Agentic AI is only as useful as the data it can access, and getting that foundation right is proving to be the harder half of the work. Years of mergers, acquisitions, and local decision-making have left many talent operations running on data and processes that were never meant to work together, and no amount of AI on top will fix what lies beneath. Some organizations are now rethinking their technology strategy in light of that problem.
So what does getting AI-ready actually involve, and what does it change about the decisions you make?
My guest this week is Lia Manafova, Talent Technology Strategy Lead at Sanofi, a global pharmaceutical company hiring at scale across more than 70 countries. In our conversation, Lia explains why the data foundation must come first, what an anchor product strategy looks like in practice, and what she has learned about making technology stick.
In the interview, we discuss:
- Why AI readiness starts with data, not AI
- Building a bridge between the business and the digital team
- The challenge of constant transformation and change fatigue
- What is an anchor product strategy?
- How the Workday, Paradox and HiredScore acquisitions changed the options
- Best-of-breed or a single source of truth?
- Keeping recruiters in one system rather than three
- Piloting with the people who will use it every day
- The case for keeping the semi-automated option
- Building an ROI story the business understands
- What does the future look like?
Follow this podcast on Apple Podcasts.
Follow this podcast on Spotify.
Matt Alder
Using agentic AI in hiring depends on having data that the AI can actually reach. And in a lot of large organisations, that data is spread across dozens of systems that were never built to work together. Getting ready for AI means fixing that first. What does that take and how does it change the tech strategy? Keep listening to find out.
Advert
Support for this podcast comes from Paradox. You know, I’ve recently heard some crazy success stories when it comes to hiring with AI. FedEx is sending offers to candidates within 10 minutes. General Motors saved $2 million in recruiting costs in a year. Chipotle reduced time to hire from 12 days to 4. And 7-Eleven is saving their store leaders 40,000 hours per week. The craziest thing is all of these companies did it by leveraging the same technology. Paradox. Acquired by Workday in 2025, Paradox pioneered conversational hiring. Whether acting as your ATS or automating the hiring process on top of your existing HCM, Paradox streamlines the entire experience. And if you’re a Workday customer, Paradox now operates seamlessly within their talent acquisition suite as the Workday Paradox candidate experience agent. Driven by a 24-7 AI assistant, Paradox can handle up to 95% of the hiring process for deskless hiring teams, or just automate specific, time-consuming tasks like screening, interview scheduling, and onboarding to allow recruiters to focus on recruiting. Paradox continues to help hundreds of the world’s top employers simplify hiring and save money, while creating great candidate experiences. Spend more time with people, not software. You can find out more by going to paradox.ai. That’s paradox.ai.
Matt Alder
809, of Recruiting Future, with me, Matt Alder.
Agentic AI is only as effective as the data it can reach, and getting that data foundation right isn’t easy. A lot of talent operations run on data and processes that were never meant to act as one, and no amount of AI on top will fix what sits underneath. Some organisations are now rethinking their tech strategy around that very problem. So, what does getting AI ready actually involve? And what does it change about the decisions that you make? My guest this week is Lia Manafova, who leads Talent Technology Strategy at Sanofi, a global pharmaceutical company hiring at scale across more than 70 countries. In our conversation, Lia explains why the data foundation has to come first, what an anchor product strategy looks like in practice, and what she’s learned about making the technology stick.
Hi, Lia, and welcome to the podcast. It’s an absolute pleasure to have you on the show. Please, could you introduce yourself and tell everyone what you do?
Lia Manafova
Absolutely. So my name is Lia Manafova. I’m based in Barcelona, and I’m with Sanofi, which is a French pharmaceutical company, leading the talent technology strategy. So our organisation is sort of a bridge between the business and the digital world when it comes to, consultative support with making decisions on technology investments, how to simplify our infrastructure and how to be more future prepared.
Matt Alder
Yeah, that’s a really interesting role. There aren’t many companies that have that role, aren’t there? How long have you been doing it and how did you get into it?
