Access: /leads — top-level menu.
Leads page

​
Definition

A lead = a customer with a fit score on a product. The Leads page asks you to pick a product first, then shows the top customers that match (ranked by overall fit score). There is no stage:lead / stage:mql concept in the system — fit ranking is independent of deal pipeline stages.

​
Page flow

1

Open /leads

The page loads.
2

Pick a product

Product dropdown. Defaults to the first product in your catalog.
3

See the best-fit customer table

Ranked by overall fit score (weighted sum of FitCriterion scores). Each row shows:
  • Customer info
  • Overall score bar (%)
  • Score breakdown per criterion type
  • Bucket (hot / warm / cold based on thresholds)
4

Search / filter

Search box filters the current ranking.

​
Friendly leads table

Columns:
  • Customer name + company
  • Overall fit (% bar)
  • Score breakdown per criterion type
  • Bucket badge

​
Human Review Modal

A “qualify by feedback” modal. One input + submit.
1

Open review

Click the clipboard icon on a lead row.
2

Enter feedback

Textarea — type your notes about this customer (e.g. “Company < 50 employees, not a strong fit”, “Showing buying signals”).
3

Submit

The system:
  1. Appends the feedback to customer.note
  2. Triggers the re-score worker — re-runs fit assessment across every criterion
4

Wait for re-score

Worker runs in the background. The table refreshes with the new ranking.
The feedback you type becomes LLM scoring context for the next run. The more specific, the better the AI learns your org’s qualification patterns.

​
Manual re-score

Customer drawer → ”…” menu → Re-score now — triggers a background re-score for a single customer across every criterion.
Stage-based pipeline (lead → mql → opportunity → customer), Exit — Handoff auto-promotion, bulk qualify with stage tags, 4-button decision (Qualify / Nurture / Disqualify / Need more info), AI-suggested recommended action — coming soon.