Which Purview Classification Type Fits Employment, Offensive Language, and Medical Emails?
Short answer
Employment offers use a trainable classifier; offensive language uses a trainable classifier; medical terms and conditions use a sensitive information type.
Question
You have a Microsoft 365 E5 subscription.
You need to label Microsoft Exchange Online emails that match the following conditions:
• Contain employment offers
• Contain offensive language
• Contain medical terms and conditions
The solution must minimize administrative effort.
Which type of data classification should you use for each condition? To answer, drag the appropriate data classification types to the correct conditions. Each data classification type may be used once, more than once, or not at all. You may need to drag the split bar between panes or scroll to view content.
NOTE: Each correct selection is worth one point.
Key Takeaways
The item tests whether you can separate semantic trainable-classifier use cases from a built-in sensitive information type; the trap is assuming all medical content requires a trainable classifier.
This SC-401 drag-and-drop item asks you to map Exchange Online email conditions—employment offers, offensive language, and medical terms—to Microsoft Purview classification types while minimizing administrative effort. The page establishes that trainable classifiers handle the first two semantic conditions and a built-in sensitive information type handles medical terms.
Common mistake: A common mistake is selecting trainable classifiers for all three conditions because they are machine-learning based, overlooking the built-in All Medical Terms And Conditions sensitive information type that directly matches the medical condition.
Expert Analysis
Why the Answer Is Correct
The first two conditions describe semantic, context-heavy email content. Microsoft Purview's trainable classifiers are built for unstructured text such as employment offers and offensive language, and the pre-trained Offensive Language classifier directly covers the second condition. For the third condition, Microsoft provides a built-in sensitive information type named All Medical Terms And Conditions, which is purpose-built to detect medical terminology without custom model training. That mapping—trainable classifier, trainable classifier, sensitive information type—minimizes administrative effort while matching each condition to the classification mechanism designed for it. The comments from gabbinu, choukou, and Eltooth consistently point to the same trainable/trainable/SIT split.
Why Other Candidate Answers Are Wrong
Choosing a sensitive information type for employment offers or offensive language would force keyword lists or patterns onto semantic content that varies widely, producing weak matches and more tuning. Choosing a trainable classifier for medical terms and conditions ignores the existing All Medical Terms And Conditions SIT and would add unnecessary model training or configuration. The all-trainable answer proposed by TheWiz is understandable because classifiers are ML-based, but it is not the minimum-effort design once the built-in SIT is considered. Similarly, treating all three as SITs would miss the built-in Offensive Language trainable classifier and the unstructured nature of employment offers.
Community Comment Notes
Several learners converged on the trainable/trainable/SIT mapping, with gabbinu explaining that employment offers are best caught by a classifier trained on examples and that medical T&Cs fit a SIT. RickySmith linked directly to Microsoft's trainable-classifier definitions and the All Medical Terms And Conditions SIT, which supports the official basis for the split. TheWiz initially argued "Trainable, Trainable, Trainable" to minimize effort, then revised after recalling the built-in medical SIT: "On further reflection, Trainable, Trainable, SIT." vixxx83 also noted that Microsoft has a built-in Offensive Language trainable classifier, reducing the need for custom training.
Official Reference
- https://learn.microsoft.com/en-us/purview/trainable-classifiers-definitions
- https://learn.microsoft.com/en-us/purview/sit-defn-all-medical-terms-conditions
These links open the official vendor documentation in a new tab.
Exam Strategy
For drag-and-drop classification items, ask whether the condition is semantic and unstructured (trainable classifier) or a predefined pattern/list (sensitive information type). Then check Microsoft's built-in classifier and SIT lists before proposing custom training, because built-ins minimize administrative effort.
Frequently Asked Questions
Why does 'medical terms and conditions' use a sensitive information type instead of a trainable classifier?
Is there a pre-trained trainable classifier for offensive language in Exchange Online?
Discussion 11 contributions
Compiled from public community threads and reader submissions, lightly edited for clarity. Handles and vote tallies are shown as recorded, and comments reflect the views of individual contributors.
From public community threads 11
-
gabbinu 👍 18Contain employment offers → Trainable classifier (Unstructured content; best caught by a classifier trained on examples.) Contain offensive language → Trainable classifier (Semantic/phrase-based detection fits a classifier, not SIT/EDM.) Contain medical terms and conditions → Sensitive info type (Use a SIT with keyword dictionary/phrases for the medical T&Cs.)
-
RickySmith 👍 53 - https://learn.microsoft.com/en-us/purview/trainable-classifiers-definitions 3 - https://learn.microsoft.com/en-us/purview/trainable-classifiers-definitions 2 - https://learn.microsoft.com/en-us/purview/sit-defn-all-medical-terms-conditions
-
choukou 👍 4Solution : 1: Trainable 2 : Trainable 3: SIT
-
Gtc_ 👍 3I would think trainable, trainable, SIT.
-
Eltooth 👍 3Trainable Trainable SIT
Show 6 more comments
-
MultiCloudIronMan 👍 2Contain employment offers Trainable classifier Detects context-rich, unstructured content using machine learning. Contain offensive language Trainable classifier Ideal for nuanced, behavioral content not easily captured by pattern matching. Contain medical terms and conditions Sensitive info type Matches structured patterns like medical codes, terms, and identifiers.
-
vixxx83 👍 1Trainable classifier For all 1. Contain offensive language Microsoft has a built-in Offensive Language trainable classifier. No custom training required. 2. Contain medical terms and conditions Microsoft has a built-in Medical Condition and Terms trainable classifier. This classifier was created specifically to identify health-related terminology and medical content. 3. Contain employment offers Microsoft provides an Employment trainable classifier. It is specifically designed to identify employment-related content such as: Job offers Employment agreements HR-related employment content
-
NeerajSap 👍 1Employment offers → Trainable classifier (built-in classifier) Offensive language → Sensitive info type (keyword/pattern based) Medical terms and conditions → Trainable classifier (built-in classifier)
-
TheWiz 👍 1Trainable, Trainable, Trainable - I believe all three of these can be accomplished with pre-trained trainable classifiers, which would minimize admin effort.
-
kollyp2000 👍 1Correct: Employment offers: Trainable classifier Offensive language: Trainable classifier Medical terms and conditions: Sensitive info type
-
TheWiz 👍 1On further reflection, Trainable, Trainable, SIT There is a built in SIT for All Medical Terms And Conditions that would work.
Reader comments & corrections
No reader corrections yet — spotted an error or have a note? Share it below.