Zscore: Solving Big Data's Dirty Problem

Enterprise data is great until you actually try to use it in an AI or BI initiative. Imperfect sources of data, myriad systems performing operations, functions creating versions of data in silos, all tend to compromise data integrity - creating what Zscore calls ‘Dirty Data’. Organizations surveyed by Gartner have stated that on an average dirty data costs them USD 14.2 million annually. Extracting value out of dirty data is an increasingly important and overlooked task that can help prevent expensive mistakes. Deleep Murali, an MBA graduate from Katz School of Business at the University of Pittsburgh, experienced this problem first hand on the many big data & analytics projects he worked on. Determined to find a solution, he decided to work on an AI-based Data Management platform that would enable businesses to improve their data quality and insights. Soon joined by his ex-colleague Ayush Banka and college-mate Rohit Kalro, the venture came to be known as Zscore Technologies.

ZScore's Data Management platform is built around a proprietary process called CLUE. It works with data from any source and allows enterprises to create workflows mirroring their business logic, making it possible to extract personalised value from data. In this ML-driven platform, clients have the flexibility to introduce custom Python or R models on top of proprietary Zscore models. This allows internal data scientists and business analysts to create their own version of a pure master data source to play with. “One of the aspects of data management is that everyone focuses on good data and ignores the dirty part, but 40 percent of data is bad. So we process the dirty data too and convert that into good data”, elaborates Deleep Murali, Founder & CEO at Zscore.
Addressing Issues from the Ground Up
In their conversations with enterprises, Zscore discovered that business analysts and data scientists actually spend 80-90 percent of their time making data usable. It turned out to be a major hidden expense for organizations, considering the costs of talent and of operating highly skilled teams. Apart from technical issues in the data, there were a host of other business imperatives such as timeliness, access to data and adequacy of information at hand to perform a task. Added on top were the reporting, auditing and compliance requirements. It was clear to the product team at Zscore that mere data cleansing was not going to solve the problem. With these requirements in mind, the team developed advanced features like business logic led data quality transformations, deep data audits, master data management, and a data firewall.

Zscore's Data Management platform is built around a proprietary process called CLUE

While working with their clients, the team also established that for initiatives like AI to succeed within organizations, it is critical that businesses trust their data and establish a sustainable data practice & strategy. However, few enterprises meet this condition, most are unfamiliar with newer technologies and options for monetizing their data assets. Such organisations would benefit from Zscore's ‘AI as a Service’ - a consulting practice providing analysis, education and appropriate guidance, aimed at enabling firms to emerge as leaders to dominate their peers and the market in a data-centric world.

With further R&D into the Data domain, Zscore as an innovator is sure to present some stellar products and solutions in the years to come.