Lia Manafova
The organisation is quite new at Sanofi. I think we’ve been established for about a year and a half. And the way the decision came about is because the technology infrastructure changes so much. There’s so much on the market that’s available. You don’t know what’s a hype. You don’t know what’s actually there to, you know, stick around and how to make these decisions. And sometimes in the digital world, the business context is missing. And in the business world, there’s not that much interest on how things are made. There’s just business outputs, which which is what how it should be, and so this kind of came around for us when we started thinking about how to make our AI investments, how to actually prepare for the future, how to move two steps ahead while at the same time staying grounded and what we need to fix and standardise and keep maintaining. So it was sort of a natural decision for us.
Matt Alder
That makes a huge amount particularly with the way that technology is at the moment. Give us a little bit more context on the company in terms of how many people you hire and the biggest talent challenges that you currently have.
Lia Manafova
Absolutely. So we’re about 80,000 people globally in over 70 countries. We hire around 20. So we have around 20K requisitions per year, so 20,000. So you can imagine how many candidates that translates to and how many screenings and interviews that we have to do per year. We also, as an organisation, we are historically a company of mergers and acquisitions. And so there’s a lot of, there used to be a lot of centralisation in sort of. In the local markets or in geographies or in business units in terms of how things were done.
And that includes talent acquisition, talent management, talent development. And about two, three years ago, we started on our transformation journey in a talent space where we wanted to make sure that we globalise and centralise our operations. We used to be very RPO reliant as well. So we wanted to bring our recruitment in-house and take a better advantage, let’s say, of our hubs. Because we did have natural organic hubs, but they weren’t used as much as, they should have been. And so that led us on a journey of talent transformation. We call it a talent operating model and, which we’re sort of coming out of now, going into another transformation. Transformations become like an average Tuesday these days.
And, you know, now focusing on similar centralisation, and globalisation, but now focused on technology and making sure that our technology reflects the simplicity of the processes that we were aiming for.
Matt Alder
What’s it like to be constantly going through transformation? How do you make sure the transformations are effective and that change really happens?
Lia Manafova
Yeah, I spoke with people from other organisations, and it seems like that’s the case for everyone these days. So Sanofi is not unique in that sense that transformations are becoming normal.
Occurrence. But I would suspect that’s probably due to the fact that things are moving so quickly these days and changing so quickly. I mean, you speak with someone five years ago, they would probably tell you a little bit about machine learning and skills, but nobody was talking about AI. Now you’re talking about AI and, you know, the work, the nature of work is changing as well. So this is not, Sanofi is not any different from our other organisation in that sense. I think what’s the biggest challenge and the biggest focus for us right now is the change itself. So making sure that it gets adopted, that people are excited about it, that you bring them on a journey with you. And in my time at Sanofi, what I found out is that you have to excite people so they don’t get sort of overwhelmed by this change fatigue that comes with the transformations. So it’s definitely difficult. We figured it out in some cases, especially when it comes to our talent operating model transformation when it comes to our other program that we did recently on a skills implementation. But sometimes it’s not really as easy to do when you go in as well with your, little project and technology implementations and all these things that add up on top of the day-to-day work that people are doing.
Matt Alder
I want to talk a bit about AI, but I want to come from an angle that I don’t see people talking about very often. But it’s one that’s really important, and I know it’s important to you as well. And it’s the concept of centralising data in a way that makes it effective to run agentic AI in a large organisation. How important is that and what are the implications for talent and talent acquisition?
