🆕✨Safetica Platform: Smart tags: Classify files with the help of AI

 

 

Introduction: What are smart tags

Smart tags help you discover sensitive files automatically with the help of AI, so you don’t have to create every classification rule manually.

Here’s what happens behind the scenes: Safetica analyzes files across your devices, finds groups of similar files based on their content, and creates a tag for each group. You then review the tags and decide which ones to add to your data classifications. Once a tag is linked to a classification (and that classification is linked to a policy), the matching files are protected. 

 

Key benefits

• Automated data discovery: Safetica finds groups of similar files without any manual setup. You do not need to know in advance what sensitive data exists in your environment.
• Broader protection coverage: Smart tags reveal files outside your existing classifications, closing protection gaps and ensuring more sensitive data is protected by policies.
• Faster setup: Review AI-generated tags and add them to classifications in a few clicks instead of building every rule from scratch.
• Continuous improvement: Smart tags evolve as Safetica analyzes new data over time.
• Protection assessment: Built-in metrics (Protection gap and Already protected) show where each smart tag adds the most value.

 

 

---

Prerequisites

• Have Safetica Client 11.31.30 or newer installed on your devices. Data collection will not run on devices with older Safetica Clients.
• Enable Data discovery: Go to Data classification > Settings > Data discovery settings and set the Status to Enabled.

  • For best results, set Searched file paths to Full local drives. Learn more about Data discovery here.

• Review Content analysis settings: Go to Data classification > Settings > Content analysis settings > Content analysis file types. Only the file types selected here will be analyzed (Recommended, All, and Custom).

 

 

---

How are smart tags created

Smart tag creation happens in four phases:

1. Data collection
2. Smart tag generation
3. Smart tag review

 

 

Phase 1: Data collection

After you enable smart tags, Safetica begins collecting files from devices across your organization. Tag generation starts once the following requirements are met:

• At least 10,000 files across the entire organization
• Files from at least 10 different devices

These requirements ensure the AI has enough variety to produce meaningful tags. The more files and devices are involved, the more precise the resulting smart tags will be.

While collection is in progress, you’ll see a Training in progress progress bar on the Smart tags tab in Data classifications > Settings > Manage classification tags, plus notifications in the Dashboard and Data classification sections.

 

 

Phase 2: Smart tag generation

Once sufficient files are collected (or you start tag generation manually), Safetica generates between 5 and 30 smart tags. This usually completes within one day. Safetica prioritizes quality over quantity. If there isn’t enough meaningful data, you’ll get fewer tags.

When smart tags are ready, you’ll see a notification in the Dashboard and the Data classification sections.

 

 

 

Phase 3: Smart tag review

Generated smart tags appear in Data classification > Settings > Manage classification tags > Smart tags. Each tag has an AI-generated name and is initially not linked to any data classification, which means no policies protect matching files yet.

 

Smart tag detail

 

Click any smart tag to see:

• Name and description: AI-generated from file content and names.
• Used in data classifications: How many data classifications already include this tag.
• File operations (last 14 days): How many recent operations involved files with this smart tag. Click Related records to go to the Data operations section to see the operations.
• Training files: The number of files used to create the tag.
• On devices: From how many devices were files collected.
• Covered by existing classifications: How many existing data classifications already cover some of the training files.
• Sample training files: A sample of up to 50 files (names and paths) based on which the tag was created.

 

Protection assessment

 

When a smart tag is not yet used in any classification, you will also see a Protection assessment with two metrics:

• Protection gap: The percentage of training files not covered by any existing classification. These files would remain unprotected if the smart tag is not used.
• Already protected: The percentage of training files already covered by one or more classifications. You can expand this section to see which classifications apply.

Example: A smart tag named CVs may be applied to some files already protected by an existing classification called HR: Resumes. If only some tagged files are currently classified, adding the CVs tag to that classification increases its coverage and ensures more files are protected.

