Why We're Witnessing The Biggest Transformation of HR in Decades
In this talk, Josh Bersin explains how the HR profession has become increasingly complex, layered with strategic initiatives on top of traditional administrative duties, and how its federated, global structure adds further overhead. He argues that traditional HR technology has largely become sophisticated record‑keeping and workflow systems that still require massive human administration. Bersin predicts a major shift in 2026 driven by AI agents and "systemic HR" – a suite of AI‑powered tools (embodied in the Galileo platform and the Systemic HR AI Blueprint) that will automate 30‑40% of HR roles and processes. He outlines a two‑dimensional grid showing which functional areas will be automated first, and stresses that automation is aimed at improving quality, speed, and business value rather than merely cutting jobs. The transformation will move HR from tactical, administrative work toward "full‑stack" strategic work, increase the value of remaining roles, and require new skills in building, stitching, and managing AI agents. Bersin highlights the importance of AI accuracy, explainability, and trust, cites examples such as Boeing and a partnership with Findem, and describes how the Imperatives research, Galileo Learn courses, and a 11‑step blueprint will help organizations and individual HR professionals navigate the change. He calls on CHROs and HR leaders to engage with the Blueprint workshops and the CHRO council to shape the future of the profession.
Chapters
You know, there's been a long discussion for many decades about what HR is supposed to do in companies and how it's supposed to do it. And this debate usually centers on HR's ability to add value in a strategic way, counterbalanced against HR's role as the administrator, the police, or the enforcer of various HR or human capital or legal practices. And we've studied this for many years. We have a whole operating model. Built around this in a maturity model. But what we've basically found is that this is a complex profession that has become ever more complex every year, and it keeps getting more complex, driven by economic and social and business trends. And there are well over 30 to 40 million people in the world who work in this profession, taking on jobs of which there are more than 250, perhaps 300 different job titles in this one domain of business. And you could argue that that makes sense because it's a very important part of a company, but you could also make an argument that it's a lot of overhead and really gets in the way of companies doing what they're trying to do. So we can have— so short the philosophical debates, let's describe why I think 2026 is the birth of a massive, massive transformation of this profession and of the HR organizations that many of you work in. Now, first of all, the history lesson is very simple that If you look at our, the webinar I recently did, and I'll send you a link here, the profession has grown in context of the laws, rules, and social norms and business practices of the companies we serve. And as HR gets better at new things, it never lets go of the old things. So all of the basic stuff on getting people hired and getting them to comply with legal regulations and having legal and defensible practices for interviewing, et cetera. Are still there, as well as the services we want to provide to employees to help them manage their 401(k)s or their vacation balances or their work schedules. That's all still there, but all these advanced ideas like social networking, leadership development, leadership models, coaching, self-directed learning, career pathways, et cetera, have been layered on top of that. So the big picture of HR is finding a way to do the new high-value things and automate or eliminate the tactical things that get get in the way. Because if you don't do that, you end up with a lot of administrative people doing administrative work, costing the company a lot of money and slowing everything down. And most HR departments are federated in their organization. In other words, there's an HR group for the business unit in Australia. There's an HR group for the business unit in Switzerland. There's an HR group for the sales team. There's an HR group for the nurses versus the administrative staff. And these HR groups are led by what are called HR business partners, which is a terrible name, but it's well known. And these HR business partners are like little mini heads of HR. Some of them do very bureaucratic administrative stuff and some of them do very strategic stuff. And so this federated model in and of itself, which mimics the model of the company, creates even more complexity because everything you decide to do has to be localized to the workforce or the geography or the region or the city or country or municipality of the people working in the company. So you can't have one policy for everybody. It just doesn't work if you're a bigger, more global company, and almost every company is these days. So there's that. And then there's this theoretical idea that HR technology is going to automate all this. Now I've been doing this for a long time. And when I first got involved in HR tech in the early 2000s, we were automating or trying to automate recruiting, training, and general HR staffing stuff. Never really worked that well. Many, many years, 3 decades in my case of attempts, but what we really did is create all sorts of record-keeping systems. So the HR technology that's been developed for the most part has been very, very fancy workflow and record-keeping systems. You put the person's name in, you put their location, you put in their