Sanofi – 2023 Digital Award Winner Shares Details of Their HR Transformation – Transcript

By Vin Kumar, Maria Elena Castillo, Christian Mueller, and Christine Van Heertum
February 27, 2024
Season 5, Episode 15

Christian Miller:

The real highlight here is the employee satisfaction. Imagine you will be promoted and the job title is not updated. The payroll is not updated. Or let’s say it needed to be maintained manually in the background. In doing the analysis, there were different actors in the process not giving their approval, a lot of country-specific configurations and a lot of small process variances.

First idea we had was to look into that data with a technology of prediction. So, for this, we played a lot with artificial intelligence and machine learning. The result was where we were able to identify different risk indicators of the process. We presented this to the operational team. What you see here are the top five processes, which will fail this week. Are you ready to bet?

Announcer:

Welcome to The Hackett Group’s “Business Excelleration Podcast®.” Week after week, you’ll hear from top experts on how to avoid obstacles, manage detours and celebrate milestones on the journey to Digital World Class® performance.

Vin Kumar:

How has pharmaceutical giant Sanofi used technology to transform the human resources operation and improve the way they manage talent? I’m Vin Kumar, a principal at The Hackett Group leading our AI and digital operations. And on this week’s “Business Excelleration Podcast,” we talk to the Head of Process Intelligence Christian Müller, Sanofi’s Workday Employee Experience and Product Manager Maria Elena Castillo, and Sanofi’s People Services Head of Employee Lifecycle and Solutions Christine Van Heertum about their cutting-edge effort. Sanofi won a Digital Award from The Hackett Group in 2023 for their impressive work. Welcome to you all.

Christian Müller:

Yes, thank you very much for having us here.

Christine Van Heertum:

Thank you. Thank you.

Maria Elena Castillo:

Hello everyone. It’s a pleasure to be here with you to share our great success.

Vin Kumar:

Christian, maybe you want to start with just a quick introduction or the description of the process that you were trying to improve using technology.

Christian Müller:

Yes, absolutely. The process we focused on was the job change process. The job change process is related to all kinds of changes within the employee life cycle – promotion, demotion, lateral move to a new position. So really a process, which is executed 25,000 times a year. It’s a process, which meets each employee sooner or later within their career. We tackled this process, improved this process, and, spoilers, we improved the process duration time by 83%.

Vin Kumar:

Maybe do we want to start with what the challenge was, what you were trying to solve, and then we can talk about what the solution was, what the value it delivered, and future plans. That’s what we have for the session today. But I will open up a question to who wants to answer – talk about what the challenge you were facing that you were trying to address.

Maria Elena Castillo:

So yes, part of the challenges that we faced before we started with the project, first of all there was a lack of full picture. We knew that the reality was that we had the strong need to improve the process – to simplify it. We knew that it was too complex – too long. There was a lot of actors in the middle of the process. But the issue that we had was the complexity because Sanofi is present in a lot of countries around the world – more than 70. The reality [is] in those countries [they] work different from one country to the other one. So we had a super bottleneck, but we didn’t know how to attack that giant.

Another thing that was very challenging to start with the project was the lack of governance. We didn’t have a specific team in charge of this end-to-end process for a worker – for an employee who had the need to move positions or to be transferred. So we had to join synergies between HR, business HR, HR services, digital. There was a combination of all these teams together, combining the technology with the functional knowledge from the operational team. So then there was an HR transformation [that] acted as facilitator of all these transformations. But we will speak about it after we share the solution.

Christine Van Heertum:

Thank you, Maria Elena. And if I may complete, one of the biggest challenge was compliance because data was not updated in due time. Of course, it had a lot of impacts not only on compliance but on payroll, because data needed to be changed with retroactive dates, meaning a lot of corrections in payroll. Employees were unsatisfied because the changes were not visible in the system in due time neither. And, of course, it impacted as well, for example, the approval flows. If an employee was still reporting to the previous manager, while it should have been approved by the new manager, so it was causing really a lot of dissatisfaction – a lot of workloads of correction – sending back the processes to the person who needed to correct it. And also, as I mentioned on payroll, and that’s where we really knock at the door of process excellence, telling them, “OK, Christian, do you have a solution to visualize the loops and the bottlenecks?” I will let Christian explain, of course, the solution he was offering us to visualize the bottlenecks and the loops.

