A Cultural Shift in Data Democratization Can Improve AI and Analytics Outcomes

By on

Click to learn more about author Tejasvi Addagada.

Democratization is more of a cultural change, that any firm should be ready to embrace. It’s synonymous to any of us having to move between continents with different political cultures. Data democratization requires breaking down data silos across divisions and functions, thus making data available and accessible, as per the access rights.

For example, in a lending institution, a risk function should be able to generate a Key Risk Indicator on home loan modifications, across geographies, loan variants, and timelines. This usually happens with the risk office drafting data requirements (DRD) while a data analyst understands where such data exists in the landscape. An architect then brings it over to a warehouse meant for such purpose. This is followed by a reporting analyst prepping, cleansing, standardizing and generating a presentable report. Days and even weeks pass by, to have such a report available. This scenario changes with democratization – where the risk office will be able to service their KRIs from data that is available and accessible by un-burdened means.

Democratization needs executive sponsorship from the source & consumption landscape, as it requires extensive training of personnel, on usage of business metadata and tools to self-service needs. Democratization across an organization requires data that is integrated and inter-operable across divisions.

The basic enablers of Democratization include reduced Turn-Around-Time to generate insights from AI models, for example to detect fraud, or improve customer journeys.

The other enabler is understanding changes to data and its structures seamlessly across the organization without having to speak to knowledge workers or going through extensive documentation.

Data democratization is a strong concept and it can be solutioned by an internal marketplace. I will speak more about this in my next blog.

Leave a Reply