Automation is especially important in controlling access to sensitive information. Data that used to live in silos is being merged into a single pool. That’s good for business because an extensive database or repository makes it easier for analysts to run queries and gain insights from the accumulated data. It’s also bad because security must now replicate barriers and segment data within cloud platforms; it was easier to secure separated systems.

The complexity increases because people want direct access to information for work. It’s easy to add users and data when data sets are aggregated in a cloud-based platform like Snowflake. But you increase the risk of human error when you give users access to that data. An administrator could try to take a shortcut, such as granting access to the entire database, rather than limiting the permissions to a specific user or a specific data set.

We need automation to check what’s in place and respond quickly. For example, if an analyst makes a mistake when creating a new data model or user group, security should flag the problem. It should prompt the user to review the process or block steps if something doesn’t look right.