Lia Manafova
You know, going back to the question you asked me before, I think this was one of the things we wanted to answer when the organisation that I’m a part of was being created, because with AI right now, people are focused so much on, just AI, just the word of it, just the hype of it, the, oh, I want AI, just make AI do things. And so what we wanted to do is we wanted to take a step backwards and say, back and say,
How do we make this happen in a practical way? And it felt at the time that as organisation, we weren’t ready for it because, there was so much data and processes and content in a background that just wasn’t there to give us trustworthy outputs for AI to use. And so that’s one of the things that we are focusing on at the moment. Yes, we are doing the exciting AI project, especially in a talent world where we’re focusing on our, we call them Sanofians. So we are all Sanofians. So we’re focusing on our Sanofians to actually bring them innovation. But at the same time, we are looking at whether our data infrastructure is ready for it, whether we actually have a true source of truth for our data.
And that people are using our systems in order to input the data that’s then used by AI. We’re streamlining and simplifying our processes as much as possible. I actually have, this is the case for everyone in the organisation, but I have a counterpart that’s focused on global processes and standardising and simplifying our processes and talent. We’re also focusing on simplifying our integration infrastructure. So my other counterpart on a digital side is looking at that and saying, why do we have so many integrations? Do we really need them? Does the data really have to flow back and forth so much? And what does that mean in terms of having the correct data right there immediately at the right time?
What we have discovered from it, probably one of the things that we’re seeing already is that.
About five years ago, it made sense to have the best of stack for technology for everything versus now when you’re preparing for AI, you have to have a trusted source of truth. And so the strategy changes a little bit to then say, should we have an anchor product that keeps all of our information, that is, our centre of where things go into and flow out of, that’s being adopted and used by Sanofians, and that reduces our integration dependency. And so this is where our strategy has shifted today from a lot of technology to just a trustworthy source of truth and anchor product.
Matt Alder
So tell us a little bit more about that. What is the product, the thing that’s sitting at the centre of all of this?
Lia Manafova
Yeah, for Sanofi, it’s several ones because we’re not necessarily…
Thinking that we’re going to solve everything with one specific product, but, we want to decrease our dependency on as many products as possible. And so in my world, which includes talent acquisition and talent management and skills for us, that’s Workday. First of all, our core employee data already sits in Workday. We have gone through several, implementations and sort of change management efforts where we have asked our employees to go on Workday and continuously update their information, which has been successful in some cases, especially when it comes to skills. And so we already have the data in there. So what we thought is, why not just simplify the processes in Workday and continue sort of relying on it as our source of truth. Additionally, what has helped us is some recent changes that’s been happening in Workday because where five years ago, we may have been missing some important links and some important features. Now with acquisition of Paradox, of HiredScore, we know that, there’s a lot of functionality that’s available and a lot of AI-focused strategy that they’re relying on that can support us and our processes and our needs.
Matt Alder
And tell us a little bit more about how you work with Paradox.
Lia Manafova
So Paradox is a recent one for us. We actually, as a part of our talent operating model that we went through recently or in the past couple of years.
One of the biggest dependencies for us was to find an interview scheduling solution. Previously, we have relied, again, a lot on RPOs, so we didn’t have to think about it much. But when it was time to take things in-house, we realised that in some cases, it takes us from 30 minutes to up to an hour to schedule an interview. Now, imagine multiply that by 20,000 hires that we do per year and all the candidates that we have to screen. It’s quite a lot of time. So if we are bringing this now in-house, how do we actually manage it? So from even my first day of joining Sanofi, I kept hearing about interview scheduling nonstop. It was also a place where we could show the most value delivered as well. So we started looking at different vendors. At the time, Paradox wasn’t a part of Workday, but already they stood out as a platform with very strong Workday partnership. And so we ran an RFP where we screened several different vendors, and Paradox really stood out for us not only because of their partnership with Workday which to us meant very strong integration but also,
The user experience was already very much on Workday, which meant that our recruiters don’t have to go to several systems to do their work. The experience was there in a flow of work as they were going through Workday recruitment process.
