A consultant enables AI-driven opportunity insights for a sales team and turns on email and Teams call auto-capture, but leaves the meeting recording and transcript capture setting off to save on storage. Reps later notice that Copilot-generated call recaps and conversation intelligence signals reflect only email content, missing the substance of negotiations that happen on video calls. What should the consultant do to fix this?
Select an answer to reveal the explanation.
Short Explanation
If the system was never told to capture the meeting itself, there is nothing for the AI to summarize from that meeting, no matter how good the underlying models are. Email gets picked up, so recaps look fine on the surface, but the actual back-and-forth that happens on a call, where most negotiation and objection-handling occurs, is invisible because that data source was intentionally turned off to save storage. The fix is simply to turn that source back on so transcripts flow into the same pipeline as email. The distractor answers all misdiagnose where the gap comes from: one assumes handwritten notes will somehow outrank the system's own capture, one assumes the AI is mostly reading opportunity fields like revenue rather than conversations, and one assumes the problem is a stale prediction model rather than a missing data source. None of those add the missing call content back in, which is the only thing actually needed here.
Full Explanation
The correct answer is C. Copilot recaps and conversation intelligence draw their substance from captured communication history, and if meeting recording and transcript capture is switched off, the richest source of negotiation detail, the actual conversation, never reaches those features. Turning that setting back on restores the transcript data that email alone cannot provide, which is precisely the gap the reps are describing. Option A is incorrect because manually typed notes are not given priority over automated capture in how recaps are generated, and relying on reps to transcribe every call by hand does not scale and still would not replicate the structured signal transcripts provide. Option B is incorrect because Copilot recaps and conversation intelligence are built from communication content, not primarily from opportunity record fields like revenue or close date, so improving those fields would not add any missing call detail. Option D is incorrect because predictive opportunity scoring and Copilot call recaps are separate capabilities; re-running a scoring model does nothing to bring transcript content into the system, and the root cause here is a capture configuration choice, not a stale score.