In this episode of the HR Digitalization Podcast, we meet Josefin Erséus, CHRO at Teleperformance Nordic. Together, we talk about how AI is used practically and strategically within HR at one of the world’s largest BPO companies. With over ten years of experience from the customer service industry and a deep understanding of how AI can be integrated into HR processes, she shares valuable insights from a global company with over half a million employees.
It is a conversation for anyone who wants concrete advice on how AI can work in an organization and what is required to implement new solutions in a responsible and value-creating way. The episode offers both strategic reflection and concrete examples, and is just as suitable for HR specialists as for managers responsible for digitalization, leadership, and organizational development.
An episode for those who want to understand how AI actually works in practice, and what is required to use it in a sustainable, human, and business-driven way.
The episode is divided into two parts and can be listened to as a whole or as separate episodes.
In the first part, the listener gets concrete examples of how AI is used in HR work at Teleperformance.
The second part focuses on the bigger questions: ethics, data security, legislation, and AI competence.
An episode for those who want to understand, influence, and help shape the future of work—listen to the episode and be inspired!
Note: This episode is in Swedish. A translated transcript is available below.
Transcription:
Anna Carlsson: In this episode of the HR Digitalization Podcast, Emira and I interview Josefin Erséus, CHRO at Teleperformance Nordic – one of the few people I’ve met who has actually worked practically with AI for several years.
The first part of the episode is about concrete experiences: how AI is used in recruitment, onboarding, coaching, and quality follow-ups. Josefin shares insights about both the advantages and pitfalls – and how it affects candidates, managers, and the role of HR.
The second part deals more with the bigger issues. With Josefin’s experience also come wise opinions and insights.
We talk about ethics, security, regulations, AI competence, and why it’s crucial for HR to build its own expertise in order to build trust in the organization and lead with purpose through the digital transformation.
You can listen to the full original version – or choose to hear the first or second half as standalone episodes.
And if you want to strengthen your own knowledge in the field
I lead the HR Association's training on AI and HR Tech, where we go through everything you need to understand in order to make informed decisions about what’s best for your organization.
Do you have an HR system, an exciting HR Tech solution, or an AI service that you’d like more people to discover?
Then you have the opportunity to participate in a sponsored episode of the HR Digitalization Podcast – preferably together with one of your clients. In a joint conversation, you can share how the solution is used, what needs it meets, and the value it has created in practice.
This is a format that builds credibility, strengthens your brand, and gives listeners the opportunity to understand what makes your offer great – at their own pace.
You can also choose to be featured in a shorter advertisement, just like this one.
Each season includes a limited number of sponsored episodes – want to be part of it?
Reach out to me at anna@hrdigi.se, and I’ll tell you more.
Anna Carlsson: Welcome to the HR Digitalization Podcast.
Josefin Erséus: Thank you so much! Great to be here!
Anna Carlsson: And today, Emira is also here.
Emira Blomberg: I’m here too.
Anna Carlsson: Exactly.
Emira Blomberg: Yes.
Anna Carlsson: And today, we’re going to talk about AI. Yesterday, I had a meeting with some colleagues in the industry, and we started talking about whether people are tired of AI. We’ll see if people want to listen to this.
Josefin Erséus: I hope not.
Anna Carlsson: No, exactly. We’ll take some other perspectives, I think. We don’t need to say more than that, but let’s start with you introducing yourself, Josefin.
Josefin Erséus: Yes, my name is Josefin Erséus. I work at Teleperformance Nordic as the HR lead there, and Teleperformance – we are a very, very large global company. We have about five hundred thousand employees around the world, so we are definitely not a small employer globally.
Anna Carlsson: Do you have any kind of ranking, like you’re among the largest in the world?
Josefin Erséus: Yes, we are the largest BPO company, as it’s called, in the world. We work with inbound customer service, primarily helping to create better customer experiences for our client companies – often very large global players that we work with. And that of course places very high demands on us to provide good service, good customer support, and ensure that we stay at the forefront when it comes to being good at delivering solutions to them.
