The next step in the implementation was taking a scalpel to the existing infrastructure. Before you build a new house, you need to understand why the old one isn't serving its purpose anymore. In Mobileye's case, the legacy system we were replacing was Blossom.
I still wasn't configuring Docebo yet. I was continuing the deep-dive mapping process, but this time, the focus shifted from the people to the plumbing.
Tracing the Data: HRIS & Onboarding
A successful enterprise LMS is never a standalone island; it is an extension of the HRIS (Human Resources Information System). I started by tracing the data flows between the HRIS and the old LMS to understand exactly what information was being pulled, and more importantly, why.
I mapped out the entire onboarding journey for new employees. When a new engineer or analyst walks through the doors at Mobileye, what is their day-one experience? What are their immediate needs?
The answer is simple: they need zero friction. A new hire shouldn't have to search for their mandatory training. The data coming from the HRIS, their role, department, location, and seniority, needs to act as an invisible engine that automatically curates their first 30 days. Understanding the exact data points currently available in the HRIS gave me the blueprint for the automation rules I would eventually build in Docebo.
The Blossom Bottleneck
Once I understood the onboarding flow, I dove into Blossom to analyze the existing content architecture. I needed to see how the organization stored and categorized its knowledge.
It became immediately clear why employees were frustrated.
The content architecture in the legacy system was rigid and deeply nested. It operated on a "file cabinet" mentality. If a developer needed to find a specific technical update or a compliance module, they had to click through layers of folders, hoping they guessed the right path. It wasn't intuitive, and it wasn't built for speed.
The Radical Shift
One of the biggest mistakes organizations make during an LMS migration is the "lift and shift", taking a broken architecture and simply moving it to a shiny new platform. If you put bad architecture into Docebo, you just get a more expensive bad architecture.
I realized that to make this implementation successful, we needed a radical change in how content was structured. We had to abandon the folder-based mindset.
Docebo allows for a much more modern, Netflix-style approach: using channels, intelligent tagging, AI-driven search, and customized widgets. Instead of forcing employees to hunt for content, the architecture had to be redesigned so the right content found the right employee at the right time.