Glossary
What is data governance?
In short
Data governance is the set of rules and roles by which a company decides what data to keep, who is responsible for it, who may see it and how its quality is checked.
Business and adoption · data governance
MMOOZZGG.com editorsChecked: 2 min read
For a child
In a school library, every book has its own place on a shelf and its own card. The librarian knows which books you can take home and which you can only read in the reading room. If a book gets torn, the librarian repairs it or buys a new one. Without such order, nobody could find the book they need. Companies put their records and tables in order in the same way.
In plain words
Any company has a lot of data: customers in the CRM, stock in the accounting system, contracts in folders, letters in email. Often the same number differs between two programs, and nobody knows which is correct. Rules remove this confusion. Each data set gets a person in charge, and the team agrees where the main version lives, who can change it and who can only view it. Quality is checked too: no duplicates, empty fields or outdated records. It is like a home first-aid kit where every medicine has its expiry date written on it and everyone knows who refills it.
For business
For AI, this is a question of money: a model answers from the data it was given, and confusion in your records turns into confident mistakes by the assistant. So data preparation goes into the budget before you choose a model. Costs depend on the number of systems, the amount of manual cleanup and whether the data includes personal data. Article 18.1 of 152-FZ, the Russian personal data law, requires a personal data operator to appoint a person in charge, issue a processing policy, run internal control and make staff familiar with the rules. Risks without order: a leak, a fine and decisions based on wrong numbers. Before a project, find out who owns each source and which systems you can trust. The section on data for AI explains what to measure and how to collect such data.
For a specialist
Data governance is the system of authority, roles, policies and processes by which an organization makes decisions about its information assets. In Russian, the same term often also covers data management itself: storage, integration and processing; governance sets the rules for this work. Typical roles: the data owner is responsible for a data set, the data steward looks after definitions and quality, and the data custodian runs the technical side. Tools include a data catalog and metadata, a business glossary, data lineage, quality rules (completeness, accuracy, timeliness, consistency), classification by sensitivity and an access matrix. AI systems add requirements for training and reference sets: source, labeling, bias checks, retention periods and anonymization. In Russia, Federal Law No. 152-FZ of 27 July 2006 sets the frame for personal data.
Example
A distributor wanted an AI assistant to answer managers about stock, but the numbers in the accounting system and the CRM did not match. First, the company appointed an owner of the product catalog and decided that the main version of stock lives in the accounting system. Only after that did the assistant start to answer without contradictions.
Related terms
Questions and answers
Sources
Not sure whether your company needs this?
AI solutions map