Anna Carlsson: So you are like… you are customer service – but for many different types of companies.
Josefin Erséus: Exactly.
Anna Carlsson: Do you have any examples to give?
Josefin Erséus: I’m not allowed to say which companies, but it’s large telecom companies, large mobile operators. It can also be smaller players, hotel chains, and similar. So it’s quite a large mix of different clients we have.
Emira Blomberg: Are you a reactive or proactive customer service – do people call you? Or do you also do outbound?
Josefin Erséus: No, no outbound calls at all – only inbound calls. Outbound is strongly associated with telemarketing. What we do with sales is that we sell to existing customers. So when someone calls in and maybe wants to cancel a subscription, then we try to retain them a little longer. Or maybe offer them another plan that suits them better. That’s the kind of selling we do.
Anna Carlsson: And that probably means you also have quite a young category of employees. What does that look like?
Josefin Erséus: Yes, exactly. You really can’t avoid the topic of staff turnover when talking about the customer service industry. Our customer service is mostly made up of young employees, who may even have their first job with us. And when it’s a young population, there is also very high staff turnover. They don’t think it’s strange to change jobs after a year. Or that they want to start studying somewhere or go backpacking or something like that. So there’s a very high turnover of staff.
Anna Carlsson: It’s an interesting category.
Josefin Erséus: A really fun population to work with.
Anna Carlsson: And you’ve worked with this for a long time?
Josefin Erséus: Yes, I’ve worked in this industry for almost ten years. Actually, more than ten years. My goodness, time flies. It’s basically been my whole career – working in this industry.
Anna Carlsson: But HR lead, HR manager – what is your role?
Josefin Erséus: Oh right, I forgot. My official title is CHRO. So that’s Chief Human Resources Officer, but I usually just say HR lead for the Nordic organization.
Anna Carlsson: So you carry a lot of responsibility for the Nordic operations?
Josefin Erséus: Yes, absolutely. Of course, it’s important that we align what we do with global initiatives and global strategies and regional goals we’ve set. But definitely a lot of freedom and a lot of flexibility in what we can do.
Anna Carlsson: What we’re going to talk about today is AI, and a lot around staff turnover and recruitment and the like. And you got interested in this digital stuff quite early on?
Josefin Erséus: Yes, I’ve thought a lot about why that happened. I don’t think it was a conscious decision on my part to work with digitalization – or with what has become AI in recent years. Rather, it was a series of random events that led me here, and it’s felt very right once I got into it. It started with me getting to work, or having the opportunity to work, with large digitalization projects at previous companies, as well as major organizational development initiatives. So really, the common thread has been working with constant change and how it affects companies and employees.
Emira Blomberg: But I also think that the AI issue naturally comes up, considering the population you work with and the job category. Especially given what we’ve heard about entire departments being laid off and replaced by AI. I mean, it’s a bit like when Henrik from Sodexo was here – he works with the Future of Work and future workplaces. There’s a real risk of becoming obsolete when disruptive things happen in society – like COVID or now the emergence of AI. And if your entire business isn’t to be questioned, then I guess you really need to think this through.
Josefin Erséus: Yes, that’s exactly right. And I think that when AI became such a big thing a few years ago – it must be nearly three years now, or two and a half – I think it was a very big challenge for all companies in this industry to try to understand: what is happening now, and how is it going to affect us? What we can see now is that the hype has died down a bit, and people are realizing that, no, we can’t just replace all humans with AI and expect it to work out. We see that we still need human relationships and human contact. There are several major players who have gone public saying that they would let AI handle all their customer service – only to go out a year later and say, no, we’ve changed our minds. We need people in our customer service, because we can see it didn’t work out as we had hoped.
At the same time, I do believe AI can really improve efficiency, especially when it comes to transactional tasks and repetitive questions. But then it’s important that companies, on a strategic level, really talk about: what do we do with the time that is freed up? Does it mean we have to lay off a certain number of employees? Or can we do something else that creates even more value for the company? Something that leads to even greater revenue than we had before? Since now we’ve freed up time to do other, perhaps more value-creating, things.