date of birth and their job level and their title, and the system does its thing, understanding who this person is and where they fit in the company. And you can run reports and analyses to understand this person's performance, role, mobility, et cetera. And over the years, these recordkeeping systems have gotten more sophisticated. They're called HCM systems. You guys know this. They manage payroll. They give the employee the opportunity to change their shift. And they have also tried to automate performance management, feedback, surveys, et cetera. The more human parts of HR, the evaluation and rating and ranking part, it hasn't done very well. There have been lots and lots and lots of performance management tools. But even if the tool, the system, the software encourages people to check in or talk to their managers once a quarter or once a month or whatever, that doesn't actually do the work of talking to them. It is up to the manager to talk to them. And so these recordkeeping systems were more like scorekeeping tools than actual solutions to problems, but we needed them. And the more complex and big your company is, the more of them you need. So these Bigger vendors like Workday and SuccessFactors and Oracle HCM and UKG and Dayforce and ADP have lots and lots and lots of business rules in them. And then you end up with a large body of HR professionals managing and tuning and administering these systems. Go into the Workday or the SAP user community groups, and you'll see there are thousands and thousands, maybe tens of thousands of questions and suggestions and challenges that people have managing these tools. And the technology side of this is tens of billions of dollars, if not more. And then the training part of it, which sort of sits on the side, is hundreds of billions of dollars. Because in addition to all this administrative work, we have to train people. So there's content and training and intellectual property in every possible domain of a company. Every single thing you do, there needs to be training or information about it. So people can learn about it and understand what is right and what is wrong, what's good, what's bad. So it's very interesting and fascinating for me. And many of you obviously love it. And some of you probably are just at different stages of learning about it. Well, this year we believe, and I'm pretty clued into this, that because AI is moving into the world of agents and super agents, and you can read about this in our Imperatives research, a lot of this is going to get wiped out and changed. Not overnight, not in one year, but it is going to happen. And these process-based database systems or processes we use for recruiting, for managing people, for mobility, for pay, for benefits, for promotion, for development are going to be replaced by agent or AI versions or new releases or new products. And many of these new systems and tools will automate away a lot of the back office work. We have an extensive model of HR in Galileo, our AI assistant. And we went into Galileo over the last 2 or 3 months and we asked Galileo to help us. And Galileo, after many, many efforts of working on this, concluded that between 30% and 40% of the job roles in HR will go away. I'm not saying the people will go away, but the roles. Now, not everybody does everything at the same time. So no one company is going to do all of this, but we have a 2-dimensional grid, which will show you which functional areas are most likely to get automated the fastest and where your opportunities are the greatest. And by the way, the benefit of automation is not getting rid of jobs. That's an artifact of the process, but the benefit is making the process easier and more streamlined and more automated and more autonomous for the employee and for the company. So we're not automating stuff with the purpose of getting rid of jobs. We're automating stuff with the purpose of making the process better or getting better data about it. So what are the jobs? What are the roles? What are the areas? I won't bother you with the details on this podcast, but what we've done is we've built Systemic HR AI Blueprint that shows you these things. And there are around 120 or 130 independent processes, each of which will probably end up with AI agents, and then a family of super agent processes that sit on top of them. Onboarding is a good example. And we have a good example of onboarding that I went through in the webinar I did this week. But a lot of this, you're not going to go buying stuff. You're going to be building stuff because you don't want to have 140 vendors each offering you independent agents. And you may not want to use the agents that come from your HCM provider because they may not be as independent or maybe creative as the outside agents, because they're mostly built to automate the things that the system already does. So there's going to be a really interesting period of time here where some of the new jobs that are created are the people building these agents and stitching these agents together and architecting these agents to automate things. Then 30 or 40% of people whose jobs change are going to be looking for new things to do. Now, we're not going to be looking, they're going to be assigned and there will be jobs managing the agents. But more importantly than that, taking, than taking care of the tools, is talking to people, talking to candidates, talking to managers, talking to employees, getting closer to the business units that you work in to add value using your HR expertise. So the general trend is up and away from administrative work towards more strategic work, which we call full-stack HR. I've been