How Lead Scoring Works
A score you cannot explain is a score your reps will not trust. LeadScore AI grades every lead against your ICP and writes down its reasons — this guide unpacks each piece of that receipt.
5 MIN READ · UPDATED JULY 2026Four Signals, One Score
The engine reads fit (how well the lead matches your ICP), contactability (can you actually reach them), engagement (what they have done), and timing (what is happening around them). Together those produce the score and its supporting grades.
The Receipt: Grade, Band, Verdict, Reason
- Fit grade — how closely the lead matches your ideal customer
- Intent band — how warm their behavior and signals look right now
- Verdict — the call to action: hot, worth working, nurture, or disqualify
- Written reason — a plain-English explanation a rep can read out loud on a call
Scores Move — On Purpose
Recency decay means truth has a shelf life: a lead that was hot in March is not hot forever. Replies warm a score, opt-outs suppress it, and do-not-contact is a hard gate the automation cannot cross.
Humans Can Overrule the Machine
A rep who knows the account can override the verdict — hot, cold, or disqualified — with a reason that sticks. Overrides feed back into the model along with your marked outcomes, so the ranking keeps getting sharper.
When a rep disagrees with a score, that disagreement is data. Record the override instead of ignoring the lead — the model learns from it.
Try It on Your Own Leads.
Every guide is shorter in practice — import a list and see the workflow run.