A sales enablement lead rolled out Copilot email drafting, next-best-action suggestions, and relationship health scoring across the sales team six months ago. Leadership is now asking for objective evidence of how many sellers actually use each feature and whether usage correlates with seller performance, rather than relying on anecdotal feedback before approving continued investment.
Select an answer to reveal the explanation.
Short Explanation
When someone wants proof that an AI rollout is actually paying off, the temptation is to reach for whatever dashboard is closest at hand, even if it's answering a different question. A model accuracy dashboard will tell you how good your predictions are in a vacuum, but not whether anyone is using them. Raw API traffic numbers will tell you the system is busy, but not which person is engaged or which feature they're touching. A queue of individual recorded calls is great for coaching one seller at a time, but it won't roll up into a team-wide adoption story. What actually answers 'are people using this, and is it helping' is the built-in usage and adoption reporting that sits at the admin level, tracking who's touching each AI capability over time. Pair that with existing performance metrics like win rate, and you get a real before-and-after story instead of a hunch. The lesson: match the report to the actual question being asked, because usage and adoption is a different measurement than accuracy, traffic, or individual call quality.
Full Explanation
The correct answer is C. Dynamics 365 Sales and Copilot provide adoption and usage analytics through the admin and Insights reporting surfaces, showing how many sellers are actively using each AI feature, such as email drafting, next-best-action, or relationship health scores, and these usage figures can be joined against seller performance data like win rate or cycle time to build a genuine ROI case. Option A is incorrect because the AI Builder model performance dashboard reports statistical accuracy of custom prediction models the organization has built, such as precision or recall, and says nothing about how many sellers are opening or acting on Copilot features day to day. Option B is incorrect because Power Platform API call volume is a technical telemetry signal, useful for capacity planning, but it does not map cleanly to individual sellers or distinguish meaningful feature engagement from background system calls. Option D is incorrect because the conversation intelligence call review queue surfaces individual calls for manual review and coaching, not an organization-wide summary of which features are adopted or by whom. For a rollout that needs to justify continued investment with data rather than anecdotes, the admin-level usage analytics are the resource built specifically to answer that question, while the other three surfaces answer adjacent but different questions.