Can You Build AI With Ever Increasing Fragments of Data Stores? – Kyosti Laiho

Building and using Artificial Intelligence and advanced analytics is gaining momentum, for relatively obvious reasons. As usual with any new data processing systems, it needs to build on data that is relevant for the purpose. At the same time, organizations keep struggling with managing the legacy – increasing set of old and new application databases, various data models, number of data marts and warehouses, big data evolution and use of increasing amount of open source solutions. At the same time, the users of these systems seem to asking for simple Excel sheets and self-service access to build their own views, to cope with their daily work, on top of everything else. So, where is the relevant data for AI and advanced analytics, how fragmented is it? What kind of data architecture and policies supports you the best – should it be controlled centralized data, a laissez faire approach, or a managed combination of both?