How data streaming transformed an HR platform provider

How PeopleSpheres built a real-time HR platform

With a decade of trading under its belt, HR software-as-a-service (SaaS) provider PeopleSpheres had a legacy data problem. Its data processes, formats and technology were holding the French headquartered company back. Legacy data is a challenge for organisations, young and old, with CIOs needing to tackle their legacy data in order to deliver on digital transformation. 

Fabien Gache, CTO at PeopleSpheres, shared how he and his team have overcome the data legacy at a HotTopics ​​C Suite Exchange event alongside Will La Forest, Field CTO at Confluent. PeopleSpheres provides a technology platform to human resources (HR) leaders that brings together all of the technologies they use to serve the needs of employees and the organisation. 

Typically, HR teams are using tools for payroll, recruitment, training, and absence recording, and each tool is from a different vendor. "It is a web-based platform where all employee data and documents are stored, managed and accessed in one place," Gache explains. "So it is the single source of truth for HR data inside the organisation, and it integrates to a wide variety of third-party HR tools." 

Legacy data issues  

HR users expect information to flow smoothly across the ecosystem PeopleSpheres provides, the CTO explains. The integration layer of the technology is therefore vital, without it HR leaders face data chaos, and often build work-arounds between systems in order to get their job done. 

Gache faced his own data chaos when he joined the organisation in March 2019. "Our connectors relied on ETL (extract, transform and load) batch processes, so it was really inefficient, as the data was processed on a scheduled basis, even if nothing had changed. So it was slow, resource-heavy, causing bottlenecks, and not aligned with the platform vision of the PeopleSpheres concept." 

The data chaos was further exacerbated by the lack of data standardisation, leading to unreliable information. There was a latency issue, further slowing down the performance of the platform. Like many of his CTO and CIO peers, Gache was also struggling with a shortage of developers.  

An event-driven model 

It was clear to the CTO that a new approach was required to tackle this quartet of challenges. Gache set out and has succeeded in creating an event-driven architecture through the adoption of Apache Kafka as the event streaming platform technology. With this new architecture, PeopleSphere has developed Connectivity, a framework and new product for real-time analytics and the management of data sources. 

Managing an event-driven architecture with a small team was going to be a challenge, so the CTO formed a partnership with Confluent, the data infrastructure leaders. Gache set up a proof of concept (POC) to analyse the various data infrastructure technologies, and how they could help his team move to an event-driven architecture. He assessed the technologies for adoption, cost of management, performance, and ease of training. "We chose Confluent because it combines the strongest technology with a real managed service. I didn't want to waste time and money managing an on-premise service when I could be developing a new product and features," he says of how the partnership has enabled the CTO and their team to focus on products for customers. 

Business benefits

Customer satisfaction was a central key performance indicator (KPI) for the partnership with Confluent and the new data model. New data flows can be delivered in just a couple of hours, whereas before they could take 15 days, he told peers. "This was just an incredible transformation for the customers." He says data is in motion, meeting the offering and expectation of customers in real-time, and the partnership has reduced the complexity of running a Kafka environment in a small firm. 

Ultimately, the customers of PeopleSphere are benefiting. A change in personal information by an employee is shared automatically across applications such as training and payroll. 

Data-driven future

With data in motion, organisations like PeopleSpheres will be able to utilise artificial intelligence (AI) in their technology products and business processes. "Companies compete on their ability to leverage their data," says Will La Forest, Field CTO at Confluent. "You are making your data trustworthy, reusable and fresh." 

La Forest says organisations must focus on the Return on Investment (RoI) that their data can deliver if they are to succeed with an event-driven architecture or AI. "High-quality, fresh data can transform all aspects of a business, but it really starts with an RoI based on a business problem in order to get the process started." 

For further information, you can download Confluent's 2025 Data Streaming Report here. 

Advice 

• Interest in AI can get CIOs the much-needed investment and resources for data infrastructure modernisation
• Define the RoI of data infrastructure modernisation, and its impact on the customer 
• Identify a number of business cases 

Join the Data Visionaries community