In this episode of the HR Digitalisation Podcast, Anna Carlsson and Emira Blomberg reflect on the major shifts currently shaping HR. The conversation takes its starting point in the HR Technology Conference in Las Vegas – the world’s largest HR Tech event – where AI is no longer a thing of the future but a reality for many organisations today.
Anna shares concrete examples of how companies are already transforming their HR processes through AI, from automating case management to revolutionising recruitment and learning. The discussion also touches on the challenges that come with these changes: change management, the need for new skills, risks related to bias and data quality, and how HR can evolve new ways of working in step with technology.
A particular focus is placed on what it means to become AI first – when HR organisations restructure to let AI handle what it does best, while people are freed up for more strategic and value-creating roles.
The episode also highlights the future of the podcast itself, as Anna shares insights about her upcoming report for strategic HR work and announces that the HR Digitalisation Podcast is entering a new phase after six years in its current form.
Listen to gain fresh perspectives, real-world cases, and insights that help you understand and navigate HR’s ongoing transformation.
Note: This episode is in Swedish. A translated transcript is available below.
Transcript:
Anna Carlsson: Hello all wonderful listeners of HR Digitaliseringspodden who have followed me for so long. Today is a bit of a special day when you hear this. Or maybe you’ve read my newsletter or something on LinkedIn. I am in the process of reshaping my life. Yes, it’s me, Anna Carlsson, who runs HR Digitaliseringspodden through the company HR Digi. In line with the changes in the world around us, and my preaching about the importance of thinking new and doing things differently, I am now doing the same myself. Today’s episode is about the future, which is already the present for some companies in the U.S. There they are restructuring their organizations to take advantage of AI in the best possible way. My personal change in how I run my business is, of course, influenced by the opportunities given by our digital world of AI and my desire to practice what I preach. So, what am I talking about then?
Well, this is the last episode in this form of HR Digitaliseringspodden. During the six years I’ve been running the podcast, I’ve had a very ambitious approach to producing a high-quality product. But for various reasons, the podcast will not sound the same or be released in the same way as before. For now, it will probably be a small break, or there might be an episode, but then it will only be me speaking straight into the mic. Maybe no intro music at all. Most likely no guests, or only occasionally. Just straight to the point and shorter. I’m inspired by Josh Bersin, who does that in his episodes. That means the quality will be different, but the speed from recording to delivery to you can increase. The other option that may become reality is that someone who listens wants to take over running the podcast. And by that, I mean project management, planning, and execution. And I would just join to record the episodes. If you’re interested in that, get in touch at anna@hrdigi.se.
The second thing I want to share before the episode starts is that I usually record a well-prepared trend episode with lots of information about the present and the future, plus tips and advice for the coming year. I now have more information than ever before, but it’s so much that it won’t fit in one episode. That’s why I’m focusing on Thought Leadership this time. What are those at the forefront doing? Instead, I’ve decided to put together a report that will be available from around mid-November.
Here I’ll gather information about everything from what you need to consider in your business, what the future looks like, which providers are out there, background material, statistics, references to reports, and more. Material you may need in your strategic planning or decision-making processes. You can already sign up if you’re interested. The link is right now in the text on your podcast player, or you can find it on my website under “News,” that is hrdigi.se. Of course, you can also hire me as an advisor in your strategic work or for inspirational lectures, training, or similar, and then access the material that way instead. So, today’s episode is about the future. A future that can look very different now with AI availability. A future that can look very different now with AI’s possibilities. A year ago, none of us really knew where this would lead, but now there are solutions, organizations that have carried out the transformation, and consultants like myself who have had the chance to lead this type of change work. All this was discussed at HR Tech in Las Vegas, which I attended at the end of September. Today’s episode shares my insights from there in conversation together with Emira Blomberg. So, as I said, going forward the podcast will come in simpler formats with only me, unless someone else wants to take over the project. And don’t forget to sign up for my upcoming report.
Anna Carlsson: Hi Emira!
Emira Blomberg: Hi Anna! I realize that it’s me who should be welcoming you, because today it’s you who will be talking.
Anna Carlsson: Yes, though I think we’ll both be talking. I mean, at least about my reflections that we’ll touch on.
Emira Blomberg: Yes, welcome home from Las Vegas!
Anna Carlsson: Thank you so much, yes, it was a fantastic trip. But I’m a bit tired today. We’re recording this on Monday. I got home on Saturday and I was so cocky. I met Johannes at the airport. Johannes Sundlo. We ran into each other and chatted a bit. I think it was to him that I said: “But I usually don’t get jet lag.”
Emira Blomberg: No.
Anna Carlsson: And now I have the worst of it.
Emira Blomberg: Yes.
Anna Carlsson: I don’t know what it is. I can’t sleep.
Emira Blomberg: But it’s still the West Coast. Not exactly, but it’s on the west side. What’s the time difference?
Anna Carlsson: 9 hours.
Emira Blomberg: Ah. No wonder, then maybe you weren’t that cocky. But at least you look fresh.
Anna Carlsson: Yes.
Emira Blomberg: Because exactly – now you did a bit of name-dropping there with Johannes Sundlo. What have you been doing? Where have you been?
Emira Blomberg: Well, we’ve been at the HR Technology Conference and Exposition in Las Vegas, the world’s largest HR Tech event. Absolutely fantastic. And this year I actually thought it was – for me, it was even more. There was something extra this year, I felt.
Emira Blomberg: Why’s that?
Anna Carlsson: So much so that we’ve gone from AI being about learning AI to there being lots who have already done a lot.
Emira Blomberg: Oh really.
Anna Carlsson: Well, maybe not lots, but those who were there and who gave presentations had done very exciting things. Plus, there are so many entrepreneurs who have done exciting things. So I feel filled up. Very filled up.
Emira Blomberg: Yes, yes, how fun – because we were at HR Tech with the same group that organizes HR Tech in Amsterdam earlier this spring. Right, it’s September now. In spring we were there, both you and I, and then we concluded that there was a lot of talk about AI, but very few who could show things they had actually done. And now you say you’ve got lots of juicy examples of those things? Wonderful!
