A consultant is helping a sales team decide which AI-driven scoring capability to apply at two different points in their pipeline: prioritizing hundreds of new inbound leads each week, and flagging established opportunities that are quietly losing momentum. The team currently applies a single scoring model to both situations and gets poor results at one of the two stages.
Select an answer to reveal the explanation.
Short Explanation
The team's mistake is trying to make one model do two different jobs, and the jobs need different kinds of evidence. Sorting a flood of brand-new leads calls for a model built on early signals, the kind of information you have the moment someone first shows up. Catching an established deal that's quietly going cold calls for a completely different kind of evidence, the pattern of activity and time spent in each stage over the life of that opportunity, which a brand-new lead simply doesn't have yet. So the fix is using the right tool for each job rather than forcing one tool to cover both. Forcing the late-stage tool onto brand-new leads fails because that data doesn't exist yet for them. Forcing the early-stage tool onto established deals fails for the mirror-image reason, it was never built to read pipeline-progression signals. And giving up on scoring the high-volume, painful stage entirely just abandons the one place where automated prioritization would help the most. Match the model to the stage it was actually trained to read.
Full Explanation
The correct answer is C. Lead scoring is built on early-funnel signals available at first contact, such as firmographic fit and initial engagement, making it suited to prioritizing a high volume of new inbound leads, while opportunity or relationship-based scoring draws on pipeline progression signals like activity cadence and stage duration, which is what is needed to detect an established deal quietly stalling. Using the model matched to each stage's available data is why splitting them produces better results than forcing one model to do both jobs. Option A is incorrect because opportunity scoring depends on pipeline history and stage progression data that a brand-new lead simply does not have yet, so applying it to leads would produce unreliable or meaningless scores despite the appeal of consistency. Option B is incorrect because lead data, while collected early, lacks the stage-duration and activity-cadence signals that stalling detection depends on, so it cannot substitute for opportunity-stage scoring regardless of how complete it seems at intake. Option D is incorrect because it discards a usable AI signal at the lead stage, where the team is already struggling with high volume, in favor of manual triage, which is the exact bottleneck the team is trying to relieve. Matching each model to the funnel stage whose data it was built to interpret is the underlying decision a consultant needs to get right here.