// case file 06 · AI automation

Lead Triage & Scoring

Took 520 messy inbound leads and turned them into a ranked list with a clear call on each (contact now, nurture, or disqualify) and the reason why.

RoleSoloTimelineAug 2026StackPython + CSV520leads cleaned

// 01 the problem

An inbound lead list with inconsistent dates, headcount ranges, budgets in mixed formats and free-text notes. No way to know who to call first.

// 02 how it works

Step 01
Normalised created dates, employee ranges and budgets into clean fields.
Step 02
Validated emails.
Step 03
Scored intent from notes (buying signals, timelines) and fit from title and company size.
Step 04
Applied disqualification rules and wrote the reason for every score.
Step 05
Produced a ranked list with a recommendation per lead.

// 03 decisions that mattered

A ranked list nobody can explain doesn’t get used.

// 04 results

520leads cleaned

167marked “contact now”

144disqualified with reasons

// 05 takeaway

Scoring is only useful if a salesperson can argue with it.

PythonCSVScoring rules

// artefacts

Let’s talk