An enterprise using Dynamics 365 Sales has enabled the AI-driven relationship health scoring feature across all accounts. A regional sales director notices that several long-standing, high-revenue accounts are being scored as low health, even though the account teams report strong relationships and frequent informal contact through phone calls that are not logged in the system. What is the most likely explanation the consultant should investigate first?
Select an answer to reveal the explanation.
Short Explanation
An AI health score is only as good as what it can actually see, and if a relationship is being maintained through phone calls that never get typed into the system, the model is essentially working blind on that part of the story. It is not singling out big accounts to punish them, and it is not some frozen number someone typed in once; it is doing exactly what it was built to do, which is read the logged activity and score based on that. The gap here is not in the AI, it is in the data feeding it. Once the team starts logging those calls, even briefly, the model has real signal to work with and the score should start to reflect the relationship the account team already knows is strong. The lesson is that an AI feature is only as trustworthy as the completeness of the activity data behind it.
Full Explanation
The correct answer is D. Relationship health and engagement scoring in Dynamics 365 Sales AI features are driven by signals the system can actually observe, such as logged emails, meetings, and activities; if account teams are having frequent phone conversations that never get logged into Dynamics, the model has no record of that engagement and will underscore the relationship despite it being healthy in reality. Option A is incorrect because there is no mechanism in the feature that deliberately penalizes high-revenue accounts, and inventing an intentional bias mischaracterizes how the scoring works. Option B is incorrect because relationship health scoring is not restricted to newly created accounts; it applies across the account base and would not explain low scores specifically on long-standing accounts. Option C is incorrect because relationship health scoring is an AI-derived signal based on interaction data, not a static, manually set value, so it changes as engagement data changes. The practical fix the consultant should pursue is ensuring sellers log their phone activity so the model has accurate input to score against.