B2B 线索先评分,再分配下一步
给代理商客户搭建询盘分级:先存原始资料,再用 Gemini 按预算与项目需求打分,高意向自动发预约链接并提醒销售,异常转人工复核。
值得借鉴的,是这一点。
先保留原始数据,再让模型分流。评分只有对应清晰的下一步,才会从“分析”变成可执行工作。
Reddit 原帖
在 Reddit 查看I got tired of my client losing hot B2B leads to slow response times, so I built an AI Lead Qualification Engine (n8n + Gemini 2.5 Flash)
The Problem My client runs an agency and was getting dozens of project inquiries every month. The issue was that their team had to manually read through every single submission, figure out who was actually serious, and write back.
Most of the leads had zero budget or vague details. By the time the team filtered out the junk and finally replied to the high-budget, ready-to-go clients, those hot prospects had already hired someone else. Hours were being wasted on manual review, and they were losing money on slow response times.
The Solution I completely removed the human bottleneck. I built an automated system that captures the inquiry, uses AI to score the lead from 1 to 10 in seconds, and intelligently routes it based on that score.

