Anna Carlsson: Yes, that’s right. Yes, it’s the same. It’s the same organization. It’s a separate team that runs the Europe event, so they were there too. They talked to providers. They want to make sure that people come to the Europe event and showcase what they’re doing so we’ll get the same juicy things here in spring. I really hope so. It’s more difficult because not as many organizations in Europe have done things yet. But we need to talk about one thing before we get into the juicy stuff. This might be the last time we sit here.
Emira Blomberg: So, bittersweet then?
Anna Carlsson: Yes, and in a way it’s a relief. Yes, for everyone listening: I’ve decided that I need a break from the podcast. So right now as we’re recording, I think it’s been a year—sorry, six years. Not one year. Six years since I released the first episode.
Emira Blomberg: And it’s been a year since I joined.
Anna Carlsson: Yes.
Emira Blomberg: But you had already been running it for five years before that.
Anna Carlsson: And suddenly I got the feeling that it’s time for something new. I’ve shared a bit on LinkedIn that I’m making changes in life and focus, and it feels good. I don’t quite have the energy to run this anymore. It’s a project, and right now the economy isn’t the best. So the sponsors haven’t really shown up, and I’ve lost a bit of energy for this area. So now it will be a break for a while, and we’ll see if it comes back—or maybe someone else wants to reach out and say they want to take over. Or, you probably shouldn’t be alone—you should probably have a company behind you, because it really is a project.
Emira Blomberg: Because you’ve built both a brand here and also quite a loyal listener base, so it would be nice if someone wanted to take over.
Anna Carlsson: Yes, it really would, and I could absolutely imagine recording episodes. That’s the small part of this project.
Emira Blomberg: The part I do. I just show up here.
Anna Carlsson: And record.
Emira Blomberg: And record. Exactly. No, but that’s been the luxury of being the sidekick. I’m so happy we got a year together, and with so many exciting and interesting guests who have been here. And you learn so much, just from doing it—and also from listening. At least we hope so.
Anna Carlsson: Yes, of course. Yes, I think so. Since the listener base is so loyal, it must be that people are getting something out of this.
Emira Blomberg: Yes, exactly. So, all of you who now want Anna to continue with the podcast—no, I’m just kidding.
Anna Carlsson: Then you’ll need to make sure there’s someone who wants to take on the project management. The planning of guests, and then I’ll be happy to continue. But in this format, I just feel that right now I need to focus on other things.
Emira Blomberg: Yes, exactly. But should we dive into this HR Tech Las Vegas thing now? I’ve had major FOMO about it. I’ve wanted to go there for a long time. I’ve been to Las Vegas once as a private person, and first of all, the whole image of it is grand, to say the least. I loved Vegas. Some think Vegas is tacky and crazy, but I just see lots of beautiful neon and the commerce. What can you say? The modern world. It’s ugly. But it’s also incredibly beautiful in its ugliness.
Anna Carlsson: In its ugliness. I’ve been there, I don’t know how many times now—maybe up to 15 times. So I really am familiar. But this time I had a friend with me, so I got to experience some things—you know, take her out and show her around. That was fun.
Emira Blomberg: But it’s also, shall we say, isn’t it kind of the Western world’s most comprehensive—like it really takes a holistic approach to HR, with a focus on the latest technology?
Anna Carlsson: Yes, it’s really about the possibilities of technology. But what I find interesting is that the first time I went, the focus was on the technology itself. Now it’s that technology is HR. You have to recognize that. That’s the shift that’s happened since the first time I went. In 2018, I went to HR Tech for the first time. Before that, I had been to other events in Las Vegas—that’s why I’ve been so many times. But in 2018, it was more about meeting exciting vendors: “What can you do?” But now, if you look at what’s happening in the world, at the megatrends, the absolute biggest, most impactful thing for us, I’d say, is AI and digitalization. And AI is digitalization—it’s just a new form. But it’s what really affects us as people—as employees, as HR professionals. That’s how it feels. I’m going to Unleash too, after this episode is released, and there the focus is broader HR. But here it’s really about what the technology enables and the experiences. Trendspotting. Vendors. I think there were 430 vendors on the floor. That’s a lot. Very interesting, and you really get an insight into what it might hopefully look like here in Sweden in ten years. Hopefully. Others might think it’s scary that technology will play such a big role in HR.
Emira Blomberg: But we had Nina Rapp here before the summer, and that episode was released fairly recently. Then there was a lot of focus on AI—as AI and the technology itself. I think what you just said was so good: that AI is digitalization. I think I even said at some point in the podcast that we won’t be talking about “AI” for much longer, or it will become irrelevant—it’s really about the technology and what we can enable with it. But for now, we still need to call it AI. That’s just how it is.
Anna Carlsson: But I’d put it this way: there wasn’t a single session about how you personally use AI. No. That’s still where we are here, I feel—learning AI. But there, it was different. That’s why I got so inspired. And then I realized, it’s really scary now. Because when an organization has truly harnessed the power of AI and made fundamental changes in its organization—meaning, “We don’t do things the same way at all, we think completely differently”—it becomes something entirely new. AI becomes a natural part. But it’s a huge shift.
Emira Blomberg: Is there anything specific—when you say these things now—is there a particular session or something that pops into your head?
Anna Carlsson: Yes, my former employer.
Emira Blomberg: Yes.
Anna Carlsson: It was very interesting because I used to work at IBM, and there they decided—when I was still there, we got our first chatbot. It worked quite poorly. We got a chatbot to help us when we were writing travel expense reports, because it was so complicated. You could ask questions, and it pointed you to the right place where the answers were written. That was magical. No person had to search the manual themselves. That’s what you used to do back then—you got a document saying, “This is how you do it.” And then you had to figure it out yourself. Or ask a colleague or your boss for help. It wasn’t easy. But that was when IBM’s journey started. And here, their CHRO spoke—it was the first day’s keynote in the afternoon. And it’s been a journey: they’ve now become AI-first. We used to talk about mobile-first. Now it’s AI-first. I’ve got some numbers. They went from a traditional model with HR business partners—and I was part of that, I was one of the managers at IBM when I left. They had business partners, specialists, and a regular service center model, with complicated systems you had to figure out. One of the worst things was transferring employees—it always went wrong. You had to know codes and lots of things.
Now, today, 94% of all cases are handled automatically by AI agents. So, for example, if a person is being transferred, you just talk to your AI and it takes care of it, finds out the information, and does it correctly. So here we are. How does that feel?
