Business glossaries are a way many businesses harmonize terms across their data. Because data is non-standardized across business units, business glossaries act as a translation layer that makes it easier for other people in your organization find data that might use properties you're unfamiliar with.
For example, a phone number column in your database may have any of the following property names:
Or any other number of options that may or may not be human-readable.
Actually standardizing the data would involve extensive reengineering of the datasets. A business glossary lets you create a standard language (also called a taxonomy or ontology) for your data without modifying the data itself.
You can create a glossary term from the dataset itself by opening the data edit popover and opening the data dictionary builder. Edit the property you'd like to provide a glossary term for.
Similar to Topics, you can either select an existing glossary term, or provide a new one. You may add several glossary terms on a single property, if needed.
You can search for datasets based on their glossary term in the main data page.