Ataccama, a data governance vendor best known for its data quality (DQ) and master data management (MDM) tools for big data, is today announcing a new release of its Ataccama ONE platform that the company says heavily leverages artificial intelligence (AI). In the current enterprise environment, data governance and data management are increasing in importance, visibility and complexity, but they have historically carried a certain baggage of tedium. AI -- and automation in general -- are thus competitive necessities for data governance and data management, as they help customers stay data-driven and regulation-compliant, as well as data leak- and breach- vigilant.
One platform, many modules
The Ataccama ONE platform includes modules for the aforementioned DQ and MDM capabilities (an interface for the former is pictured in the figure at the top of this post), as well as for data catalog, business glossary and business rules management on the governance side. Data management-wise, ONE sports modules for data integration, preparation, discovery and reference data management.
Ataccama likens its AI capabilities to those found in autonomous vehicles and, in fact, use the "self-driving" moniker to describe the functionality. Rather than adding these self-driving features to each of ONE's module in a piecemeal fashion, Ataccama said it chose to add a central AI module that, in turn, is used by each of the others. The AI Core module, as the company calls it, provides services for data fingerprinting/classification, anomaly detection, domain detection, relationship discovery and AI matching. Ataccama says it uses active learning technology and is built with explainable AI as a core tenet.
Three pillars and a get-out-of-the-way credo
The AI Core module is one of the three pillars to Ataccama's self-driving features. A second one is metadata, which, in Ataccama's use of the term, goes well beyond mere database/table/dataset schema information, and encompasses semantic details about the data as well. The third pillar is policy management with automated enforcement, which address access control and data retention, among other data governance and management constraints.
Ataccama says its self-driving capabilities are based on the principal of letting customers define the requirements (the "what") for functions such as data cataloging, quality, cleansing, deduping, and MDM, while letting the system determine the implementation (the "how") of meeting those requirements. Importantly, though, representatives of the company also explained to me in a briefing that customers may apply any detail of the implementation manually, should they wish or prefer to. In other words, the automation is there to support data stewards, IT personnel, data scientists and data engineers (as well as line of business managers and business analysts), but not to get in their way when they have a precise management/governance measure in mind.
Ataccama One Gen2 is running in Beta, at certain customers. The company says the release will hit general availability (GA) in February, 2021.