Christian Müller:

Yes, thank you very much for this. I think the real highlight here is the employee satisfaction, as you said, Christine. So imagine as an employee you will be promoted and the job title is not updated, the payroll is not updated, or let’s say needed to be maintained manually in the background. A huge risk on employee satisfaction, so really something not one of us wants. And having also the team in the background, being very reactive instead of proactive in the daily operations. That was the situation we found when we started our analysis. So we looked with our team and what we found very fast was data, of course. The underlying system was Workday, and the processes are documented pretty well with time stamps. So we decided to look into the process with a technology of process mining, extracting the data, visualizing the process and doing the analysis. What we have seen here is that there were different actors in the process not giving their approval, a lot of country-specific configurations, different initiators initiating the process. So the process was considered as management self-service. But in real life, we have seen that there are many other actors starting the process and that the managers are the smallest population initiating these processes. And we have seen a lot of small process variances – process going into loops and bottlenecks within the workflow.

So the first idea we had was to look into that data with a technology of prediction. The idea was to take the data from the past two years and applying prediction analysis to predict how the cases will run in the future, so a little bit like the weather forecast. We wanted to see, based on data from the past, how the particular process will behave in the future. So, for this, we played a lot with… And this is also what we wanted. We wanted to play with artificial intelligence, and we did both in parallel. We worked with machine learning, artificial intelligence and also with a statistic model to see how we can get the insights out of it.

Yeah, so the result was a very nice dashboard where we were able to identify different risk indicators of the process. I remember pretty well when we presented this to the operational team. So usually when I’m showing results of process mining projects to our customers, everyone is super relaxed because it is past data. Whatever we do, we cannot change it. But it was very nice when we showed, “Hey, this is a dashboard and what you see here are the top five processes, which will fail this week based on data. Are you ready to bet? We can tell you the risk indicators. We have identified the background.” So the motivation was totally different, and we had some iterations with the operational team, getting feedback from them to improve this model and to even make the prediction and the forecast more reliable.

What we also figured out was that I would say there were top five exercises, which were solving 80% of the problems. So we had a very nice brainstorming session where we were dreaming about automation, bot automatic emails, reminder emails to be sent. But that was also the moment where we looked at each other and we said, “Hey guys, what we are doing here?” So we really had a lot of fun with this AI and analytics topics, but what we really see here is that the process has really a very, very bad design on the ground, and we should not put now more automation prediction and all of this pretty cool stuff on it when we can solve the problem on the ground.

It was not easy for us to stop all of these technical analysis things and artificial intelligence things because we, as I said, had really fun with playing around with that. But the answer for all the problems was a real redesign – a user-centric real design on the ground of the process.

Vin Kumar:

And Christian, that was what was very interesting when we are looking at when you won the award was you did all this prediction. Everybody wants to do these things – predict a problem before it occurs. You’re using all these cool technologies. But then you resisted following that path and coming back to simplifying it and putting up simplification – simplified process and technology – to enable that simplified process rather than, quote unquote, “chasing the cool – the next technical solution.” Maybe this is a great segue to talk about what was the solution? Why did you decide on that solution? I know you’ve already started talking about it. I think that will help the listeners to understand how you went about it.

Christine Van Heertum:

Yeah, I think what we can say is really that, as you mentioned, we decided really to stop, step back, clean the table, I would say, and really redesign the process. But for that we really needed to first show to our stakeholders – the deciders – where we need to improve and to simplify.