Anna Carlsson: This is such an incredibly interesting question, and I’ve had the chance to discuss it too – the human versus AI debate: where should we position ourselves? And you guys had the advantage of getting started early, right? Did you start before the whole hype around AI really took off?
Josefin Erséus: Yes, I would say we did. We were actually one of the first companies to be involved when OpenAI was in its early stages. At that time, they needed a lot of data – a lot of real-world situations to relate to in order to train their models. Teleperformance was one of the companies that collaborated with OpenAI to help develop the language model that we now know as ChatGPT. So we were very, very early with AI as a company. And absolutely – in almost every process we have today, there is some form of AI tool or model that helps and supports. It’s so natural for us now that we don’t even have to say we’re using a tool – it’s just there.
Anna Carlsson: Do the employees like it? What do the employees think now? They must be quite used to it? Or is it because they are part of a generation that grew up with digital tools? Does that make it easier for them?
Josefin Erséus: I think it’s very welcomed that we have this kind of solution. For example, we have some very cool tools at the global level where a customer service agent can receive a call – which is of course recorded with the customer’s consent – and then they get real-time tips on how to adjust their tone, how to handle the situation better. They can even receive coaching suggestions during the call. It’s a very powerful support tool that doesn’t replace but rather strengthens the service provided to the customer.
Anna Carlsson: So that you can respond at a higher level. That’s always the point – to meet the person who is reaching out. I talked about this back in 2015 when I started working with AI, and now you have it as a reality. It’s very interesting.
Emira Blomberg: I mean, you’ve said that you collaborated with OpenAI, and I imagine you’ve done a lot of things. And I can imagine that many listeners haven’t done nearly as much. One thing I’ve been reflecting on – I was at HR Tech in Amsterdam, when we were there – is that many people talk about the end result. Like, “This is how it’s going to be,” and that some skills will become obsolete, and so on. But very few people talk about the actual implementation. Also, how did it even start? What was the first thing you did?
Josefin Erséus: Oh, well I haven’t been with the company that long. I was involved when we started, but if I talk about an initiative or implementation that I’ve been part of from the beginning, it would be when we implemented AI in the recruitment process. That was super interesting, because the reason we wanted to bring AI into recruitment was that, since we have such high staff turnover, it places big demands on us to find replacements for those who leave. But we also have customers with highly fluctuating volumes – like, say, a telecom company that is going to broadcast the European Championship, and suddenly everyone wants that telecom service. So then we get a huge influx of calls that we need to handle.
Of course we plan for that, but it means we need to scale up quickly – and it’s not unusual that we recruit, say, 60 to 80 people per month. And then you can imagine, it’s not just 60 to 80 applications – it could be thousands of applications we receive. So it becomes very hard to have the right level of staffing in the recruitment team to handle that inflow.
So what we did was to find a tool called Hubert AI, and what this tool does is that it screens all the candidates who apply. All the thousands, let’s say, who apply to us get a link sent to them where they start a conversation with an AI called Hubert. We’ve set up criteria in a profile card that defines what should be asked, and so on, and then we can very quickly see how well a candidate matches the role – or not. Which gives us a quick overview of who we should proceed with.
So it’s very, very efficient. And what we saw – from the moment we implemented this tool to when it was fully in place – it only took six weeks. That’s an extremely short time to implement. Usually, it can take years to implement this type of large system. But what we saw from day one was immediate efficiency gains. Honestly, it’s almost embarrassing to say it – but we saw a 65% efficiency increase in our processes.
Emira Blomberg: Yeah, we’ve talked about Hubert before, I think.
Josefin Erséus: I thought it was great.
Emira Blomberg: Yeah, it’s been around.
Anna Carlsson: Yes, the CEO was a guest – that was maybe a year and a half ago. Back then, it was hard to find people who could talk about AI at all. And he was great. I think it’s a fantastic example of Swedish innovation.
Josefin Erséus: Totally agree.
Emira Blomberg: I was on the jury for Innovation Company of the Year at the Recruitment Awards, where Hubert won that category.