talking about full-stack HR for more than 10 years. And those of you that have been in HR for a while, or maybe that are new, are going to see that this profession is going to get better and more interesting and more creative and more interconnected and multidisciplinary as well. As far as the company itself, the amount of money spent on HR administration may go down. It may not, because you got to pay for the agents too. The agents are not free and we're paying for agents by the token now. So they're not as sort of license-free as they used to be. But, you know, ideally a lot of bigger companies would like to reduce the number of staff in HR because what happens in many companies, and this is very very common is if the HR leadership is not looking at the HR operation carefully, they end up with a bunch of people in all these business units doing administrative work, replicating work that other people are doing. And pretty soon you look around and we've got too many, a whole bunch of HR people. Now, the metrics on here are interesting. The average benchmark for the number of employees to HR ratio is around 100 and has been this for a while. It varies a lot. We've talked to companies that are 50. Which means they have twice as many HR people. We've talked to companies that are 150, which means you have a third less HR people, but based on the research we've been doing with Galileo and the simulations and model for systemic HR, it's going to go up to 200 to 300 to 400. So we're going to need fewer people per unit of work or unit of employee. But, but that doesn't mean the number of people in the HR profession is going to go down. The 25, 30, 40 million people who do HR stuff, Listen, if you're an interview scheduler, yeah, I mean, you're not really doing HR anyway. You're just doing scheduling. And if you're a recruitment coordinator, you're just doing coordination work and you could go coordinate something else. But for those of you that are really HR professionals, I don't think the number is going to go down that much. It might go down a little, but actually the opposite has happened is that because of AI, the number of people in HR has gone up a lot. In fact, the HR job postings have gone up by almost 60% in the last 3 or 4 years, higher than almost any other job because of all the human capital issues we have. Have around AI. But anyway, that's not really the goal. The goal is to let the company move faster and more effectively to do what it's trying to do. If you're Boeing, and I understand, I was just reading Boeing shipped almost 1,000 airplanes last year. Can you imagine the number of things that have to happen to ship 700, 800, 900 airplanes in a year? The engineering, the manufacturing, the packaging, the shipping, the customer training, the customer enablement, the supply chain management, all of those things that make an airplane have human components to them. And you know this in Boeing's case because the culture at Boeing, you know, went sideways and that's come out in Congress. So, so I don't know the HR function of Boeing super well. I know a little bit about it. They have a lot of human capital stuff to do. They have to train people. They have to build a culture of safety. They have to move people from role to role. They have to hire people. They have to retain people. They have to identify new technologies and new tools. They need to manage their supply chain. All of those have human capital elements, all of that stuff. So if you automated a whole bunch of the HR department, maybe Boeing would have shipped 1,100 or 1,200 airplanes. Because other things would have gone faster and more effectively. So that's where we're going here. We're not doing this for the purpose of shrinking the administrative budget of HR. Now I'm, you know, not trying to over-glorify what we do. We are a cost function. We add value in many, many ways, but we don't directly contribute to revenue. But even that is not necessarily true in the world of AI. If you're a restaurant chain or a food service company of some kind, and you can hire staff faster to fill demand or schedule staff faster, as I talked about in the article I wrote about UKG, to meet demand, your company's going to increase revenues. So these AI systems that we're in the middle of starting to build are not cost reduction systems. They could be revenue generation tools, or they could improve the quality of hire, or they can improve the quality of skills or leadership, et cetera. So at a high level, this 30-40% shift away from tactical work is not a cost reduction effort. It's a value improvement, value increase effort. And you're going to all see that when you do it. And so I'm not trying to create an inflammatory headline when I talk about jobs and roles changing, not at all. I think most of you who have administrative roles like this probably don't enjoy them that much. And frankly, would like to do more high-value work and make more money. And by the way, I think the average salaries in HR are going to go up because the work that's left as these agents come to pass is higher value work. By the way, managing the agents takes work too. And that's a high value job too, because an agent can move very fast. And that leads me to another topic. In this AI-powered systemic HR world, when we have more automated agents, just like we would in a manufacturing line, accuracy and quality is even more important than ever. It's like if you're a, you're in a manufacturing plant at say Tesla or any other company, and one of the machines is off by 1 millimeter in its welding or its bolting, or its torque is a little bit off. You may end up with 100 cars that have a dent