Transform
a custom service
to a product
and global phenomenon:
Amdocs EMS

Amdocs Error Management System (EMS) was a product that was copied and customized in each implementation. Customization provided a notable revenue stream, but I proposed we could make significantly more money if we sold faster and more per quarter. This simple observation resulted in a new generic product I designed and built. We eliminated customization by drastically enhancing the focus on developing a graphical rules definition module. We incorporated a Machine Learning (ML) module by Gadi Pinkas.

​​​

The primary challenges in this paradigm shift were corporate culture and its evident success, migrating existing customers to the new product without impacting their processes (or resorting back to customization), and moving the development from Israel to North America, where we had no team or knowledge of the existing codebases.

​

I partnered with Sales at the time of inception. This early collaboration was pivotal in the product's success. In less than a year, EMS emerged as the company's most successful door opener globally. The revenue generated from EMS surpassed all expectations, growing significantly above previous years' performance.

​​

EMS significantly reduces revenue leakage and the labor required to resolve data processing errors. The system identifies records that failed to process or failed in transit between systems, presenting them to users for resolution. Users can apply grouping rules for mass resolution. The ML module learns and proposes corrective action as manual work is performed. In future cycles, these ML-proposed actions can be turned into automatic resolvers, optimizing efficiency and accuracy and reducing time to revenue realization.

​

Key technologies:

• web UI

• Java

• Oracle DB

• Supervised learning

• ETL