Josefin Erséus: That’s so fun. Yeah, it’s a really cool company. I think they’re fantastic.
Anna Carlsson: But what I think a lot of people are wondering – if you’re doing this so fast, how did you know that the selection was right? Did you have a lot of data that you could train the AI on? Because it goes back to the question of: how do you actually get started in practice? You probably can’t start from a completely manual process that’s based more on a recruiter’s gut feeling.
Josefin Erséus: Exactly.
Anna Carlsson: What was it like before?
Josefin Erséus: Well, in our industry, data is king – or queen. Yes, king and queen. But it’s true. We have an incredible amount of data, and we also measure performance down to the minute, basically, for our customer service reps. We can see customer satisfaction, how quickly they resolve issues, and so on.
That also allows us to test what personality traits the top performers have. Who thrives the most, who feels the best, who creates the best dynamic in this industry or in their teams. And based on that, we could extract these profiles and see who performs best and thrives most in our company – and then create a profile card that matches that.
Then of course it was also important to look at whether there was any bias in the population we were targeting. Were we, for instance, tending to recruit more men than women because certain personality traits were more typically male-coded than female-coded? So it was really important to evaluate that objectively before putting the tool into full use.
Emira Blomberg: That sounds exemplary, because I think that’s often something people miss—checking for bias. That you create a profile that’s just a copy-paste of what you already have, which isn’t necessarily what you’ll need in the future either.
Josefin Erséus: Absolutely. I think it’s so interesting that you bring that up, because I think it was Amazon, about ten years ago or something like that, who tried to be first with this kind of screening of all candidates who submitted a CV. And it was a self-trained AI model, and it was trained on old data. And when they put it into use, it quickly became clear that only men were being invited to interviews and moving forward in the process.
And when they tried to go back and fix that—when they realized it was a problem—they couldn’t, because the bias was so deeply embedded in the model that they couldn’t train it out. So they had to scrap the whole project.
Anna Carlsson: Start completely from scratch again.
Josefin Erséus: Exactly.
Anna Carlsson: But when you’ve worked with this selection process—which I think is fantastic—it also means that everyone gets a chance, no matter who they are or what background they have. There’s no judgment. Do you use AI further along in the process as well?
Josefin Erséus: Yes, to some extent, I would say. For onboarding and coaching, especially at the regional level. So it’s not just for the Nordic region, but on a larger scale.
We’ve mainly worked with a tool called Centrical, which we’ve seen deliver very successful results. This system is integrated with all our telephony systems and other tools we use, which means it can tailor coaching initiatives. It can personalize how to boost your development or competence in the right direction.
You can see your KPIs in a gamified way, so that those working in customer service feel a bit challenged and think it’s fun. And at the same time, the manager can see how the team is doing and what small pushes or coaching efforts are needed for a specific individual to improve performance.
Emira Blomberg: Is it a performance management system?
Josefin Erséus: Yes, you could say that it is.
Emira Blomberg: Mhm. Exciting.
Josefin Erséus: Yes, really.
Emira Blomberg: So it’s both onboarding, learning, and coaching?
Josefin Erséus: Yes, exactly. It’s all in one.
Anna Carlsson: It’s an AI-based tool?
Josefin Erséus: Yes.
Emira Blomberg: It’s a great example of enhancement—as you said earlier—that with the help of AI, we enhance and boost our own performance, rather than being replaced by AI.
Josefin Erséus: Yes. And I think it’s really important that when using AI, it shouldn’t feel like a spotlight shining down on you, only to check that you’re doing right or to catch you doing something wrong, or just to feed data into a dashboard.
It should work like a compass that shows direction and helps guide us in the right way. And that’s where I think these tools are really helpful.
Anna Carlsson: But in that recruitment process—if we go back to that—once you’ve got a selection of candidates from Hubert, what happens next? Do you do personal interviews?
Josefin Erséus: Yes, exactly. And that’s where the human contact comes in, which is also super important. But what we noticed when we implemented it was that, firstly, the candidate experience became much better. The candidates felt it was smooth and fast, and they felt involved from the very beginning.