Emira Blomberg: Yes, but I’m wondering—cases. What kind of cases? All cases?
Anna Carlsson: All HR cases. They have a service called Ask HR. They started experimenting with chat back in 2017–2018, when there was already some generative AI, but not much. It wasn’t anywhere near the level that came later. But they started testing with chat, and there were lots of different chatbots. They played around with it to test the technology. And then they decided, since they really needed to save money in HR—it was bleeding money—they needed to do something drastic. So they decided, now we’ll use AI. They started sending all cases to their chat. That way, they could learn what cases were being asked about and everything. And at first, it didn’t work. People had to step in and help in different ways. When they started, their engagement index was +32. When they began sending everything to AI, so that the AI could learn and they could build this environment, it dropped to -17. And gradually, they changed things and sent all cases over. As I said, 12 million interactions per year. 94% of cases are resolved entirely by the system without being sent on to HR staff. And this has led to a 40% reduction in HR support costs. But of course, that raises a lot of questions.
Emira Blomberg: But I think it becomes—we’ve talked about this before: what is HR really? HR is such a broad spectrum of responsibilities and tasks. Some things are more compliance-focused HR, or their own profession. Other things are about facilitating and supporting. And I think this really highlights that we have these generators and managers. Everyone who used to generate information—maybe that’s not needed anymore. Who knows. But it would be very interesting to see from the inside too. You said they were bleeding money—so I assume they’ve also let people go?
Anna Carlsson: Yes, that’s actually the point I wanted to get to—how did they do this then? What happened to the HR organization? I’ve seen it from the outside. I have colleagues who complained that they didn’t want to stay, because they didn’t want to be part of the new model. And that’s what happens in transformation—it doesn’t always work out well for everyone. What they did is they didn’t fire people outright. And at the event, many talked about how AI is sometimes used as an excuse to let people go. But often that’s not the case. They did reduce staff, but they did it through attrition, as people moved on. They didn’t hire replacements, so they reduced costs that way. It wasn’t a big program where people were laid off, but rather in other ways. Maybe in Sweden someone was affected, I don’t know—I’ve only listened to the head responsible for the whole area. The exact effects in each country, I don’t know. But what they did is that any cases the AI can’t solve become the challenges. Those lead to exciting conversations. Suddenly, an HR person has to step in, talk, and help solve the manager’s issue or whatever it might be. And that requires much broader competence. Before, people worked in silos: “I know Compensation and Benefits,” or “I know Talent Development.” And now suddenly it’s, “What challenges do managers have?” and then they staff specialists accordingly. Not siloed in the old model, but rather: how do managers really need support? That’s how they’ve built the support that exists behind the AI.
In large organizations, you usually have a service center—you call or send cases in. There’s first-line support, often fairly junior, who sort cases and send them to specialists. Always several levels. But here, all cases the AI can’t solve go directly to a competent HR professional. That’s the model.
Emira Blomberg: Exactly. Then I think it’s interesting that they trained this model by sending in all cases.
Anna Carlsson: And by calling them in.
Emira Blomberg: And they understand that it’s needed for the AI to learn. And in the beginning, the numbers dropped drastically. But then I wonder—do they do the same with the people? Because people also need new skill sets and training. Was that something they talked about? Because right now, there’s so much talk about how we train these models so they can take on tasks. But that also means, as you say, that we humans get new tasks. Was that discussed?
Anna Carlsson: Yes, it was in the same context that they talked about how they had put people together with those kinds of skills so that you can enrich each other, so if you have a narrower competence and need a broader, even broader, more general one, then you need to sit together with others and work together with others for knowledge transfer. But what she talked more about was actually how you also have to bring in juniors. Because that’s also something we’ve discussed a lot. What happens when AI arrives and we no longer have these junior roles? There they talked about them sitting together. They get to work on advanced cases as apprentices rather than having to do the—what shall we call it—administration? More routine-case activities, because those no longer exist. But they still have to learn, so she said we don’t see that disappearing in any way; rather, we still bring in those who come from school, and they then develop in that way instead.
Emira Blomberg: So we’re almost a bit back to how it was before.
Anna Carlsson: The apprenticeship period.
Emira Blomberg: Exactly. And hopefully perhaps a more sustainable, in some way, workforce planning structure as a result. Who knows, really. Because of course it’s been very difficult for companies to plan competence and staff when there are lots of people who are specialists in their areas and need to interact with many other people. And that the specialist knowledge is so human and doesn’t really sit in any structural capital. Of course that becomes very disruptive. But oh, how exciting! Okay, IBM—anything more you want to say about that particular thing that you…?
Anna Carlsson: Yes, because I think a lot about what we hold on to—how we’ve done things up until now. I have a lot—I’ve been thinking a lot about it, also while she was presenting. It’s so easy to want to continue doing things the way you’ve always done them. It’s built into us as humans. And this change—it’s tough. When you think about the poor metrics and everything you have to… You have to have a very clear target image if you’re going to make a change. You can’t have anything but a clear goal in front of you—“this, I know, will be good, and now it is good.” Then people might not have liked it all the way, and that’s the change curve you have to go through. And well—they’ve been at it since 2017, so that’s eight years now.
Emira Blomberg: Exactly. A target image is also a clear “why”—that they have to know why—and that there is some kind of sense of urgency. That’s also super important for a change effort to be successful.
Anna Carlsson: And in their case, they’d been given a requirement that “you must reduce costs.” I think there are many costs you could work on in organizations when you have access to data and can work more—well, insight-driven—rather than just relating to “this is how we do things.” And it’s tough. Breaking patterns. Really tough.
Emira Blomberg: Exactly.
Anna Carlsson: But there were more examples—one you might like. Recruitment has been a very, very big topic. I mean, if you think about which processes can be digitalized and agentified: there was a hotel in Las Vegas that had been built. They’re one of those fancy hotels that were to be built and needed to recruit people—2,000 people in three months—by using a lot of digital tools in this, i.e., AI tools. Agents that primarily handled booking interviews and going through practical matters. What was also discussed around the use of AI was that people have increasingly moved away from discussing whether you should review CVs, and instead look at other things where you really get impact from AI. There was a bit of a divide over whether people wanted to do that or not.
