AI-900 Practice Question: Describe features of Natural Language Processing workloads on Azure

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

• ✗

  NER identifies grammatical parts of speech like nouns and verbs

  Why it's wrong here

  Named entity recognition (NER) does not perform part-of-speech tagging; that is a separate NLP task that labels each token with a syntactic category such as noun, verb, or adjective based on its context. NER instead scans text for proper nouns and other expressions that denote specific real-world entities and assigns semantic categories like Person, Organization, Location, or Date. For instance, in 'Microsoft hired Satya Nadella,' POS tagging labels 'Microsoft' as a proper noun, but NER further classifies it as an Organization. Confusing these two tasks is a frequent misconception on the AI-900 exam.

• ✓

  NER identifies and categorizes named entities like people, organizations, locations, and dates in text

  Why this is correct

  Named entity recognition (NER) is an information-extraction capability that locates and categorizes named entities into predefined types such as PERSON, ORGANIZATION, LOCATION, and DATE. For the sentence 'Bill Gates founded Microsoft in Seattle in 1975,' a typical NER model tags 'Bill Gates' as Person, 'Microsoft' as Organization, 'Seattle' as Location, and '1975' as Date. This matches the entity types supported by Azure AI Language NER and is the exact capability the question is testing.

• ✗

  NER generates new names for products based on brand guidelines

  Why it's wrong here

  Generating new product names is a creative, generative task that falls outside the scope of NER. NER only recognizes and categorizes named entities that already appear in the input text, labeling them with types such as Product, Organization, or Location. Producing novel brand-compliant names would require a generative language model (for example, GPT) fine-tuned with brand guidelines, not an NER model. Thus, NER has no mechanism to invent or suggest names.

• ✗

  NER converts names into anonymous placeholders for privacy

  Why it's wrong here

  NER is purely an extraction and classification process; it does not alter or rewrite the original text. Privacy anonymization is a downstream task that might use NER to first locate sensitive entities and then replace them with generic placeholders like '[PERSON]' to protect identities. The NER model itself only outputs entity labels and offsets — it never generates anonymous replacements. Therefore, converting names into placeholders is an application built on top of NER, not NER's function.

Go deeper

Related to this question