A customer support team receives thousands of open-ended feedback comments each week and wants software that automatically pulls out the main topics and terms mentioned in each comment, such as 'shipping delay' or 'packaging damage', without judging whether the comment expresses a positive or negative opinion. Which natural language processing capability should the team use?
Select an answer to reveal the explanation.
Short Explanation
Think about what the team actually wants out of each comment: the topic, not the tone. They don't care yet whether 'shipping delay' was said angrily or calmly, they just want to know shipping delay came up. That's a completely different job from scoring emotion, which would tell you how someone felt but not what they were talking about. It's also broader than hunting for proper nouns like a company or product name, since ordinary phrases like packaging damage aren't names at all, they're topics. And figuring out what language the comment is written in doesn't get you any closer to knowing its subject matter. The tool built for pulling the core topics and terms out of free text, leaving the emotional judgment for a separate step, is the right fit here.
Full Explanation
The correct answer is B. The team wants the main subjects of each comment surfaced, such as 'shipping delay,' with no judgment about whether that topic was described positively or negatively, which is exactly what key phrase extraction does: it pulls out the salient terms and topics from a block of text. Option A is incorrect because sentiment analysis produces a positive, negative, neutral, or mixed score for the emotional tone of text, which is the one output the team explicitly said it does not need. Option C is incorrect because language detection identifies which language a piece of text is written in, a preprocessing step unrelated to surfacing topics within comments already assumed to be readable. Option D is incorrect because named entity recognition tags specific proper nouns like people, organizations, or locations, which is narrower than general topic terms; 'shipping delay' and 'packaging damage' are ordinary phrases, not named entities, so this capability would miss most of what the team needs. Key phrase extraction is the workload built specifically for topic surfacing at scale.