Emira Blomberg: Right, yes—because we’re following the Workday lawsuit with interest as well, of course. Speaking of AI and recruitment: for those who aren’t up to date, there’s an ongoing lawsuit against Workday in the U.S., where a man initially claimed that he was discriminated against because his application documents were not moved forward in the process due to his age. And normally you can’t sue the vendors; you sue the companies that use the tools. But this is—if it’s not the first time in history, then at least a historic case—where a vendor is actually the target of the lawsuit. Is that how you say it?
Anna Carlsson: Yes, exactly. Because there, it’s the tool itself that they believe has, so to speak, misjudged.
Emira Blomberg: Exactly.
Anna Carlsson: We’ll see. I don’t know what has happened now, if anything else has happened.
Emira Blomberg: No.
Anna Carlsson: I’ve written about it in the newsletter as well.
Emira Blomberg: Yes, exactly.
Anna Carlsson: But it is—it is interesting.
Emira Blomberg: You could imagine it’s scared some people off as well. I agree with you that a lot of focus has been on CVs—well, maybe not CVs—but also that there’s so much else that eats up time in a recruitment process.
Anna Carlsson: I have—this was the one—I was flipping through my computer while you were talking. There are some recruitment companies that have made recruitment very digital. They’ve marked: what is done in person, and what do you call it—human?
Emira Blomberg: Yes, exactly.
Anna Carlsson: In person. Career site, job search, send application, talent communities, candidate screening, interview scheduling—all of that is digitalized. In person: interview. The hiring decision is also personal, and then everything is digital again—send offer letter, background check, onboarding forms. And then we get to the personal bit again, which is day one.
Emira Blomberg: Yes.
Anna Carlsson: So there’s a lot that you can really automate, and then the AI can do an incredible amount.
Emira Blomberg: Yes, indeed—it’ll be really exciting to see how/if this takes off. But this case you just described—who were the drivers behind it?
Anna Carlsson: The recruitment one? It was someone who was building a hotel and needed to get staff for the hotel.
Emira Blomberg: So they built their own solution for it? Or did they have vendors? Or…?
Anna Carlsson: Yes, they had vendors. That’s also what’s exciting—I’ve seen so many interesting, real agent technologies. New vendors have emerged who—who have this, who build the flows completely from start to finish, that run themselves. And I visited a vendor where they work—if listeners know Lovable. You know Lovable—you know how it works, that you…
Emira Blomberg: “The last piece of software,” or whatever they say. That they build the last piece of software.
Anna Carlsson: Perhaps that’s what they say. I think of it as low code/no code, so you basically write “this is what I want to do,” and then the system—maybe you can get a system?
Emira Blomberg: They mean that no one will ever need to build a system again. Right? We noticed that. I think that’s their slogan. I shouldn’t—well, it’s the slogan I’ve seen floating around. I think it’s catchy.
Anna Carlsson: It is. And that’s exactly what I got to watch with one of these—one of these vendors said, “Now I’ll show you how it’s done.” And he just wrote, “I want to do this, and that data exists in that system,” and so on. Then the onboarding flow for that was built. It’s absolutely crazy and very interesting. And then it works. I mean, we should also be aware that you need integrations to the data that will be used. But what I also discussed with many of these vendors that have the new AI agent systems is that with that type of system, when it comes to GDPR, the data remains in the systems it uses; the agents do not become that risk. That also means you should be able to buy it from other types of vendors, because no data is stored. I think that’s pretty cool.
Emira Blomberg: Yes, it is—and we’ve also discussed this. The listeners of course know that I work at a software company. We’ve also talked about AI functionality, etc. But I’ve come to the conclusion that the future will consist of two different, more clearly categorized—almost two different types of systems. One is these agents, or user interfaces, or the way you interact with different data sources—that’s one category. The other category is data quality and data sources—because the AI will never be better than the data is. And I think that’s the other track to pursue. Everyone is running down the agent track now, in my view. But ensuring that there is a high-quality data source to pull from—that’s actually at least as important.
Anna Carlsson: And when everyone was standing there showing off their fancy tools at an event where I was a speaker, I talked about the importance of data quality. But the thing is—it was also interesting. I mean, the USA—how many years ahead can they be, 5–10 years? They don’t talk about data because it’s so natural that they’ve solved it. It’s a given that you have your data in place, that you have your systems in place. You have master data, you work with data, and data influences the business and is part of the strategic work. It’s so obvious that… it just wasn’t talked about.
Emira Blomberg: No.
Anna Carlsson: There were sessions about new tools that themselves have AI—AI chat, generative AI interfaces—so you can easily talk to your data, making it much easier to visualize and create stories behind your data. But it wasn’t like last year, when that was the new thing. Back then, people still talked a lot about data and the real importance of working with your data and the different lessons learned. But this time it was very, very little.
Emira Blomberg: But that’s also… speaking of Sundlo, whom you mentioned earlier—it was him at HR Tech Amsterdam in his talk who said “Map your processes.” That maybe we’re still there, after all. That we don’t really have a handle on things. We don’t have a handle on our processes. We don’t have a handle on our data. Is that where the shoe pinches, or what do you think?
Anna Carlsson: Yes, that’s back to what I talk about so much. And I’m kind of in the middle of this whole analysis. And I can mention that I’m going to produce a report that will point to all these different aspects, and also sources and various vendors that you should be able to access. I don’t know if it will be ready when this episode is released—maybe not. But I will probably include a link if you’re interested in getting hold of this report. But it’s… it comes back to our history and our specific way of working in Sweden. Back to our model that’s unique to us: that we allow— we give so much freedom, or freedom with responsibility. Sorry, not freedom, but perhaps it is the freedom to have responsibility as a manager and responsibility for your employees. That whole structure—the Swedish model—and how we negotiate and work with unions and the like. That has shaped how HR’s role has been formed, and a hundred years ago it was important to be a behavioral scientist and understand the psychology behind things to help employees. And that’s what remains. That’s how we’ve built our structures and processes. So it’s not natural for us to use and think disruptively in what we do; rather, we’re perhaps more seekers of security. And—this human-to-human contact. Then the first thing you do is not to start digitalizing.
Emira Blomberg: Did HR used to be called something else in the U.S.? Before it was called HR—has Human Resources there also gone through…?
Anna Carlsson: That’s a question I can’t answer.