And the data set in Workday, which was a big one for us, again, decreases and simplifies our processes and integration dependency. And the user experience with Olivia was a very nice one. We don’t call our interview scheduling AI Olivia. We call it Sanofi Assist. And so for that, it was a very sort of futuristic and nice and easy user experience. Obviously, with a new generation entering the workforce, they don’t want to be sitting there at their email trying to schedule an interview. They just want somebody to text them and say, hey, we’re scheduling an interview for you. Are you available? And all the busy work happens in the background when done by AI. And so that one was a big one for us. We started our implementation journey and literally like a month ago, a month later, it was announced that it’s now a Workday company, which to us was a bet that we took and it worked out in our favour.
So, we’re now further down in our implementation journey, and I can talk a little bit about that as well.
Matt Alder
Yeah, please do. And I suppose also reflecting back on what you were talking about earlier in terms of the change management around transformation, tell us about the implementation, but also the adoption and the change management and everything that’s come around it.
Lia Manafova
In terms of implementation, we specifically chose a pilot to start with, and that was for a reason. We wanted to make sure that it gets adopted from the start. And so when we thought about what could be our pilot, we decided to focus on our hubs. Hubs are where our talent services and talent support, including interview scheduling function, sits. And so if that’s the case, then why not start with the people who we’re going to be bringing in to use Sanofi Assist or Paradox product from the start to then, you know.
Brag about how amazing their experience was, and then be able to start hiring the candidates and sort of promote the product itself in an office. We also noticed that, and that wasn’t necessarily on purpose, but a lot of people who were coming in as a part of this function used Paradox on their previous company. So what we did from the start is we asked them to join our core project team and share their experience and help us shape how Paradox is going to be set up for us. They brought in a lot of great ideas. They challenged us as well in things where maybe we were promoting a little bit too much automation where we weren’t ready enough for it. And I’m thinking specifically of an example here where I was the guilty party. I remember I was sitting in our design workshops and the question of automated versus semi-automated interview scheduling came up. And I remember thinking, why do we need a semi-automated? Just press a button, make it work. Like you don’t need to sit here and think, you know, and configure things.
And I remember there was a lot of pressure to say, no, no, we want to keep the semi-automated option.
In the future, now looking back on it, you know, in the future that helped us with adoption because people did have that option to then go and say, you know, no, maybe I need an additional round of interview where I need to invite someone else to interview who wasn’t supposed to be a part of it originally. And so while a lot of people have been using automated option, having this semi-automated has helped us with adoption as well.
We went live with hubs at the start of the year. The adoption has been incredible. I mean, in one of our hubs in Hyderabad in India, the adoption has been 93%, in the first three months alone.
One of the numbers that I had to triple check before I said it to anyone, but that’s true, is our interview scheduling went from 34 minutes down to eight seconds for automated scheduling, which is a very easy, it’s a very easy return on investment to show, like it’s a very, you know, the platform does the job for me.
And so what we did right after we were done with hypercare which we ran for six weeks is we went with global implementations so we’re now just getting ready to go live globally.
What we have done in the meantime as well as we have worked with our pilot population to say, okay, what couldn’t we do for go-live? What couldn’t we squeeze in? Let’s continue improving the product and let’s continue enhancing it in partnership with Paradox to actually get us to the point where, you know, where you think it’s perfect and it works for you. We’re now going to go through the same exact cycle globally. Obviously, when you do the global rollout, there’s way more wishes and there’s way more requirements that the command languages start playing a big role. And so we’re going live soon with a product that we think meets our needs. And then we’re going to continue enhancing it in this partnership with Paradox. So that’s sort of the short term that’s happening this year. But we also have long term plans as well.
Matt Alder
Cool. And can you tell us a bit more about your long term plans?
Lia Manafova
I was hoping you would ask.
