How Servier set up a truly innovative Digital Purchasing Assistant

Corcentric

“Angie”. This is the beautiful name given to the Digital Purchasing Assistant for the pharmaceutical group Servier, in reference to the pilot of the project, Angèle Cardoso from Méotec, who led this project with great success. Born out of a collaborative effort between the Servier team and Corcentric, this ambitious project aimed to deploy an innovative Digital Purchasing Assistant to a panel of users. Unlike a chatbot, it is a bidirectional assistant, capable of answering users’ questions in real-time as well as pushing task reminders via Microsoft Teams, and can be operated from within Teams.

Sylvain Feraud, Director Group Purchasing Services at Servier, Matthieu Barthelemy, Project & Deployment Manager at Servier, and Julien Nadaud, Director of Corcentric’s Innovation Lab, are answering our questions.

 

How was this project born and what was the goal?

Sylvain: During a lunch meeting, Julien Nadaud came to me with the ambitious idea of a Digital Purchasing Assistant and I immediately imagined the possible use cases for our users, the helpfulness of such an assistant, and therefore the precious time saved that users of the Corcentric solution could dedicate to serve our mission: to provide innovative therapeutic solutions for the benefit of our patients.

By conversing in natural language with the assistant, our 2,500 users could free themselves from using a traditional software interface, get the answers to their questions in real time, and be reminded when actions need to be performed! This is what motivated us to work together on a proof of concept.

 

How does this assistant work?

Matthieu: It works via two channels, full integration into the Corcentric Procurement solution and via Teams without having to connect to the application.

The assistant answers questions about:

 

    • Standard operating procedures: “How do I purchase under contract?”
    • Management rules: “How can I have access to such authorization?”
    • The data set: “Show me all my orders” or “What is the spend of the last three months with this supplier?”

 

When a question is not understood by the assistant, it is sent directly to the local team who is responsible for answering it and improving the knowledge base so that Angie learns quickly!

We were able to appreciate its proactive side that sends notifications in Teams to remind us of our tasks to be performed. As the assistant improves, we discover new use cases for it. It is even able to converse in more than a hundred languages.

 

How did the project go?

Matthieu: Three elements were critical to the success of this project: the team, the choice of the user panel, and the technology.

The first element is the commitment of the team. It required a pilot in the cockpit to set up the project team, of course, but also a rally of our ‘early adopters’, to prioritize Angie’s learnings, and not aim for the moon from the outset.

Secondly, we quickly needed a panel of users large enough to help us better understand different forms of the same question. Therefore, we called on the 500 team assistants, who use the application daily. Tasked by the requesters, they are already familiar with Corcentric’s platform. As they already have the answers to their questions, they are more demanding and know which questions are the most relevant to ask and which questions will be the most frequently asked.

Finally, we started from a blank page, and working with Corcentric, adopted a flexible method to improve the Digital Purchasing Assistant via small sprints. The goal was to “learn as you go”, allow for backtracking, and get the most out of these next-generation technologies.

The project is a real success. Angie continues to learn thanks to the involvement of the team and the assistants, who see a real added value in it every day.

 

What do you see as the potential for transformation?

Sylvain: eProcurement tools are often undergone by business users and are perceived as cumbersome administrative tasks that require a lot of research and clicks. Being able to converse directly with an assistant to buy articles and services can change the perception of this process: we go from constraint to empowerment. We’ll probably be wondering in a few years how we could do without it!

 

What is the future of the Corcentric Digital Assistant?

Julien: Thanks to the deployment of the Digital Assistant, which does not restrict the functional scope of the application, we can understand the real expectations of our users. Our goal is to constantly meet these expectations by using all the technologies available to us today (AI, Machine Learning, etc.). We want to make the Corcentric eProcurement solution efficient and easy to use.

The assistant should be seen as a collaborative expert with access to all internal and external data, but we will also ensure that it anticipates and predicts what needs to be done until it is able to perform a number of tasks on behalf of users who can then focus on their job with confidence.