Emira Blomberg: I was just thinking—here it’s “personalvetare.” I found that bit of trivia interesting. That behavioral scientists and the personnel department and those things—well, “personnel department.” Personnel. It’s not the “managers’ department,” right? And then came “Human Resources,” and sometimes—we’ve discussed this too—what should it be called? Sometimes names do matter.
Anna Carlsson: Yes, indeed, because it has an impact.
Emira Blomberg: Absolutely.
Anna Carlsson: But I actually have some numbers here about maturity. I’ve done some work on this. What I do every year is review—what are we talking about? What I talk about is how to think about maturity. I’ve had a model before, but I’ve redone it this year. I’m thinking more about digital maturity in another way—what actually is included when you’ve reached different levels. And we have—if you look at the lowest level, it’s that you’ve digitalized; you’re working on your HR systems. There are still a lot of organizations doing that, and about 19% when you look at different data sources that I’ve analyzed—and don’t ask me which ones right now, because I don’t quite have the structure yet. I have a presentation I’m going to give—so, still 19%. That’s quite a lot. That’s a fifth of Sweden’s organizations that haven’t really done the basic work. Then we have an “emerging”—or I should take the middle level first. There, you have—okay, you have your systems in place, you’re collecting data, you’re working with innovation and are on your way—you have things in place and can really—you could develop your AI strategy and really get into this and automate. About 30% of Swedish companies are there. Then we have somewhere in between—when you’re on the journey, where you’ve implemented some of the systems and are—maybe you haven’t started much with data yet. That in-between. That’s also 30%, so there’s a big chunk—30%—that should be able to dive in and become AI-driven, AI-first. But there are also many—what is it then, 31? God, I can’t count today with my jet lag. Thirty-one—well, that makes it 50% that still aren’t… not there yet.
Emira Blomberg: No, this thing about becoming AI-first—does that mean that within HR, AI should be able to answer this before a human does? Or how do you define it?
Anna Carlsson: Yes—well, it doesn’t have to just answer, because there are different types of agents. It’s about… it’s much more about having a broader perspective. One of the things they talked a lot about was how—when you’re going to become AI-first—you have to start looking at what people do in their job roles.
Emira Blomberg: Yes.
Anna Carlsson: What can be removed and handed over—what can be automated—and what remains that is important for humans? The human-oriented parts.
Emira Blomberg: Is that what AI-first is?
Anna Carlsson: It’s that you make use of AI.
Emira Blomberg: More than human resources?
Anna Carlsson: Well, more than—depends on what you’re doing.
Emira Blomberg: This is also a challenge, because I think it gets so fluffy and buzzwordy sometimes. Like “remote-first”—that I can understand.
Anna Carlsson: Yes. Because we’ve had that for a while too.
Emira Blomberg: It’s clear. It’s a geographical thing—either you sit there or you sit there. But this—if it’s not AI-first, then what is it? Then it’s human-first, I suppose. Is that how it is? Is that it?
Anna Carlsson: Yes, it has to be, because it’s either AI or a human.
Emira Blomberg: Right. Then I assume the concept of AI-first means that most of the things we do—the majority of all activities we do—should be performed by AI.
Anna Carlsson: It doesn’t have to be the majority, but the things that are suitable for AI should be performed by AI.
Emira Blomberg: I get it.
Anna Carlsson: Because it’s not an end in itself to try to automate things that don’t benefit from it.
Emira Blomberg: No.
Anna Carlsson: But, on the other hand, we do have a lot of routine work that gets done. I had lunch with an old colleague—I told you about her, someone I have great trust in—and she said that, in many HR organizations, they’ve created a model for answering questions where managers simply ask HR, and HR answers. And that—that’s human-first.
Emira Blomberg: Yes.
Anna Carlsson: But is that good value? Is that a good use of that competence’s time? No. The change is to make it so you actually have to search yourself—or there is a service. If there’s an AI service you can ask, then it becomes much easier. I’m very happy to go to an AI service when I need help with things, because it knows things. But the journey to get there is the tough part. And then the idea is that you’d rather go to AI to ask questions. And when the AI doesn’t have the answer, or it’s too complex, then you go to a human. And there’s nothing that says that this role—well, it depends on what’s included in your role. For an HR Director, for example, you shouldn’t—shouldn’t cede—an HR Director should have a role that is strategic: working with their leadership team, seeing where the challenges are, and helping them reach the organization and reach the goals you’re aiming for. That role benefits from having access to AI, so they themselves don’t need to know a lot of details, but can work on the strategic level. Right?
Anna Carlsson: And when everyone was standing there talking about their fancy tools at an event where I was a speaker, I talked about the importance of data quality. But the thing was—this was also interesting. I mean, the USA—how many years ahead can they be, 5–10 years? They don’t talk about data because it’s so natural that they’ve solved it. It’s a given that you have your data in place, that you have your systems in place. You have master data, you work with data, and data affects the business and is part of the strategic work. It’s so obvious that… it just wasn’t talked about.
Emira Blomberg: No.
Anna Carlsson: There were sessions about new tools that themselves have AI—AI chat, generative AI interfaces—so that you can easily talk to your data, making it much easier to visualize and create stories behind your data. But it wasn’t like last year, when that was the new thing. Back then, people were still talking about data and the real importance of working with your data and the different experiences they had drawn. But this time it was very, very little.
Emira Blomberg: But that’s also—speaking of Sundlo, whom you mentioned earlier. It was him, at HR Tech Amsterdam in his talk, who said “Map your processes.” That maybe we’re still there, after all. That we don’t really have a handle on it. We don’t have a handle on our processes. We don’t have a handle on our data. Is that where the shoe pinches, or what do you think?
Anna Carlsson: Yes, that goes back to what I talk about so much. And I’m kind of in the middle of this whole analysis. And I can mention that I’m going to produce a report that will point to all these different things, and also sources and various vendors that you should be able to access. I don’t know if it will be ready when this episode is released—maybe not. But I’ll probably include a link if you’re interested in getting hold of this report. But it’s—it’s back to our history and our particular way of working in Sweden. Back to our model that’s unique to us: that we allow—we give so much freedom, or freedom with responsibility. Sorry, not freedom, but maybe it is the freedom to have responsibility as a manager and responsibility for your employees. That whole structure, the Swedish model, and how we negotiate and work with unions and the like. That has shaped how HR’s role has been formed, and a hundred years ago it was important to be a behavioral scientist and know the psychology behind things to help employees. And that’s what remains. And that’s how we’ve built our structures and our processes. So it’s not natural for us to use and think disruptively in what we do; rather, we’re perhaps more security-seeking. And—this human-to-human contact. Then the first thing you do is not to start digitalizing.