But also, our recruiters felt that this was so exciting to be part of, because it wasn’t really the administration and sifting through CVs and reading loads of documents that made them want to become recruiters. They wanted to meet people and determine if they were right for the role. So now they could really focus on that, which was also a huge lift for them to be part of that change.
Now I’ve lost track—what was your question?
Anna Carlsson: It was about the next step, that it becomes personal. So, in the next step, there’s a personal interview. Do the recruiters also get tools to help them, like “this is the kind of profile Hubert has identified,” so they know what to look for?
Josefin Erséus: Exactly.
Anna Carlsson: Like a type of assessment?
Josefin Erséus: Yes, exactly. That’s exactly what it does. It gives an assessment of the candidate who submitted their answers or interacted with Hubert. And after that, you can see if there’s something unclear, or if you want more examples of previous experiences or situations they’ve faced at past jobs, and so on.
It’s a very easy way to get an overview of the candidate’s profile.
Emira Blomberg: Maybe we should just point out why it’s so successful to implement AI in this kind of process. I think there are some clear criteria here for listeners who might be considering it themselves.
Like the fact that there are large volumes of applications, and that humans are really bad at reading CVs—especially when the CV doesn’t contain relevant information. In your case, it’s young people who often don’t have prior experience, and so on.
So just to be clear: CVs are relevant in some recruitments—it’s not one size fits all. But in this case, you have large volumes, CVs are often irrelevant, and people are bad at reading them. It’s heavy and administrative. So the result is just better and better and better—better quality and more efficiency.
Josefin Erséus: But I also think it feels more ethically correct. Because what we do is give everyone the same opportunity to answer the exact same questions. It doesn’t matter if you’re good at writing a CV or not, or presenting yourself in a certain way—everyone gets the same chance to do the exact same thing. And that feels really good, to be able to offer that opportunity to our candidates and those who want to work with us.
Anna Carlsson: So they don’t need to write so much—they should be able to talk. Do they speak to Hubert, or…?
Josefin Erséus: No, it’s written text. So you don’t speak—you’re not interviewed by the AI—but you write. But it’s like a dialogue where the system encourages you to give examples and lets you request clarifications in a simple way. It doesn’t feel like a formal interview—it feels more like a relaxed conversation where everyone is welcome.
Anna Carlsson: Do you use it—thinking about how different roles require different processes—do you use similar processes for other types of roles, or do you have completely different processes depending on the role?
Josefin Erséus: It depends. When it comes to roles that aren’t mass recruitments, we still use the traditional recruitment process. But for the large volumes of incoming candidates, we use Hubert.
Anna Carlsson: And what about follow-ups with those who apply? That’s also something that’s often discussed—that candidates want feedback on where they are in the process. Do you have that in all your processes, or is it manual?
Josefin Erséus: Yes, that’s definitely a key point. We work a lot to ensure that everyone who applies receives some kind of response. In the process where we use Hubert, it’s actually built into the tool that the candidate receives feedback, which is incredibly valuable.
It gives a summary after the interview, and it might say something like:
“Thank you for participating – we will proceed with your application,” or
“Thank you for your participation, but we have chosen to move forward with other candidates this time.”
So it happens automatically, and it’s both fast and consistent, which improves the candidate experience.
In the more traditional processes, where we don’t use AI, the follow-up is still manual, which of course takes more time. But we are working on improving that too – we know it’s one of the most important parts of leaving a good impression as an employer, even if the candidate doesn’t move forward.
So it’s something we’re focusing on and constantly trying to improve.
Anna Carlsson: Because I think that’s something we talk about a lot in the HR world—how important it is to provide feedback. But also how hard it can be to do it when the volumes are large.
Josefin Erséus: Yes, exactly. That’s why I think it’s so good that it’s integrated into the tool. Because then it happens automatically, and you can ensure that everyone receives something.
And even if it’s not always a detailed personal message, it’s still something. It’s a sign that we’ve acknowledged their effort, and I think that really matters to candidates.
Emira Blomberg: Especially today, when people talk about “ghosting” in recruitment processes—when no one gets back to you at all. That’s still surprisingly common, unfortunately.