that the customer finds later before you even know it happened because it happened so fast. Well, think about a recruiting tool that inadvertently brings in the wrong candidates or gives somebody the wrong pay or has a poor candidate experience and then run it at scale for a high-volume recruiting company. And all of a sudden, not only is it great, but it could be really bad too. So the quality of AI in HR actually has to be very, very high. I mean, I'm not saying it doesn't have to be high in other areas, maybe not as high in marketing because it's a little more amorphous, but we have to make sure things happen really accurately and the best possible way, because they're going to happen at scale and you're not going to have humans watching over every single transaction. Now, in the human elements of AI, that happens too. You know, some recruiters are good and they bring in great candidates. Some recruiters aren't good, so they don't bring in great candidates. Some L&D people are great and they end up with highly engaged, highly skilled people. Some L&D people aren't so great, so you don't end up with quite the same experience. Well, we have the same issue here. In fact, I just had a really fascinating meeting with a vendor. I'm going to be telling you a lot about them called Findem. And we're talking about this is that the explainability and accuracy and authoritative trust of these systems is really, really important because the decisions that these systems are making are very large decisions if they operate at scale. So the work of managing the AI and the content and the agents is very high-value work too. So, so generally speaking, we're talking about a, not just upskilling, but upgrading of many, many of these 250 to 300 job titles in HR. Now we spend a lot of time on this with a lot of clients, and we're really going to be spending a lot of time this year on the blueprint and on helping companies go through this transformation. All of this information is embedded in the Imperatives research we launched this week. The Imperatives research is embedded into Galileo. We are not shipping it as a PDF. Unless you license Galileo, because the PDF doesn't give you the actionability that we want you to have. We have built several courses in Galileo Learn to let you step through this and get to know what these 11 steps are. You can literally go into Galileo now, effectively this week, and you can tell it about your company and it will diagnose your opportunities for AI amongst these agents. And as we finalize and really develop more details into the blueprint and we're meeting with clients now. This will all go into Galileo and you will be able to use Galileo to go through this exercise yourself. Plus Galileo is a development tool. You can ask Galileo to give you a development plan. You can give Galileo your resume, your bio, your LinkedIn link, whatever you want to do and say, give me a 6-month development plan to become up to speed on all this new agentic work and superagent work that's going on. And it'll show you what to do, not only using resources from us, but using resources of your own. So we're really doing everything in our power to facilitate this transformation and empower you as a professional, as a leader, as an individual to really be a part of this. We do not want you to feel threatened by this. You shouldn't feel threatened by it, but don't ignore it because it's maybe the biggest thing that's happened since the beginning of the internet and the beginning of Cloud-based HR. And for those of you that are VPs and CHROs, we're more than happy to talk you through this. We have a CHRO council of members, and we're going to be going around the world doing workshops at all of the major conferences. We are more than happy to invite you to the Blueprint workshops. If you would like to be part of the Blueprint workshops, reach out to me or any one of the people in our company around the world. Our company, by the way, is growing not quickly, but a lot, but a little, and we are now bigger. We have new people in Europe. We have new people in the Middle East, new people in the Far East. I'll be on the road a lot this year. So we really want to help each one of you listening to this podcast understand and be a part of this transformation. And I think 2026, as I've talked about a lot this week, is the year of enterprise AI. And we're going to look back on time 4 years from now when 2030 rolls around, we're going to say, wow, things are really different. You know, I have a very interesting article I'm working on about the future of the work experience itself that I will share later when I'm a little further along with it. And the way I built the piece is I went back 50 years and I looked at all the things that have happened since the 1970s when I got out of college or even a little bit before that. So, so we're in the beginning of a really fascinating change in this profession and in this industry and in our companies. And I want all of you to really be an active part of it and a well-informed part of it. That's what we do. That's what we're here to do. So, uh, stay tuned for more. I'll write an article about this to make it clear. You can go to our Imperatives page, joshberson/imperatives on the internet, register. You can see a replay of the webinar I did this week and just get Galileo and you'll have access to all of this IP in any form that's best for you. Have a great weekend, everybody. Talk to you next week.
More from Josh
Goodbye Boring Courses, Hello AI Learning | Josh Bersin on the Future of HR and L&D
Irresistible Keynote: The Post-Industrial Age, The New Economy For Workers, Companies, and HR teams.