河北发布生态文明建设年度评价结果

What’s changing 

• Multi-model Support: When AI classification first launched, the product supported training a single model for a single label field.  Now, customers can train up to five unique models.  Common use cases for multiple models are:
• Models for different labels
• Models for different fields of a single label
• Multiple models for the same label & field combination, with different training datasets curated for separate audiences 

• On-demand Training: Training AI classification models can be an interactive process.  With the former version of the product, the models would train on a predefined schedule.  Now, the administrator can decide when to train the model, initiating the training process on demand – enabling organizations to move at their own pace! 

• Refreshed UI: We’ve redesigned the AI classification experience from the ground up with a new onboarding flow and model details page.  With the redesigned UI, Workspace Administrators will now see richer insights into the status of model training, metrics on their training data, model recall scores, and a history of their model versions. 

Who’s impacted 

• Admins

Why it matters 

• Powered by privacy-preserving AI models that can be uniquely trained on specific customer needs, AI classification automatically identifies, classifies, and labels files in Google Drive. This helps organizations standardize data classification and achieve labeling consistency at scale. Labels can then be used to trigger rules on files that can and cannot be shared through data loss prevention (DLP) controls, lifecycle management policies, as well as audit and reporting use cases. These latest enhancements give admins the flexibility to train models when they need to and for the specific and dynamic needs of their organization.

Getting started 

Rollout pace 

Availability 

Resources