A consultant is setting up conversation intelligence for a sales team that uses Microsoft Teams for all customer calls. The team wants reps to receive an in-call alert whenever a prospect brings up a competitor by name, so the rep can respond immediately rather than after the fact. The consultant has enabled call recording and transcription, but reps report the alerts never appear during live calls. Which configuration step most likely resolves this?
Select an answer to reveal the explanation.
Short Explanation
Real-time alerts like this only work for words the system has actually been told to listen for. Recording the call and turning it into a transcript is just step one, it gives the system text to search through, but it does not automatically know that a rival's name matters. Think of it like setting up a smoke detector: the sensor can be running perfectly, but if nobody's told it what smoke smells like, it won't go off. Here, the fix is adding the competitor names to the list of tracked keywords so the live transcript gets checked against them as the call unfolds. Stretching out how long recordings are kept only affects whether you can pull up an old call later, not whether something pings a rep in the moment. And having reps tag mentions after the call is backwards, by then the moment to react has already passed, and the detection isn't built to learn from manual tags anyway. The keyword list is the missing piece, not the recording pipeline.
Full Explanation
The correct answer is B. In-call keyword alerts only fire for terms the system is explicitly configured to track; recording and transcription supply the raw conversation data, but without the competitor names entered as tracked keywords in conversation intelligence settings, nothing triggers an alert even though the words are being spoken and transcribed. Adding the names to the keyword list is what connects the transcript stream to a real-time notification. Option A is incorrect because retention period controls how long recordings are stored for later review, not whether real-time detection happens during the call; a longer retention window has no effect on live alerting. Option C is incorrect because conversation intelligence keyword detection does not rely on reps manually tagging transcripts to train recognition; it matches configured keyword lists directly against the transcript as the call happens, so manual tagging after the fact would not enable a live alert anyway. Option D is incorrect because predictive opportunity scoring and conversation intelligence keyword alerts are separate capabilities that do not compete for a shared processing pipeline, and disabling scoring would remove a different feature entirely without addressing the missing keyword configuration.