16:27Speaker 3: and really make sure they're the right fit and give
- 16:30Speaker 3: them a sense of power that they can improve their
- 16:34Speaker 3: careers by coming to work for your company.
- 16:36Speaker 2: They're sort of coaching salesy kind of people. That's what
- 16:39Speaker 2: they want to do.
- 16:41Speaker 3: So if they can offload some of that mundane work
- 16:43Speaker 3: to an agent, they can spend more time doing the
- 16:45Speaker 3: high-quality work that they really are good at. Same thing
- 16:48Speaker 3: is true with training. Same thing is true with business partners.
- 16:51Speaker 3: Employee experience people. I mean, all of these jobs. So
- 16:57Speaker 3: I think there will be jobs that are no longer
- 17:00Speaker 3: needed and some call centers will be smaller. But I'll
- 17:03Speaker 3: give you an example of where this didn't work out.
- 17:05Speaker 3: So a large bank that's been working on a big
- 17:07Speaker 3: global onboarding agent system for a while. Originally, when they
- 17:12Speaker 3: built the started the project, the goal was to eliminate
- 17:15Speaker 3: this big call center. that did global onboarding because they
- 17:18Speaker 3: do onboarding in 130 countries. So they have a whole
- 17:21Speaker 3: bunch of, you know, rather country-specific regulatory things to do.
- 17:27Speaker 3: So they started building the agent. They took all the
- 17:29Speaker 3: documentation they had. They started rolling it out. What they
- 17:32Speaker 3: found is that no matter how hard they tried, most
- 17:35Speaker 3: of the work these people were doing was not written down.
- 17:39Speaker 3: So there was no way to train the agent to
- 17:41Speaker 3: know what these people knew unless we interviewed all these
- 17:45Speaker 3: people and asked them hundreds of questions. about, well, what
- 17:48Speaker 3: do you do in this case?
- 17:49Speaker 2: What do you do in this case? What do you
- 17:50Speaker 2: do in this case? What do you do in this case?
- 17:52Speaker 3: And eventually they're, I think, getting to the point where
- 17:54Speaker 3: they're starting to see maybe headcount reduction. But by training
- 18:00Speaker 3: this thing into a super agent, they're going to have
- 18:03Speaker 3: scale and quality data about their workforce that they didn't
- 18:07Speaker 3: have before as a result. So maybe the original goal
- 18:11Speaker 3: of cutting the size of that call center is somewhat realized,
- 18:14Speaker 3: but the better goal is faster hiring, more accurate hiring,
- 18:19Speaker 3: more legally compliant hiring, and so forth. So most of
- 18:22Speaker 3: these projects seem to go in that direction.
- 18:26Speaker 1: Yeah, well, and I recall, I don't remember how many
- 18:29Speaker 1: months ago it was, but it was relatively recently, you know,
- 18:34Speaker 1: the story out of IKEA with, you know, their updates
- 18:39Speaker 1: with their call center, you know, they automated with AI
- 18:43Speaker 1: and And they had the choice that so many organizations have.
- 18:47Speaker 1: Do they lay off all of those people? Now they have,
- 18:51Speaker 1: you know, so many of the customer service calls being
- 18:55Speaker 1: handled by AI. Well, they did, you know, something totally different.
- 19:01Speaker 1: They upskilled their people. They essentially made them into kind
- 19:07Speaker 1: of like mini interior designers.
- 19:10Speaker 2: So that now... So now... So they can sell more stuff. Yeah.
- 19:16Speaker 1: So now when people call in, they go through the
- 19:19Speaker 1: initial screening with the AI. They get passed along to
- 19:21Speaker 1: this kind of mini interior designer who can provide better
- 19:26Speaker 1: customer service, who actually can sell more stuff. They actually
- 19:29Speaker 1: saw an increase in sales, quite a dramatic increase in sales.
- 19:34Speaker 1: They increased their business. They didn't lay people off. It
- 19:39Speaker 1: enabled them to grow, right? And it wasn't a reduction.