Josefin Erséus: Yes, and I think it affects the employer brand more than many realize. Even if someone doesn’t get the job, they still walk away with an impression of the company. And if they don’t get any feedback, or if the process feels slow or confusing, that impression can be negative. And that’s unnecessary, because often it’s not about intent—it’s about a lack of time or tools.
Anna Carlsson: Or structure.
Josefin Erséus: Exactly, and with AI you can build in that structure in a smart and scalable way.
Anna Carlsson: So smart. And what you describe also makes it clear that AI doesn’t have to be some big, scary change. It can actually be a way to improve things that we already want to do better but haven’t had the resources for.
Josefin Erséus: Exactly. That’s how we’ve tried to think about it. Not replacing people, but supporting and improving the work we already do. For example, in recruitment—we don’t want to remove the human part. On the contrary, we want to free up time so that our recruiters can focus on the parts that really matter: meeting people, evaluating soft skills, making informed decisions. And if we can use AI to sort applications more efficiently or identify patterns that help us make better decisions, then that’s fantastic.
Emira Blomberg: Yes, it’s about prioritizing where human time and energy are best spent.
Josefin Erséus: Yes. And about making sure that the people who work with these things actually get to do what they’re passionate about—and what creates the most value.
Anna Carlsson: Do you notice that in employee satisfaction too? That it affects how people feel about their jobs?
Josefin Erséus: Absolutely. Especially among the recruiters. Many have said that it’s more fun now. They don’t feel like they’re drowning in admin work—they feel like they’re doing real recruitment again. And that’s valuable, both for their well-being and for the results we get.
Emira Blomberg: It sounds like a win-win. Better candidate experience, more efficient processes, and more satisfied employees internally.
Josefin Erséus: Yes, and also better results for the business. Because when we recruit the right people more quickly and give them a better start with onboarding and coaching, they perform better. And that, in turn, leads to more satisfied customers.
So it all hangs together. That’s what’s so powerful.
Anna Carlsson: And that’s the business case. That’s what makes it easier to get buy-in from management as well.
Josefin Erséus: Exactly. When we can show concrete numbers—on efficiency, on customer satisfaction, on employee satisfaction—then it’s not a “nice to have” anymore. It’s a strategic investment.
Emira Blomberg: That’s something we should say more often, I think. That AI doesn’t just belong to IT or HR. It’s a business issue.
Josefin Erséus: Yes! That’s so important. It’s easy to fall into the trap of thinking “Oh, AI—that’s something for the tech department,” or “That’s HR’s thing.” But this affects the whole company. It affects how we work, how we deliver value, how we grow.
Emira Blomberg: That’s such an important mindset shift. And I think it’s something that HR can really drive—helping the organization move from fear to learning.
Josefin Erséus: Yes, absolutely. HR has a huge role to play here. Both in building AI competence within the HR function, but also in supporting managers and employees throughout the organization.
It’s about equipping people with the tools they need to understand, to influence, and to feel safe in the digital transformation.
Anna Carlsson: And in that, it’s also important that HR itself gets the chance to develop. That we get to build our own skills—both in data, in technology, and in strategy.
Josefin Erséus: Yes. And that’s something I’ve tried to prioritize in my own role too. Making sure we’re not just “support” but part of the strategic conversation. Because the people perspective is so critical—especially now.
Anna Carlsson: And I think that’s a great note to end on. That we in HR need to take our place—not just react, but lead. Especially when it comes to AI and digital transformation.
Josefin Erséus: Yes, I really believe that. We have a unique opportunity to contribute with something that no other function can—an understanding of people, behavior, and culture.
And if we combine that with tech-savviness and strategic thinking, then we can really be a driving force in shaping the future of work.
Emira Blomberg: Thank you so much for joining us, Josefin. It’s been incredibly inspiring to hear how you work with these issues in practice.
Josefin Erséus: Thank you! It was so fun to be here and to talk about something I’m so passionate about.
Anna Carlsson: And thanks to everyone who listened—see you next time!