And so long term Paradox plays very much into our strategy that we have with Workday being our anchor product. Right. So when we looked at Workday on its own, like I said, maybe there are some functionality that really Workday didn’t have, wasn’t really even planning to develop. Now you have Paradox, now you have HiredScore which are additional functionalities available and products that are available in Workday ecosystem. So when we are comparing ourselves, in terms of our needs either current needs or future needs we don’t no longer have to just say well Workday doesn’t really play in this space so let’s go and look at someone else. We are now looking at what else can Paradox do, the AI assistant on a career side potentially or their assessment or potentially,
just recently I had a conversation with someone who was telling me that wouldn’t be nice if there was like a CRM that was specifically for frontline workers or blue collar, contingent workforce and I was like wait hold on let me see what’s available. And so all these things, they have expanded our options. So I don’t necessarily know where our partnership with Paradox is going to lead us, but it’s really good that there is these options available that we can review now without leaving the Workday ecosystem to go and look at somewhere else and.
To then have to think about, again, about integrations and the data and the adoption and all that that comes with it.
Matt Alder
You mentioned sort of ROI there and kind of a really obvious saving in terms of time. What else do you look for in technology products in terms of results that can be translated into ROI and business value that people kind of outside of the talent function would understand?
Lia Manafova
Specifically, the one that I found very relevant in Sanofi space is, I’ve noticed that our recruitment people, and I think that’s a general idea maybe about all the people who deal with recruitment, but they’re very practical and very sort of, you know, very practical in the way they do their recruitment. I need to go through certain steps. The less you’re going to make me go through any administration steps, the better I can be as a recruiter to focus on being an advisor to the candidates and to actually create a great candidate journey. So what I’ve noticed in the TA space for us is that when we implement a platform where somebody has to exit their system and then go and log into a different system to do their work, the adoption drops. And so while Paradox provides an opportunity to exit and go to Paradox to review maybe some of the analytics,
to be able to provide a bit more flexibility to recruiters, in a day-to-day work, you really don’t have to do that. And so that helped us with adoption a lot versus if I’m comparing it against something that we maybe implemented where you do have to go to a different platform, especially if it’s not a required step. Like it will help you, but it’s not a required step.
The adoption just drops. So this being able to work within the Workday for everything end to end really helped us with adoption of the platform.
Matt Alder
Just a final question about the future. So we’re already seeing AI doing really interesting things and driving a lot of change in talent. What do you think the future is going to look like? I mean, if we had this conversation again in two years’ time, what would we be talking about?
Lia Manafova
User experience, I think. I remember it was like two years ago or maybe a year ago when I went to Workday Rising, and they had a phrase, you know, there’s always this like hype phrase in the market. And the phrase that they were using was future of UI is no UI. And so I think that’s become more and more relevant these days to where the technology kind of comes to me versus me having to go and log in and do something in it. And it meets me at my day-to-day workflow. So it could be my Teams. It could be maybe it’s becoming more and more popular, in organisations to build their own internal AI agents. So maybe it’s an agent who is approaching me proactively, sort of like my digital twin meeting me in a workflow. And I think in two years, it’s gonna be more and more relevant because the risk that we’re running is that unless we go that path, we’re gonna end up in the same world that we are today with too many technologies, but now it’s gonna be too many agents. And so being able to sort of bring ourselves to, okay, what’s our agent source of truth?
Surface the information for me there, knowing whether I need this information as a recruiter, as a manager, as an employee, maybe a candidate and all the magic of, you know, the processes and the flows happens in the background. And I don’t even care whether I’m working with a Workday system right now or Paradox or anything else. I just know that I’m interacting with my digital twin and doing what I need and supporting me.
Matt Alder
Yeah, I completely agree with you. It’s like one agent to rule them all because this has to be simple for the adoption and for it to work. Lia, thank you very much for talking to me.
Lia Manafova
Matt, this was a pleasure. Thank you so much for inviting me. It was a really nice chat.
Matt Alder
My thanks to Lia. You can follow this podcast on Apple Podcasts, on Spotify, or wherever you listen to your podcasts. You can search all the past episodes at RecruitingFuture.com. On that site, you can also subscribe to our weekly newsletter, Recruiting Future Feast, and get the inside track on everything that’s coming up on the show. Thanks very much for listening. I’ll be back next time, and I hope you’ll join me.