With the use of process mining, it was really the visualization of the bottlenecks and that was an eye-opener. Without that, I’m not sure we would have been able to change the mindset and tell the stakeholders we really need to simplify – get rid of unnecessary steps – because simplification starts with removing unnecessary steps. And when everything is clean, then we can automate and move to the digital parts of it. But Maria Elena, of course, digital, you can talk better on that part than I will do. Please do.

Maria Elena Castillo:

OK, thank you. So yeah, as Christine was mentioning, this removing steps were part of the challenge that we had to face, because as we mentioned at the beginning, the processes were different country to country. From the digital side, we had to analyze each of the steps, where all those steps were for the different countries and remove localization things. So the approach that we took and the decision was to go for a global approach. We harmonized among all countries how the process was supposed to be. Not only that, but also as Christine was mentioning, we removed a lot of administrative tasks that used to be assigned to HR business partners. And that’s why we were saying that it was part of an HR transformation because sometimes you can put a great tool – a great technology – but if you do not do a change of management plan and you don’t communicate and also you don’t change the roles and responsibilities, it’s very hard to put it in practice and to last this project, to make it real, and to last long and reliable in the time.

So we removed a lot of steps. We took out steps that were in HR’s hands, and we put it and transferred it, not all those steps that were there, and we simplified only into one – to HR services. We empower managers to be responsible for their own data in the system. So they were responsible to keep well-updated organization of their scope. We had to do a strong communication here as well. The change of management again was key because they need to train all managers and to prepare them to what the system was going to look like starting from now.

To do so, we simplified the system in a way that first screens what the manager sees, we simplify with essential information that we knew that manager may know, and we take out all the fields that were very administrative things that manager in general doesn’t know, and we transferred the responsibility to the HR services. We also put at disposal of managers a bot where they were able to ask directly and interact with this artificial intelligence directly instead of trying to look and figure it out how to start this process that we’re talking about.

Christine Van Heertum

Correct. And perhaps what we can also add is that one of the key element as well of the success was really to have all the needed actors around the table asking them as well to test the process before implementing it. Because, of course, if we simplify a process, we implement it without first checking with those who will use it, we might end up with a process which is not sustainable. So that was also part of the approach – to make sure, for example, that the operational team is buying the new process.

Vin Kumar:

Maybe a question for you, Christian, is what technology components did you use to solve for this simplified process? I know it was based on Workday, but maybe you want to share something on it?

Christian Müller:

We used QPR process mining to do the analysis. And when it came to the prediction topic, we also developed things. We looked with Python and played with a couple of other technologies to build the prediction model, let’s say, and loading it back into the QPR tool. What we also did once the final design of the process was made in BPMN, we loaded this as well to the process mining tool and used the conformance functionality of the tool. So really, we had already the as-is picture – how is the process executed in real life? And after the user-centric approach, we really know how do we want to have the process being executed. It was a huge rollout. It took more than a year. So a lot of configuration country by country in Workday. And the technology again helped us to monitor is the process expected how it should be processed? So how is the change management by country?

Vin Kumar:

Excellent. Thank you, Christian. Maybe we can move to what was the value delivered? I think it’s apparent, and I know, Christian, you shared something earlier on the percentage of manager self-care, self-service. But anyone wants to talk about what was the value your employees, your stakeholders, your functional owners got from this solution that you put and transformation you did?

Christine Van Heertum:

Yeah, I can give some figures definitely. As you mentioned, of course, employee satisfaction is much higher because data now is completed in due time by approximately 95%, which is really fantastic. Managers are as well very satisfied to finally be able to do the changes by themselves without the administrative burden. The manager satisfaction is around 75%. But most importantly, I would say that we also freed human resources of the administration by 95%, which means that they can now focus on strategic partnership. This is really key. And there is also, of course, as I mentioned already, the compliance. We have fewer red flags from the auditors when they are checking if the process is followed correctly. And the conformance is also really helping us in fine-tuning some minor rules. So really those are the added value that the simplification is bringing because it keeps bringing the added value.