- 19:45Speaker 1: And that's exactly what you were talking about earlier. So
- 19:48Speaker 1: it's a different mindset, right? Then I think it's different
- 19:51Speaker 1: than when you have like an Oracle. That's just like
- 19:54Speaker 1: we were going to lay off 40 percent. We're going
- 19:56Speaker 1: to send out these emails at 6 a.m. and just
- 19:59Speaker 1: let go of people like that's a totally different way
- 20:01Speaker 1: of thinking about your workforce.
- 20:03Speaker 2: Right.
- 20:04Speaker 1: Then, you know, like how.
- 20:06Speaker 3: Well, there are companies, I mean, honestly, I'm sure there
- 20:08Speaker 3: are companies that could afford to lose 40% of their
- 20:11Speaker 3: people and maybe that's just, you know, they just hired
- 20:13Speaker 3: too many people and they're.
- 20:14Speaker 2: Not managing well.
- 20:16Speaker 3: But, you know, my, you know, belief is that in
- 20:19Speaker 3: most cases, if you look at a project that looked
- 20:22Speaker 3: like a cost reduction project, but reframe it as a
- 20:26Speaker 3: customer intimacy growth project, you will find that the second
- 20:32Speaker 3: opportunity is sitting in front of you, but you weren't
- 20:34Speaker 3: thinking about it that way. And I know in the
- 20:37Speaker 3: case of HR people, people don't go into HR to
- 20:40Speaker 3: do administrative work.
- 20:41Speaker 2: They don't like doing it. Nobody likes doing it.
- 20:44Speaker 3: They go into HR to help people and to help
- 20:46Speaker 3: the company grow. So this tends to be the heart
- 20:50Speaker 3: and minds of what HR people have been trying to
- 20:52Speaker 3: do their whole careers. And we're trying to make it
- 20:54Speaker 3: easier for them to do that.
- 20:56Speaker 1: Yeah, yeah, absolutely. So we've been talking about it all along,
- 21:00Speaker 1: but the super agent architecture. So again, we're taking individual
- 21:05Speaker 1: AI agents. They're being grouped into these eight functional super
- 21:09Speaker 1: agent families. So the whole idea here is that you're
- 21:12Speaker 1: going to have these HR domains where they're grouped together functionally,
- 21:18Speaker 1: they're working together. So you're not having these chaotic, you know,
- 21:23Speaker 1: individual agents off doing their, their own unmanaged kind of
- 21:27Speaker 1: free for all their own, their own thing.
- 21:29Speaker 3: Well, if you, yeah, the, The example that I think
- 21:32Speaker 3: is like if you compared SAP strategy versus Workday, SAP
- 21:36Speaker 3: is building hundreds of agents. Workday built 17. Workday basically
- 21:40Speaker 3: built 17 super agents. SAP built 80 or 90 individual agents.
- 21:46Speaker 3: I'm sure they're both going to end up in the
- 21:47Speaker 3: same place. But so you go to an agent and
- 21:51Speaker 3: you say, I need to grow my team to do
- 21:55Speaker 3: more work, to sell more stuff, write more code, whatever.
- 21:58Speaker 1: Yeah.
- 22:00Speaker 3: The agent doesn't know what the word grow means. Does
- 22:04Speaker 3: it mean we're going to make them taller? Does it
- 22:08Speaker 3: mean we're going to have more of them? Does it
- 22:10Speaker 3: mean we're going to improve the skills of them? We
- 22:12Speaker 3: need to teach the agent to think broader. So the
- 22:15Speaker 3: idea of the super agent is the super agent is
- 22:18Speaker 3: the place where you add what Workday calls it the
- 22:21Speaker 3: decision intelligence agent. You give it bigger questions and let
- 22:25Speaker 3: it go to the specialists to give you scenarios for answers.
- 22:30Speaker 3: because most of the HR problems that I run into
- 22:32Speaker 3: have many, many possible solutions. I mean, every HR issue
- 22:37Speaker 3: has dozens of potential solutions. We can fire the person.
- 22:41Speaker 3: We can give them a coaching plan. We can train them.
- 22:45Speaker 3: We can move them to a new job. We can
- 22:46Speaker 3: give them a new manager. We can pay them more money.
- 22:49Speaker 3: We can pay them less money. I mean, you know,
- 22:50Speaker 3: come up with 50 ideas on how to get, make
- 22:53Speaker 3: this stuff work better. Um, The agent should be able
- 22:57Speaker 3: to go and get the information from other agents on
- 23:00Speaker 3: what the scenarios look like and then give you an
- 23:03Speaker 3: intelligent set of recommendations on what the right solution would be.