Emira Blomberg: Did HR used to be called something else in the U.S.? Before it was called HR, has Human Resources there also gone through…?
Anna Carlsson: That’s a question I can’t answer.
Emira Blomberg: I was just thinking—here it’s “personalvetare.” I thought that bit of trivia was interesting. That behavioral scientists and the personnel department and those parts—well, “personnel department.” Personnel. It’s not the managers’ department, right? And then came Human Resources, and sometimes—we’ve discussed this too—what should it be called? Sometimes names matter.
Anna Carlsson: Yes, yes, because it has an impact.
Emira Blomberg: Absolutely.
Anna Carlsson: But I actually have some numbers here about maturity. I’ve worked a bit on it. What I do every year is review—what are we talking about? What I talk about is how you can think about maturity. I’ve had a model before, but I’ve redone it this year. I’m thinking more about digital maturity in another way—what actually is included when you’ve reached different levels. And we have—if you look at the lowest level, it’s about having digitalized—you’re working with your HR systems. There are still very many doing that, and about 19% when you look at different data sources that I’ve analyzed—and don’t ask me which ones right now, because I don’t quite have the structure yet. I have a presentation here that I’m going to give—so, still 19%. That’s quite a lot. That’s a fifth of Sweden’s organizations that haven’t really done the basic work. Then we have an “emerging”… or I should take the middle level first. There, it’s that—okay, you have your systems in place, you collect data, you work with innovation and are on your way—you have things in place and can really—you could make your AI strategy and really get into this and automate. About 30% of Swedish companies are there. Then we have somewhere in between, when you’re on the journey—where you’ve implemented some of the systems and are… maybe you haven’t started much with data yet. That in-between. That’s also 30%, so there’s a big chunk that’s 30% that should be able to jump in and become AI-driven, AI-first. But there are also many—what is it then, 31? God, I can’t count today with my jet lag. Thirty-one—well, that makes it 50% that are still not… not there yet.
Emira Blomberg: No, this thing about becoming AI-first—so within HR, is it that AI should be able to answer this before a human does? Or how do you define it?
Anna Carlsson: Yes—you don’t just have to answer, because there are different types of agents. It’s about… having a slightly bigger perspective. One of the things they talked a lot about was how—when you’re going to become AI-first—you have to start looking at what people do in their job roles.
Emira Blomberg: Yes.
Anna Carlsson: What can be removed and handed over, and what can be automated, and what remains that is important for the human? The human-oriented part.
Emira Blomberg: Is that what AI-first is?
Anna Carlsson: It’s then that you make use of AI.
Emira Blomberg: More than human resources?
Anna Carlsson: Well, more than—depends on what you’re doing.
Emira Blomberg: This is also a challenge, because I think it gets so fluffy and buzzy sometimes. Like “remote-first”—that I can understand.
Anna Carlsson: Yes. Because we’ve had that for a while too.
Emira Blomberg: It’s clear. It’s a geographical location—either you sit there or you sit there. But this is more—if it’s not AI-first, then what is it? Then it’s human-first, I think. Is that how it is? Is that it?
Anna Carlsson: Yes, it must be, because it’s either AI or a human.
Emira Blomberg: Right. Then I assume the concept of AI-first is that most of the things we do—the majority of all activities we do—should be performed by AI.
Anna Carlsson: It doesn’t have to be the majority, but what is suitable for AI should be performed by AI.
Emira Blomberg: I get it.
Anna Carlsson: Because it’s not an end in itself to try to automate things that don’t benefit from it.
Emira Blomberg: No.
Anna Carlsson: But, on the other hand, we do have a lot of routine work that’s performed. I had lunch with an old colleague—I told you about her, someone I have great trust in—and she said that in many HR organizations, they’ve created a model for answering questions where managers simply ask HR, and HR answers. And that—that’s human-first.
Emira Blomberg: Yes.
Anna Carlsson: But is that good value? Is that a good use of that competence’s time? No. The thing is to change it so that you actually have to search yourself—or there is a service. If there’s an AI service you can ask, it becomes much easier. I’m very happy to go to an AI service when I need help with things, because it knows things. But the journey there is the tough part. And then the idea is that you’d rather go to AI to ask questions. And when the AI doesn’t have the answer, or it’s too complex, then you go to the human. Well, there’s nothing that says that this role—it depends on what’s included in your role. Because an HR Director, for example, shouldn’t… shouldn’t give up—an HR Director should have a role to be strategic and work with their leadership team and see where the challenges are and help them reach the organization and reach the goals you’re aiming for. That role benefits from having access to AI so they themselves don’t need to know lots of details, but can work on the strategic level. Right?
Emira Blomberg: Yes, absolutely. Or—yes, it becomes big. And I think this report will be valuable because we need to become concrete.
Anna Carlsson: Yes.
Emira Blomberg: We really need to become concrete and define what we mean by different things. Because that’s probably the hardest part in all change projects. And also—this is the mother of all fuck-ups, sort of. “Well, we thought this.” “Well, but that’s not how it was.” “You thought that; I thought this.” Yes, exactly. And then you haven’t defined and teased out what you actually mean by things. So it’s really about becoming as concrete as you possibly can.
Anna Carlsson: I like models, because for me it becomes the same—then I understand how it can be put into boxes. And one of the things that’s also being talked about is different types of AI. What should we call… AI agents, I can say too, because that wasn’t entirely… We talk more like this: “We have AI.” What does that mean then? Then trying to put it into a structure. I myself got a model where it’s not just agents; we have general—what shall we say? Now I have to find the right word here. I haven’t had time to talk about this before.
Emira Blomberg: I understand. But they’ve also talked about generative AI, conversational AI, predictive AI.
Anna Carlsson: Now I’d like to see a bit more about what role they have. The one that’s used the most is the conversational—what is it called in Swedish? That’s what I was trying to find.
Emira Blomberg: No, I actually don’t know.
