A sales enablement team wants to extend Conversation Intelligence so that call sentiment scores and keyword-trigger counts feed a custom Power BI dashboard used for quarterly coaching reviews, updated automatically as new calls are analyzed. Which approach should the consultant use to extend Conversation Intelligence for this purpose?
Select an answer to reveal the explanation.
Short Explanation
The fastest, most reliable way to build a coaching dashboard that stays current is to go straight to where the analysis is already being stored, rather than recreating it somewhere else. Conversation Intelligence writes its scoring and keyword counts into records you can connect a reporting tool to directly, so the dashboard updates itself as new calls come in. A weekly manual export undercuts the whole point of an automatically refreshing dashboard and invites mistakes. Assuming two Microsoft products will just snap together with zero setup is wishful thinking — you still have to build the actual connection. And spinning up a whole separate AI process to re-listen to calls and re-score them from scratch just duplicates work that has already been done, with the added risk that the two sets of numbers will not even agree with each other.
Full Explanation
The correct answer is D. Conversation Intelligence writes call insights, including sentiment scores and keyword-trigger counts, into Dataverse tables, and those tables can be used as a source for a Power BI dataflow or report so the dashboard refreshes automatically as new calls are processed. This satisfies the requirement for an automatically updating coaching dashboard without duplicating any analysis work Conversation Intelligence already performs. Option A is incorrect because a manual, weekly export process is neither automatic nor timely, and it introduces human error and delay that conflict with the stated goal. Option B is incorrect because there is no zero-configuration embedding between Conversation Intelligence and Power BI simply because both are Microsoft products; a real data connection still has to be built and configured. Option C is incorrect because building a separate Copilot Studio agent to re-analyze call recordings duplicates work Conversation Intelligence has already done, risks producing sentiment scores that disagree with the native ones, and adds unnecessary cost and complexity instead of reusing existing insights.