- 23:07Speaker 3: And the more these things work through a company's evolution
- 23:11Speaker 3: over time, the smarter they are that in this case,
- 23:16Speaker 3: a redeployment is usually a better solution than hiring from
- 23:19Speaker 3: the outside because the last 10 times we hired from
- 23:22Speaker 3: the outside, eight of the people didn't work out and
- 23:24Speaker 3: we had to hire somebody else So whereas, you know,
- 23:27Speaker 3: that might be in the head of a HR business
- 23:29Speaker 3: partner or it might not. So a human may not
- 23:33Speaker 3: have that insight. So there's a lot of, I mean,
- 23:36Speaker 3: I think a lot of HR is sort of art
- 23:38Speaker 3: and science mixed together and it's probably too much art
- 23:41Speaker 3: because we don't have enough science.
- 23:44Speaker 1: The alchemy of HR, right?
- 23:48Speaker 3: So I mean, and I've seen this in recruiting and
- 23:50Speaker 3: recruiting because that's where the most mature AI has been implemented.
- 23:54Speaker 3: And recruiting is, These recruiting systems are very smart about
- 23:59Speaker 3: who's going to be successful. Like Delta Airlines has been
- 24:02Speaker 3: running a bunch of AI in their recruiting. And they
- 24:04Speaker 3: told me the other day that the most successful flight
- 24:08Speaker 3: attendants are not coming from customer service or hospitality jobs.
- 24:13Speaker 2: They were schoolteachers.
- 24:15Speaker 1: Wow.
- 24:17Speaker 2: Because I don't know exactly why, but you could guess.
- 24:19Speaker 3: School teachers are kind of good at taking care of
- 24:22Speaker 3: people that don't behave and being nice.
- 24:24Speaker 2: To people and listening and teaching and being patient.
- 24:28Speaker 3: That kind of personality would be great to have somebody
- 24:31Speaker 3: on a plane with you. But that may not have
- 24:34Speaker 3: been obvious from the skills model that some IO psychologists did.
- 24:37Speaker 1: Yeah, super interesting. Well, Josh, this has been a great conversation.
- 24:44Speaker 1: I know we've only scratched the surface. There's so much
- 24:46Speaker 1: more In the report and in the initiative all the
- 24:50Speaker 1: way around, I encourage people to check that out. As
- 24:53Speaker 1: we wrap things up for today, I just wanted to
- 24:55Speaker 1: give you a chance to share with the audience how
- 24:56Speaker 1: they can connect with you, find out more about your team,
- 24:59Speaker 1: more about the report, and give us the final word
- 25:02Speaker 1: for today.
- 25:03Speaker 3: Well, almost everything we do is on joshperson.com, J-O-S-H-P-E-R-S-N.com.
- 25:09Speaker 2: Or our Galileo agent.
- 25:11Speaker 3: Our Galileo agent has everything we've ever done in it.
- 25:15Speaker 3: It's linked from that website. But for people that really
- 25:19Speaker 3: want to dig into this more, Jonathan, we launched an
- 25:22Speaker 3: HR certification program called GHRE, Global HR Excellence Certification. And
- 25:28Speaker 3: so take a look at that on our website. It's
- 25:30Speaker 3: a 12-week in-depth sort of master's degree where you get
- 25:35Speaker 3: your hands on using this, building agents, implementing systemic HR
- 25:39Speaker 3: and real case studies. And that's another way for people
- 25:42Speaker 3: to learn about this. So those are some of the
- 25:43Speaker 3: things that that we can offer.
- 25:47Speaker 1: Wonderful. Again, Josh, thank you so much. It's been a pleasure.
- 25:50Speaker 1: I encourage the audience to reach out, get connected, check
- 25:52Speaker 1: out the website, check out the Academy, check out Galileo,
- 25:56Speaker 1: all the things. And as always, I hope everyone can
- 25:59Speaker 1: stay healthy and safe, that you can find meaning and
- 26:01Speaker 1: purpose at work each and every day. And I hope
- 26:03Speaker 1: you all have a great week.