Anna Carlsson: “Samtals-AI,” we say “conversational AI,” right? I’ve never heard anything else. No, actually. Conversational AI is when you talk to your AI. That’s what’s used the most. And now I can’t find the numbers here, because I had those too. But they’ll be in the report later. Yes, because in a recent survey they looked in the U.S. at what people actually use their AI for. And there it’s that people are simply working with text in various ways. We’re not using AI for anything more advanced—especially at such a high level as 81% had “shadow AI,” meaning they bring their own AI. That’s a bit risky, because then you don’t have control over what you’re doing. Yes, but then I’ve seen a few other types—and those are specific agents that perform a specific task. Like with recruitment, working with a specific part. There was one that won a startup competition, and they are an agent that helps the applicant easily fill in applications—because it’s not as if our recruitment systems are always so simple—so you just talk to an AI that helps you fill in based on your information. I don’t mean parsing a CV, but rather talking to get a much simpler interface. There are a few different variants of that from before, but this one was super smart, as a service regardless of recruitment platform. It’s a specific agent that does something specific. It could be interviewers or coaches or whatever it might be. But then there are platform agents, and that’s an HR system, for example, that has an agent helping you in that system. And it can also be a Copilot that is a platform agent. So there are different variants in terms of how much you do. Then there was another variant that I call a knowledge agent, and that was Willis Towers Watson—are they called that? They’re called WTW now. So I was standing there talking to them. And I said, “It’s you.” Yes—they have developed an agent, that is, a service that allows their data and information and experience to be used by small companies. Previously, it was very expensive to buy their data on salaries and the like. Now they can sell it to small companies through an agent that brings it out and does exactly the same. Josh Bersin also has his Galileo, which he talks about as an HR knowledge agent—based on all their knowledge. So I’ve captured a few different variants.
Emira Blomberg: So exciting—when you said this about how you talk to the AI and tell it how to do things. I also find this fascinating, because so far, for us ordinary mortals, AI is still very much written and very little verbal.
Anna Carlsson: It’s a habit.
Emira Blomberg: Yes, but is it? I think this is interesting—will we also replace speech, or, you know—do you understand what I mean—finding…
Anna Carlsson: Talking to the system.
Emira Blomberg: Yes, exactly. For a while there has been resistance to this—that has to be said. Maybe it’s because personally—when they introduced those voice-controlled automated customer service calls, you just got fed up with those that didn’t understand anything—like, “Is it business or private?” “I didn’t hear what you said.” You don’t want that. You just want to talk to a human, or you want to write something. “I don’t want to go through this.” But how—that transformation will be so exciting to follow. How we’ll accept, or not accept, for example in recruitment contexts, talking to an AI screener. Is that something we want to do?
Anna Carlsson: According to studies, we do. We feel safe in that environment. No one judges us when we talk to an AI, so there is— I haven’t looked at it now, but I have old—or older—research. Not old. Let’s not say old. But I mean, in this—this thing about talking. I talk to my devices, and I— but it’s so embarrassing. I’m like, “I want to talk because I don’t have the energy to type.” And then I talk to ChatGPT. Or I write by dictating to my computer. But then I have to do it at home. I can’t sit on the subway and talk. My partner usually shouts, “Are you saying something?” “No, I’m talking to the computer.” So I talk a lot. But it has developed, and it’s us who have to help these things develop.
Emira Blomberg: Exactly. But then I also think again—let’s not nerd out too much—but I do think a lot about this. And I think there are different screening tools now that replace CV review and such, but it’s still the candidate filling things out in writing. Maybe that’s because they can do it on demand—when they’re sitting on the bus. You don’t chat away with an AI on the bus.
Anna Carlsson: On the bus.
Emira Blomberg: And things like that. Then we’ve also done some surveys: if you could be judged or assessed by a human or by an AI, which would you choose? Even though you know the human is more biased, candidates still say “human,” based on wanting to feel that you have some kind of influence—maybe that you can still manipulate this person. I don’t know—unlike the AI.
Anna Carlsson: But then it’s good—you can manipulate the person.
Emira Blomberg: No, but that’s what I mean—recruitment has been very, very much about how good you are at selling yourself. It’s been more about that than about how suitable you are for the job.
Anna Carlsson: Exactly.
Emira Blomberg: If we’re being blunt. And that behavioral change—I don’t know. Some will win from it and others will lose. And those who are the current winners in this won’t want to give up that power.
Anna Carlsson: That’s mainly what has come up—that those who are good at speaking for themselves and writing well are the ones who don’t want to lose power.
Emira Blomberg: Exactly.
Anna Carlsson: And suddenly, others get the power—those who may be more suitable. Yes, there was also a lot about when to use AI for review and when not to. And again, it’s about the maturity of both candidates and the recruiting department. But also how we—what kind of role it is. Where is it relevant? Where are the large volumes and where are the smaller volumes? Where should we place our emphasis? I had, on the podcast a long time ago, a fintech company—and I thought they had digitalized everything, but no. It was—no, but they brought everything in digitally, but they always did a call because they were so keen that those who had made it through the first steps would get a call. So yes—but there are other things. You can—you—it’s so easy to get stuck on that specific part which is an assessment thing and which we’re not used to, and some will hopefully benefit from it who maybe haven’t benefited before. But there’s so much else—I’m thinking a lot about the planning of everything that needs to be done, and all the material that needs to be evaluated, and those parts—getting through the process, quickly producing contracts and all that.
Emira Blomberg: I think there are a lot of opponents out there who will fight to hold on. I also think about managers in organizations who might never have become managers if we had made more data-driven decisions, and Learning & Development initiatives that might be directed in the right way. I’m not saying all Learning & Development initiatives have been misdirected, but we do have quite a few sad examples of large costs for learning and development without much effect. Then maybe we should also mention Sana Labs—should we talk about whether they were there?
Anna Carlsson: No, they were at the Workday event happening at the same time. Not so strange, since the acquisition was announced when I was sitting having breakfast with two women who were HR Tech leads at large companies. One of them just yelled, “Workday has bought Sana.” So that was super cool. And yes—this again ties back to digitalization and the importance of looking at new solutions like a company such as Workday does, and SuccessFactors. Yes, that’s a whole discussion too, but we can take it here—that I was also at some sessions about investments and what people invest in.
Emira Blomberg: Yes.
