CRM Lead-Qualifying Automation

Overview

A CRM is only useful if it's current and the team can trust the priority order. Through Agenture I built qualifying automation that did the unglamorous work — reading, scoring, and filing every lead — so small teams worked a clean, ranked pipeline instead of a stale spreadsheet.

The problem

Inbound leads arrive in different shapes from different channels, and qualifying them by hand is slow and inconsistent. Records go stale, duplicates creep in, and the team ends up working leads in arrival order rather than by likelihood to close. The highest-intent enquiry can sit behind a tyre-kicker simply because nobody scored them.

What I built

• Automatic intake + enrichment — every lead, whatever channel it came from, was parsed into the same structured fields and enriched before it hit the CRM.
• Rubric-based scoring — a defined rubric (budget, fit, intent, urgency) combined deterministic rules for must-haves with LLM judgement for nuance, returning a score with its reasoning, not just a number.
• Tiering + routing — leads were tiered (A/B/C), with A-tier surfaced for immediate action and lower tiers routed into nurture.
• CRM hygiene — dedupe, field normalisation, and status updates ran continuously so the CRM stayed clean without manual upkeep.

Impact

• Focused the team's time on the leads most likely to convert, rather than whoever happened to come in first.
• Kept the CRM clean and current automatically, removing recurring manual data entry.
• Made qualification consistent and explainable instead of a gut call that varied by who was on.