Overview
About Vitalliance
Vitalliance is a French network of 140 home help and personal services agencies. Its 7,000 qualified Caregivers support elderly people or people with disabilities on a daily basis to enable them to live well at home.
Opportunity | People are at the heart of Vitalliance's work
People are at the heart of Vitalliance's work. For 20 years, the home care expert has enabled elderly people and people with disabilities to continue to live in their own homes thanks to personalized support from qualified employees. Firmly established with 140 agencies in France, Vitalliance has 7,500 Caregivers serving its 11,000 customers, 24/7. Whether they need help with meals, hygiene or relationships, this assistance offers real emotional and physical support to the most vulnerable members of society.
The network of home care and personal service agencies helps 11,000 customers who are highly dependent, due to age or disability, to live well at home. Whether it's a few hours of household chores per month or 24-hour support for those most at risk, Vitalliance's 8,000 employees, spread across 140 agencies in France, represent essential physical and emotional support.
A heterogeneous and complex IS
However, technology is essential to enable them to carry out their various missions. Vitalliance, which has just celebrated its 20th anniversary, chose from the start to build its own computer system in order to have a tool that suits its needs perfectly. The company has therefore developed software internally to manage all of its activity: customer management, recruitment, contracting, scheduling, operational monitoring of missions, billing, cash, etc. Some forty people make up its IT team, focusing in particular on developing and managing this critical base.
At a time when the cloud did not exist, Vitalliance deployed its own internal data center, before outsourcing its infrastructure to a colocation site a few years later, before experimenting in the cloud with Amazon Web Services. When Thomas Lier, Head of Build and Architecture, and Matthieu Brun-Bellut, CTO, joined Vitalliance 3 years ago, they therefore discovered a heterogeneous information system. “Some parts of the IS had been migrated to the cloud, others were not, which created complexity,” explains the latter. We quickly realized that we did not have the resources we needed to manage a data center: upgrades, security, backups, capacity planning, etc.”
Solution | On the way to Infrastructure-as-Code
The IS must also withstand peak loads, especially at the end of the month, when the billing process consumes a lot of resources. “It was obvious that we had to switch to 100% on the cloud to gain flexibility and agility,” says Matthieu Brun-Bellut. We also wanted to move towards containerization, continuous integration and automatic deployment, and the Amazon Web Services (AWS) cloud seemed to us to be the simplest and most mature solution to achieve this. Our business also required us to use an HDS certified partner, which is the case with AWS.” The decision was therefore made to migrate the entire IS, in an Infrastructure-as-Code (IaC) approach.
After restoring order to the services that had already been migrated (segmentation of user accounts, segregation of prod/non-prod resources, governance with AWS Control Tower, partitioning networks with Amazon Virtual Private Cloud), Vitalliance redeployed the services which were hosted in its data center and based on a VMware environment on the AWS cloud. The final, and most critical, step is to migrate this core business software, consisting of a SQL Server database, Microsoft IIS servers, and a file server. “We switched the DB to Amazon RDS with a simple export/import, via a backup on Amazon S3, explains Thomas Lier. On the server side, we no longer have fixed virtual machines. We moved to Amazon EC2 with auto scaling and application AMIs, as well as all essential services such as Amazon VPC or AWS Route 53. Even certificate management is automated. In any case, there was no added value for the company in administering these types of services.”
A more resilient organization, and more efficient applications
The entire migration was carried out in 6 months, between early 2023 and the summer of that same year, with no interruption of service for users. “100% of the infrastructure is now deployable and modifiable with code,” says Thomas Lier delightedly. “Now we can create an environment in a few minutes instead of 1 to 3 days previously.”
Another benefit of the migration is that backup management has been simplified considerably. “We have defined policies on AWS Backup,” explains Thomas Lier. “Now, all we have to do is assign a tag to a resource and the associated backup and retention policy is automatically applied. This simplifies compliance while taking the burden off team management. And in addition, it prevents us from keeping certain data longer than necessary, which helps to reduce costs.” As a result, the organization is now much more resilient. “Before, we estimated the time needed to rebuild the IS in the event of a major problem at around a week,” continues Matthieu Brun-Bellut. Today we can do it in half a day. And a half-day of interruption is a delay that we can manage from a business perspective because all activities for customers are planned in advance.”
But for users, the most visible benefit is undoubtedly to be found in terms of response times. “First of all, the entire IS is streamlined in the same place, which reduces latency times,” highlights Thomas Lier. Then, the RDS Performance Insight service allows us to see which queries generate the most load on the database in a very simple way. Thanks to this visibility, we immediately identified 2 or 3 problem requests that we had not identified until then. So we were able to optimize them to improve response times.” At the end of the month, when all the pay and billing processes had to be launched, some teams preferred to work in the evening, or even at night, to avoid slowdowns. “Since the migration has been completed, it is no longer an issue and no one works at night anymore”, explains the CTO.
Outcome | Putting data into the hands of businesses
For the past six months, Vitalliance's IT teams have therefore been able to focus on projects with higher added value for the company. And that value, today, is in the data. This is why they very quickly became interested in the Amazon QuickSight business intelligence service. “Information on our caregivers and their interventions was stored in an unstructured MongoDB database, while customer data was in a SQL database, and therefore structured. It was complex to cross reference the two, and the businesses had to go through us to get the information they wanted. Now everything is centralized and they can generate the reports they need independently.”
The IT teams don’t intend to stop there and are currently working on two other projects related to artificial intelligence. First of all, a machine learning project, which aims to assess the risk of an employee leaving. Predictive analysis could allow Vitalliance to identify the first signs of employee disengagement and thus take steps to understand the source of any potential problems and improve quality of life at work for the individual. “We started this project in September and we already have a model that has a success rate of over 90%, reveals the CTO. If we manage to reduce turnover by 20% thanks to this, it will be a huge gain for the company.”
Generative AI in the crosshairs
Second project: facilitating the search for information through generative AI. Each agency in the Vitalliance network has a 400-page manual that contains a large amount of information, especially with regard to legal or HR matters, used to manage employee or customer contracts. “However, many employees prefer to call a support team or their manager directly rather than searching for the answer to their question in this manual, which can actually be a bit indigestible,” acknowledges Matthieu Brun-Bellut. But, as we open between 20 and 30 new branches per year, our teams spend a lot of time answering the same questions.”
Vitalliance therefore used Amazon Bedrock to start from an existing model and adapt it to its needs in order to offer its teams a generative AI capable not only of answering their questions, but also of specifying the source of its information. “The explainability of AI is an important point”, notes the manager. “Our teams must be able to know whether the answer comes from a collective agreement or a company agreement, for example, because one takes precedence over the other. The question we are currently asking ourselves about this project is whether we are going to provide a single crawler for all teams or several more specialized crawlers.” The only certainty is that, for Vitalliance, innovation now has no limits.
It was obvious that we needed to go 100% on the cloud to gain flexibility and agility.
Matthieu Brun-Bellut
CTO of VitallianceGet Started
Did you find what you were looking for today?
Let us know so we can improve the quality of the content on our pages