Christian Müller:

In addition to my role, I’m also a manager. I’m heading a team of 10 people, so I can also share from the manager’s perspective. What I really appreciate was also the redesign of the form of initiating the process. In the past, I remember there were a lot of questions asked from me as a manager I needed to look at first. While later on in the process, people on an HR services and HR side already had this information and were able to edit. The number of approval steps was incredibly reduced, and now it’s two and in some cases was up to eight approvals and really the acceleration of the cycle time. That’s what we also have proven and measured more than 83% improvement in the cycle time of executing a job change process. I can also speak as a manager that there is really a huge improvement. It’s really something the managed population can feel, and it really feels like giving time back to the business.

Vin Kumar:

I know you won the award in 2023. This was kind of put in place. Anything on what, any changes, any upgrades you’ve taken to the version 2.0 and what’s the kind of the future in line for employee services that you’re thinking about? The question everybody’s asking today is, “Are you incorporating any of the AI stuff?” I know you did a lot of things on the predicted side. Anything on the generative AI that you’re planning to add that you want to share and talk about?

Christine Van Heertum:

From a process perspective, I can tell you that we keep trying to improve based on the managers’ and other users’ feedback. We are already looking into another business process to be streamlined and simplified. We keep feeding the Q&A because, of course, there are new cases popping up here and there. So it’s really in constant evaluation, amendments, etc. But as I mentioned, that’s from a process perspective, and I know that Maria Elena is keeping a line on any digital possible improvement on that part as well.

Maria Elena Castillo:

Yeah, definitely. We are also, once we finish with the project, now as a second phase, what we are trying to put the focus on is in make it better – ­the look and feel of the notifications. Because something that we didn’t mention, but it was part of the project as well, not only [did] we remove steps – that is the administrative steps to confirm or to approve – we replaced those steps by sending notification by sending an email from Workday to the actors who should be informed about the process. So now what we are doing is improving at make it fancy the message to be easy to be read by the oldest stakeholders who should be involved. So we are reshaping those in terms of Workday and specifically for the changes that we were mentioning. But with Christine, we are working on other business processes as she was mentioning, and the approach will be something similar – to reduce steps, check if we can replace manual steps by automatic notifications and so forth.

Christian Müller:

On a process intelligence side, we really have tasted blood with all of these data and artificial intelligence stuff. So we are leveraging this lighthouse project also to other end-to-end processes within the company. We are now working together with a purchase-to-pay on the accounts payable process and looking here how we can support with prediction when it comes to processing invoices. We also pretty much look into process insights – so how we can use all of the new opportunities of artificial intelligence to get insights of the process about how to improve, how to measure, and how to manage to make the processes really ready for the future.

Vin Kumar:

From a jury perspective, what we liked about this particular application and why you got the award was … and it’s sometimes very hard to simplify – make it easier to do it and intuitive to do it. It is a complex process in a large organization like yourselves to do these things, and how you are able to just not chase the latest technology – chase a technical solution – but you really simplified it. And to do that, you had to apply a lot of technology to do it. But the objective was all about simplification, making it easy for your managers, for the employee that experienced to do that, and how you incorporate technology behind the scenes they didn’t even know they had. It’s not to do that. So those were some of the reasons why we felt this particular application deserved the Digital Award. We are looking forward for you to be participating in this. It’s a great story that we have. And again, thank you very much for all of you to one, take the effort to participate. Two, share it with our clients to do it. And again, thank you very much for sharing your journey.

Maria Elena Castillo:

Our pleasure. Thanks to you.

Christine Van Heertum:

Thank you.

Christian Müller:

My pleasure. Thank you very much.

Announcer:

Thanks for listening. You can find audio, a transcript of each episode and other resources at podcast.thehackettgroup.com. If you liked this episode, please share it. You can also subscribe on your favorite listening app, so you never miss an episode. We welcome your feedback by rating this or any episode. Or send us an email at podcast@thehackettgroup.com.

The Hackett Group is a leading benchmarking, research advisory and strategic consultancy that enables organizations to achieve Digital World Class® performance. Learn how we can assist with your transformation journey at www.thehackettgroup.com.