 

 

---

How to work with smart tags

How to add smart tags to data classifications

To protect files that match a smart tag, add the tag to a data classification. There are two ways to do this:

Option A: From the smart tag detail

1. Open the smart tag’s detail in Data classification > Settings >Manage classification tags > Smart tags.
2. Click Use tag.
3. Add the tag to an existing data classification or create a new one.

Option B: From the data classification rule editor

1. Go to Data classification and open or create a classification.
2. Click Create new rule.
3. Select the Smart tags element.
4. Choose one or more smart tags from the list.

 

Combining smart tags

If the AI generated several specific tags (for example, “Invoices – Safetica services” and “Invoices – Partner billing”) for a broader topic (here “Invoices”), you can combine them in a single “Invoices” data classification.

You can also mix smart tags with other classification elements like keywords.

 

How to rename smart tags

Go to Data classification > Settings > Manage classification tags > Smart tags and click the edit icon next to the tag you want to rename.

 

How to delete smart tags

If a smart tag isn’t useful, you can delete it:

1. Go to Data classification >Settings > Manage classification tags > Smart tags.
2. Click the Delete icon next to the tag you want to delete.
3. If the tag is used in any classification, Safetica will show you which classifications and policies are affected so you can review the impact before confirming.
4. Choose automatic deletion (Safetica updates everything for you) or manual deletion (you remove the tag from classifications yourself first and then click Refresh to re-check dependencies).
5. Provide feedback on why the tag wasn’t useful. This helps Safetica improve future tag generation.

 

 

---

How to view operations with tagged files

Once smart tags are created, any file that matches a tag is automatically tagged with it. You can view operations with tagged files in three ways:

• From the smart tag’s detail, click Related records to see the Data operations section filtered based on that tag.
• In the Data operations section, use the Classification tags filter to display operations tagged with a specific smart tag.
• In an individual file’s detail, check the Classification tab to see smart tags assigned to that specific file.

 

 

---

How to deactivate smart tags

You can deactivate smart tags at any time by clicking the Deactivate smart tags button in Data classification >Settings > Manage classification tags > Smart tags.

 

 

---

Security and privacy

Smart tags are designed with security and privacy at their core:

• On-device processing: Content analysis happens directly on the device. Safetica never reads or stores actual file contents. Instead, the device generates abstract mathematical representations (feature vectors) that capture content patterns. Only these vectors, along with file names and paths, are used to generate smart tags.
• AI-generated names via Azure OpenAI: Smart tag names and descriptions are generated under an enterprise agreement using Azure OpenAI. This ensures that customer data is not exposed outside the secured environment and is not used to train AI models.
• Consistent standards: The entire process follows the same security and privacy standards used throughout cloud-hosted Safetica.

 

 

---

FAQ

Q: Which files are analyzed during smart tag creation?

A: All files with extensions defined in your Content analysis settings (Recommended, All, or Custom). The final groups of similar files are selected based on internal scoring that considers factors like group homogeneity, number of training files, or keyword presence and relevance.

 

Q: Smart tag creation didn’t start even though 10,000 files were collected. Why?

A: The system also requires files from at least 10 different devices to ensure sufficient diversity. If that requirement isn’t met, tag generation may wait for additional files before starting tag generation. You can click Create now to start sooner. 

 

Q: Can smart tags disappear after they are created?

A: Yes. Smart tags that aren’t used in any data classification may be deleted or replaced during continuous tag generations. To keep a smart tag, add it to a data classification.

 

Q: Why are some smart tags very specific?

A: Smart tags tend to be generated for granular groups (for example, “Invoices – Safetica services” tag rather than just “Invoices” tag). To build a broader data classification, you can:

• combine multiple related smart tags
• mix smart tags with other classification elements like keywords
• add smart tags to existing data classifications to strengthen their coverage

 

Q: Why should I give feedback when deleting a smart tag?

A: Your feedback helps Safetica’s AI model improve, so future tag generations are more relevant to your environment.