Data Classification & Sensitive Information TypesBuilt-in vs Custom SITs

Built-in vs Custom SITs

30 mins

Understanding the Concept

Built-in SITs are pre-configured by Microsoft and cover common sensitive data types globally. They're maintained and updated by Microsoft, requiring no configuration to use.

Custom SITs allow organizations to define patterns specific to their business - employee IDs, project codes, proprietary data formats, or regional identifiers not covered by built-in types.

You can also clone and modify built-in SITs to adjust confidence levels, add keywords, or fine-tune detection for your environment.

Key Points

  • Built-in: Ready to use, Microsoft maintained, global coverage
  • Custom: Organization-specific patterns and keywords
  • Clone option: Modify built-in SITs for fine-tuning
  • EDM (Exact Data Match): Match against actual data tables
  • Trainable Classifiers: ML-based classification for complex content

SIT Types Comparison

Step 1

Built-in SITs

300+ ready-to-use patterns (SSN, Credit Cards, IBANs)

Step 2

Custom SITs

Organization-specific patterns (Employee ID, Project codes)

Step 3

Exact Data Match

Match against actual sensitive data tables

Step 4

Trainable Classifiers

ML models for content types (contracts, resumes)

Why This Matters in Real Organizations

Every organization has unique data that built-in SITs won't cover. Custom SITs ensure your proprietary information is protected. Conversely, over-relying on custom SITs when built-in ones exist wastes effort and may be less accurate.

Common Mistakes to Avoid

Creating custom SITs for patterns Microsoft already provides
Not testing custom SIT accuracy before production
Making regex patterns too broad or too narrow
Forgetting to update custom SITs when data formats change

Interview Tips

  • Explain when you would use each type
  • Describe the process of creating a custom SIT
  • Discuss EDM for exact matching scenarios

Exam Tips (SC-401)

  • Know the process to create custom SITs
  • Understand EDM requirements and setup
  • Know when trainable classifiers are appropriate

Course Complete!

You've finished all lessons

←Previous|→Next|HHome