Feature

Not Every Mention Is a Lead. SignalPipe Knows the Difference.

Two pipelines — keyword gate → multilingual semantic scoring → sarcasm detection, then a 3-judge AI drafting swarm — surfaces real buyers in any language and rejects everything else.

The hardest part of social selling isn't finding mentions — it's knowing which ones are worth your time. A keyword alert on "looking for a CRM" will return irrelevant posts, sarcastic comments, competitor discussions, and the occasional real lead buried underneath.

SignalPipe runs two sequential pipelines. The 3-stage scoring filter — keyword gate, multi-factor semantic scoring (embedding similarity, urgency, specificity, and keyword density), sarcasm detection — eliminates ~85% of posts cheaply. Urgency and specificity are detected in English, Spanish, French, German, and Portuguese — so posts like "urgente", "dès que possible", "dringend", and "agora mesmo" score correctly without any extra configuration.

Leads that survive scoring go to a 3-judge AI drafting swarm (Skeptic, Analyst, Optimist) that evaluates the lead and writes a contextual reply draft. The result is a stream of scored leads where nearly every post is genuine buying intent — each with a draft ready for your review.

How it works

1

Keyword gate

Fast, cheap pre-filter. Any configured buy-signal keyword must appear before a post is scored. Eliminates ~85% of API costs.

2

Multilingual semantic scoring

Embedding-based semantic similarity is measured against your anchor sentences. Urgency and specificity are detected in English, Spanish ("urgente", "migrando de"), French ("dès que possible", "on change de"), German ("dringend", "wir wechseln"), and Portuguese ("agora mesmo", "trocando de") — no extra setup required.

3

AI drafting swarm

Three judges (Skeptic, Analyst, Optimist) evaluate each scored lead independently via ensemble weighting. Low-intent leads are suppressed by ensemble veto before a draft is written. The Analyst's draft is used by default.

4

RL weight learning

Every approval and rejection you make adjusts the per-product RL weight multiplier. Over time, similar leads are scored higher or lower based on what you've found valuable.

5

ES / FR / DE / PT coverage

Buying-intent phrases in all four major European languages are baked into the scoring model. A Spanish-language post expressing urgency scores just as accurately as an English one — useful for any SaaS with international traction.

See it in action

Join the waitlist. First 100 get Starter free for 3 months.

No credit card. No commitment.

Related