Anna Carlsson: And where the money goes. That’s it—they see very clearly that when larger companies acquire other companies, it’s very much about both getting competence and solutions that are built without the old history—like Sana. Why does Workday buy Sana? There are several different reasons. And the one they bought earlier—what are they called? Those recruitment companies—Paradox, I think. I forgot the name. It’s about that. And it’s the same with SuccessFactors buying SmartRecruiters as well. It’s about both getting the competence and something that’s conceived for—because it’s harder—and it affects the market a lot right now, with the desire to get hold of AI competence and AI solutions. As investors and as larger companies doing acquisitions. So it’s interesting.
Emira Blomberg: Yes, indeed.
Anna Carlsson: Then I have a few other things. Then we should wrap up, I think. I also thought a lot about what we—what wasn’t talked about. Regarding what I said about there not really being a focus on data. There were a few sessions about the value of analytics, and I met a very interesting company that links learning—that’s why I thought of it now—learning to results, and had a tool focused solely on proving the value of learning.
Emira Blomberg: Fun!
Anna Carlsson: But what people didn’t talk about—what was huge in previous years—was skills. But at the same time, we do talk about it in connection with AI, that we have to start looking at skills and tasks.
Emira Blomberg: Yep.
Anna Carlsson: Competence and tasks. Trying to look at that. And that’s where—back to this AI-first idea. If you’re going to be able to—what tasks are actually not necessary for a human to do? And yes, that will affect us enormously. But people aren’t talking about skills tools as much as we used to.
Emira Blomberg: And I can understand that, because I’m a bit skeptical. Or—like this. I’ve said before that I worked at a large— a big insurance company that had come very far with skills mapping and had created their profiles and so on. And they’re super ambitious about it, and it worked very well. But wow, what a production it is! And someone has to run that machine and be passionate about it and have the stamina. And then I also think that what they did, which was successful, was that they stopped at a level where you don’t go down into micro-skills, because that’s what I think I’ve seen with these skills companies that profile themselves as skills mapping and such things. It becomes too detailed. It’s at such a level of detail—how quickly can you learn this skill? You know—do we know it? There are so many layers before you get value from these tools. So I think they’re far too… The idea is good, but I don’t think that’s… that it’s the solution.
Anna Carlsson: No, again—it’s about what task you’re trying to solve. And you can’t—it’s constantly about looking at what you want to achieve and what tools you need to achieve it. And for some, that may be a skills tool. But it depends. It’s not something everyone should have. But we do need to be able to look at our job roles and structure them somehow. But there are also solutions that run in the background. Instead of being a tool you have to use, you want to be able to pull out data and model it to find out what we have and which jobs—what tasks we actually perform.
Emira Blomberg: Yes, and then I also really believe more in this apprenticeship thing. I really do— the apprenticeship setup. And I also think that—I’ve worked with psychometric tests and the Big Five, or five-factor test, and I can say that it’s a spring for Openness, which is one of those five factors—meaning you’re interested in new thinking, that you get energized by your inner creative process and that you need to nourish it all the time. Abstract thinking, questioning, and above all that you feel a desire for what’s new. You feel a desire. You want more. You want… These are the types who have those abstract discussions over a glass of wine at a dinner. And then it becomes two camps. One corner sits and talks decking and wood types, and the other corner sits and talks—I don’t know—something very meta, at least.
Anna Carlsson: I think it’s too complex and too fast-moving. It’s going too fast. We can’t keep up if we’re going to map skills and constantly make sure we’re doing skills planning. Instead, I believe more in—finding those people who can quickly shift their mindset and who are the kind of multitaskers you can use for many things. In talent management, people talk about jackdaws and hummingbirds: that hummingbirds are those who need very, very specific conditions in their environment to succeed, while jackdaws can be dropped anywhere and they’ll make their way. And I think a new kind of talent is being born that’s a bit like that.
Anna Carlsson: But at the same time, people shouldn’t feel scared listening—like there’s only one kind of profile.
Emira Blomberg: Oh no, no, no—yes and no. That really wasn’t the intention. No, but I think—again—I think it’s about people, and now someone out there will sit and say, “We can certainly break a person down into components and deconstruct,” but I still think there’s so much—there’s so much here. Behavioral changes. Who holds the power? What incentives do people have to resist things versus actually promoting things? There are so many layers, and the solution isn’t to buy a smart tool—the solution is what you’re talking about as well: change management, the “why.”
Anna Carlsson: And then I have to add—but then we really have to stop talking—that the money being spent on change work has increased. Very happy to see that. This is a report that comes from an organization called Sapient Insights. It hasn’t been released yet, but they always release preliminary information at the event. I’ll get access to it, but not until the end of October. Therefore, my report won’t be out until I’ve dug a bit more into that report, actually. But it’s great that it’s happening—that people are investing. It’s a must, right?
Emira Blomberg: Yes, absolutely. Isn’t that what it becomes then? Because that’s what remains, simply.
Anna Carlsson: We’re now simply—if you look at what will happen going forward in the long term—it’s that we’ll have completely different structures and parts of the organization, how we work, and what’s done by AI and what’s done by us. Oh dear, it’s going to be stressful.
Emira Blomberg: But again, change moves much slower than you think. That classic thing that we overestimate it in the short term but underestimate it in the long term. But in ten years—it’ll be interesting to see what’s happened.
Anna Carlsson: 2035. Oh my. Thank you—for this conversation—you’ve probably gotten wiser from what I’ve learned.
Emira Blomberg: Yes, I think it’s interesting that there are such concrete cases. And then I think—we’ll put in a plug for that report later, where it will be a bit more concrete.
Anna Carlsson: I’ll do everything I can to make a concrete report with useful material that you can use in your planning.
Emira Blomberg: Great—fun!
Anna Carlsson: But is this the last one?
Emira Blomberg: We’ll see.
Anna Carlsson: I’m keeping the door open.
Emira Blomberg: Exactly. I’m going to RecFest in Nashville anyway. Then maybe we’ll do something. We’ll see. We’ll see. Are those the closing words?
Anna Carlsson: We’ll see.
Emira Blomberg: That’s better than saying thanks and goodbye. Right?
Anna Carlsson: Yes, it is. It’s better to say “until next time,” in some way.
Emira Blomberg: Yes, absolutely.
